document resume se 056 829 author gott, …document resume ed 390 642 se 056 829 author...

DOCUMENT RESUME

ED 390 642 SE 056 829

AUTHOR Gott, Richard; Duggan, SandraTITLE Investigative Work in the Science Curriculum.

Developing Science and Technology Education.REPORT NO ISBN-0-335-191143-6PUB DATE 95

NOTE 145p.

AVAILABLE FROM Open University Press, Celtic Court, 22 Ballmoor,Buckingham, England MK18 1XW, United Kingdom(paperback: ISBN-0-335-19143-6; hardcover:ISBN-0-335-19144-4).

PUB TYPE Books (010)

EDRS PRICE MF01/PC06 Plus Postage.DESCRIPTORS Educational Change; Elementary Secondary Education;

Foreign Countries; *Investigations; ScienceCurriculum; Science Experiments; Science Instruction;*Science Process Skills; Science Teachers;*Scientific Methodology; *Technology Education

IDENTIFIERS England

ABSTRACTTeaching is essentially a personal and professional

business in which lively, thinking, enthusiastic teachers continue toanalyze their own activities and mediate the curriculum framework totheir students. This book aims at encouraging teachers and curriculumdevelopers to continue to rethink how science and technology shouldbe taught in schools. It is based on research on students' abilitiesto carry out investigations in both primary and secondary schools, aswell as the authors' experience of using investigations in theclassroom. It describes in careful and authoritative detail just whatdoing investigative work means, how students tackle and aevelop suchwork, and how it can be encouraged and assessed. Chapters include:(1) "Changing Views of Practical Science"; (2) "AlternativePerspectives"; (3) "Investigations: What are They?"; (4) "Pupils'Performance of Investigations in Secondary Schools: An Overview"; (5)

"The Performance of Investigations in Secondary Schools: A DetailedLook"; (6) "Investigations and Teaching"; (7) "IncorporatingInvestigations into a Scheme of Work"; (8) "Assessment"; (9)

"Investigations and the UK National Curriculum"; and (10)"Postscript". (JRH)

*

*Reproductions supplied by EDRS are the best that can be made

from the original document.******************************i.A**:,*****i;"***.AAA*********i *******

*

*

DEVELOPING SCIENCE ANDTECHNOLOGY EDUCATION

INVESTIGATIVEWORK IN THESCIENCECURRICULUMRICHARD GOTT ANDSANDRA DUGGAN

**PERMISSION TO REPRODUCE THISMATERIAL HAS BEEN GRANTED BY

(-4

TO THE EDUCATIONAL RESOURCESINFORMATION CENTER (ERIC).-

BEST COPY AVAILABLE

2

U S DEPARTMENT OF EDUCATION011...0 of Fclurahoo,o nosearch and lenr,datment

EDUCATIONAL RESOURCES INFORMATIONCENTER (ERIC)

h°i's document has been reproduced asreceived from the person or orgarwationoriginating itMinor changes have been made toimprove reproduction quality

Points of view or opinions stated in thisdocument do not necessarily representofficial OERI position or policy

Investigative Work in theScience Curriculum

DEVELOPING SCIENCE ANDTECHNOLOGY EDUCATION

Series Editor. Brian Woolnough,Department of Educational Studies, University of Oxford

Current titles:

John Eggleston: Teaching Design and TechnologyRichard Gott and Sandra Duggan: Investigative Work in the Science CurriculumDavid Layton: Technology's Challenge to Science EducationKeith Postlethwaite: Differentiated Science TeachingMichael J. Reiss: Science Education for a Pluralist SocietyJon Scaife and Jerry Wellington: Information Technology in Science and Technology EducationJoan Solomon: Teaching Science, Technology and SocietyClive Sutton: Words, Science and LearningBrian Woolnough: Effective Science Teaching

4

Investigativ( Work in theScience Curriculum

RICHARD GOTT and SANDRA DUGGAN

Open University PressBuckingham Philadelphia

-t)

Open University PressCeltic Court22 BallmoorBuckinghamMK18 1XW

and1900 Frost Road, Suite 101Bristol, PA 19007, USA

First published 1995

Copyright © Richard Gott and Sandra Duggan 1995

All rights reserved. Except for the quotation of short passages for thepurpose of criticism and review, no part of this publication may bereproduced, stored in a retrieval system, or transmitted, in any form orby any means, electronic, mechanical, photocopying, recording orotherwise, without the prior written permission of the publisher or alicence from the Copyright Licensing Agency Limited. Details of suchlicences (for reprographic reproduction) may be obtained from theCopyright Licensing Agency Ltd of 90 Tottenham Court Road. London,W1P 9HE.

A catalogue record of this book is available .from the British Library

Library of Congress Cataloging-in-Publication Data

Gott, Richard, 1946-Investigative work in the science curriculum/Richard Gott.

Sandra Duggan.p. cm. (Developing science and technology education)

Includes bibilographical references and index.ISBN 0-.335-19144-4 ISBN 0 335-19143 6 (pbk.)I. Science Study and teaching- England Curricula. 2. Curriculum

planning- England. I. Duggan. Sandra. 1949- . II. Title.III. Series.Q183.4.G72E534 1994

507' . 1'241 dc20

This book is based on research commissioned by the National CurriculumCouncil and carried out by the Exploration in Science Team at the Universityof Durham. The research was directed by Richard Gott. Ken Eoulds ofSeaham Comprehensive School, County Durham was the project leader.responsible for the design and implementation of the project, much of its analysisand the ensuing research report.

Typeset and illustrated by PanTek Arts. Maidstone, KentPrinted in Great Britain by St Edmundsbury Press. Bury St Edmunds. Suffolk

Contents

Series editor's prefaceForeword

What is this book about?The structure of the bookAcknowledgements

GlossaryAbbreviations

CHAPTER 1

CHAPTER 2

CHAPTER 3

911

11

11

11

1315

Changing views of practical science 17The emergence of practical science 17

Early beginnings 17The Nuffield schemes 18The 'processes and skills' movement 19Investigations 20

Practical science today 21Summary 23References 24

Alternative perspectives 25A 'content' perspective (an epistemological view) 25

A model of science 25Developing the model 28

The thinking skills approach and CASE 34Piaget's developmental model 35The CASE project 36

Concepts of evidence and CASE 38Summary 39References 39

Investigations: What are they? 41Problem-solving and investigations 41The Assessment of Performance Unit 41

f

CHAPTER 4

CHAPTER 5

CHAPTER 6

INVESTIGATIVE WORK IN THE SCIENCE CURRICUI UM

Skills, concepts of evidence and the APU problem-,,olvingmodel 43

The Open-ended work m Science ( OPENS) project 43

The Assessment of Practical Work in Science ( A PWIS ) 47

Investigative work in science 47

A note on the limitations offocusing on the variable siructure of investigations 48

Summary 49

References 50

Pupils' performance of investigations in secondary schools: An overview 51

Factors affecting performance 51

The NCC saniple, design, methodology and analysis 52

Sample 52

The design 53

Methodology 55

The analysis 55

The effect of substantive concepts 56

The effect of the level of difficulty of the concept 57

The effect of procedural complexity 57

The effect of age 58

The effect of context 59

The effect of openness 59

The relative significance of these factors 60

Motivation, expectations and perceptions 60

Teachers' perceptionsEvidence from HMI surveys 63

A note on the effect of teaching about investigations and concepts of evidence 63

Summary 64

References 65

The performance of investigations in secondary schools: A detailed look 67

Skills 67

Concepts of evidence 69

Design 69

Measurement 73

Data handling 75

Evaluation 83

Summary 84

References 85

Investigations and teaching 87

Int roduc t ion 87

Changing the overall lOcus ofinvestigations 87

Targethig specific concepts- (q- evidemv within whole invevigationsDesign 89

Measurement 90

CONTENTS

CHAPTER 7

CHAPTER 8

CHAPTER 9

Data handlingEvaluation reliability and validity

The reinforcement of concepts of evidenceAssociated with designAssociated with measurementAssociated with data handlingAssociated with evaluation

Targeting conceptual understandingProgression and differentiation by outcomeSummaryReferences

7

91

93939393959595969898

Incorporating investigations into a scheme of work 99Progression 101

Some practical considerations 103A note on introducing the language of investigations 103Organising investigations in the science laboratory 104Prompt sheets 104

The role of the teacher in the investigation 105Reporting-back session 106

A note on resources and safety 106Swnmary 107References 107

Assessment 109What is to he assessed? 110

Traditional assessment 110Assessing procedurdl understanding 111

Assessing the recall and use of skills 112Assessing the understanding and application of individual concepts

of evidence in short assessment tasks 113Assessing the synthesis of concepts of evidence 113

Using investigations for assessment purposes 113Choosing the investigation 113Collecting evidence for assessment 115Making judgements 121

The understanding df individual concepts qf evidence withW investigations 123Using simple task scores 123

7/w synthesis qf concepts of evidence within investigations 124Summary 125References 125

Investigations nnd the UK National CurriculumThe development of the National Curriculum

The present National Curriculum77w research evidence and tlw present curriculum

127

127

129

131

8

CHAPTER 10

Index

INVESTIGATIVE WORK IN THE SCIENCE CURRICULUM

The ways forward 132

A structure based on the complexity and sophistication of evidence 132

A structure bpsed in a norm-referenced framework 136

GCSE assessment of Scl 137

Summary 137

References 137

Postscript 139

Science is hard? 140

Science is irrelevant? 140

Science is boring? 141

A bad image? 141

An alternative science? 142

References 143145

Series editor's preface

It may seem surprising that after three decades ofcurriculum innovation, and with the increasingprovision of a centralised National Curriculum, itis felt necessary to produce a series of books whichencourages teachers and curriculum developers tocontinue to rethink how science and technologyshould be taught in schools. But teaching cannever be merely the 'delivery' of someone else's'given' curriculum. It is essentially a personal andprofessional business in which lively, thinking,enthusiastic teachers continue to analyse their ownactivities and mediate the curriculum frameworkto their students. If teachers ever cease to be criti-cal of what they are doing, then their teaching,and their students' learning, will become sterile.

There are still important questions which needto be addressed, questions which remain funda-mental but the answers to which may varyaccording to the social conditions and educationalpriorities at a particular time.

What is the justification for teaching science andtechnology in our schools? For educational orvocational reasons? Providing science and tech-nology for all, for future educated citizens, or toprovide adequately prepared and motivated stu-dents to fulfil the industrial needs of the country?Will the same type of curriculum satisfactorily meetboth needs or do we need a differentiated curricu-lum? In the past it has too readily been assumedthat one type of science will meet all needs.

What should be the nature of science and tech-nology in schools? It will need to develop both themethods and the content of the subject, the way a

scientist or engineer works and the appropriateknowledge and understanding, but what is the rela-tionship between the two? How does the student'sexplicit knowledge relate to investigational skill,how important is the student's tacit knowledge? Inthe past the holistic nature of scientific activity andthe importance of affective factors such as commit-ment and enjoyment have been seriously under-valued in relation to the student's success.

And, of particular concern to this series, what isthe relationship between science and technology? Insome countries the scientific nature of technologyand the technological aspects of science make thesubjects a natural continuum. In others the curricu-lum structures have separated the two, leaving theteachers to develop appropriate links. Underlyingthis series is the belief that science and technologyhave an important interdependence and thus manyof the books will be appropriate to teachers of bothscience and technology.

There have been few changes in school scienceteaching over the past few years that have beenmore significant than the change in the type ofpractical work being done. Moving from the ubiq-uitous 'cookery book' type practical, in which thestudents followed the teacher's recipe, through the'guided discovery' experiment and the experimentsdesigned to develop and test specific practical skills,we have now the practical investigation where thestudents are expected to design, deliver and evalu-ate their own experiment. This move, which hadbeen advocated and practised by a minority sincescience teaching began, has now been enshrined in

1 1

10 INVESTIGATIVE WORK IN THE SCIENCE CURRICULUM

the National Curriculum for Science in Englandand Wales. Though, in this form, it stresses a purescience model of investigation as distinct from amore technological problem-solving type of investi-gation, it has done much to introduce genuinescientific activity into schools.

Among the most influential in bringing aboutthis change have been Richard Gott and his col-leagues, first through his APU work, then in hisActive Science textbooks, and subsequently by hisown research work into the way students andteachers actually do investigations. And it is on thiswork that this scholarly and perceptive book byRichard Gott and Sandra Duggan has been based.Much of the early discussion about the 'best type'of practical work for schools was based more onrhetoric than on research. This important book istimely, for it spells out in careful and authoritative

1 4

detail just what doing investigative work means,how students tackle and develop such work, andhow it can be encouraged and assessed. It shouldgo a long way to underpinning investigative practi-cal work so that, despite the logistical difficultiesinvolved in .uch work, it becomes securely based asthe most important type of practical work inschools. Such holistic investigational practicalwork, whether it is of a pure scientific nature ormore technologically problem-solving, is of vitalimportance. It develops not only the students'knowledge, understanding and practical skills butalso the all-important affective aspects of enjoy-ment, motivation, commitment and self-confidenceI hope and believe that this book will help todeliver such in both science and technology lessons

Brian E. Woolnough

Foreword

What is this book about?

Some years ago now, one of us (R.G.) was respons-ible, with others, for the assessment of scienceinvestigations for the Assessment of PerformanceUnit (APU) team based at Leeds University. Onmoving to Durham, the ideas developed at Leedswere taken into a local school and, over a number ofyears, incorporated into a workable curriculum.Concurrent with these developments came theNational Curriculum. The team at Durham wascontracted by the National Curriculum Council(NCC) to undertake research into pupils' ability tocarry oul investigations in both primary and sec-ondary schools. The aim of the research was,primarily, to inform any subsequent revisions to thestructure of the 'investigations' element of science.The experience of using investigations in the class-room, together with the extensive research materialgathered for NCC, form the basis of this book.

The structure of the bock

This book has evolved in the writing. It arose out ofa desire to publish the results of the research.Reflecting on the findings of this research has led usto develop our thinking and moved us forward intwo ways: in developing the theoretical model whichhas been latent in our thinking, and in consideringhow the implications of the research can be appliedin the classroom.

The first chapter locates the changing role ofpractical work hisiorically. The reader is advised

that it would be unwise to omit the second chap-ter, which focuses on the theoretical framework,since it discusses the thinking which underliesmuch of the rest of the book: The research find-ings are confined to Chapters 4 and 5, whileChapters 6. 7, 8 and 9 are concerned with teach-ing and assessment.

Acknowledgements

We would like to acknowledge the assistance andsupport of a number of people during the process ofwriting this book: Ken Foulds and RosemaryFeasey for their discussions of the NCC projectwork; Helen Costello for the useful discussions andher support and advice throughout the developmentof the hook; all those people who have taken thetime and trouble to read and comment on the manu-script, including Ruth Jarman (Queen's University,Belfast) and members of thc Exploration of Scienceteam in Durham. We would also like to thank PhilipAdey for his comments on Chapter 2; our respectivepartners for their comments on the manuscript andtheir forbearance; and, last but not least, the rest ofour families for their patience.

The authors and publishers would also like toacknowledge the following for permission toreproduce material: Taylor and Francis, NewtonD.P. and Newton L.D. (drawing on p. 141) andthe Northamptonshire Inspection and AdvisoryService (Table 7.4.). Figures 1.2, 1.3 and Table 8 3are reproduced by permission of Simon andSchuster Education, Hemel Hempstead, UK.

1 a

Glossary

Application

Categoric variable

Concepts

Conceptualunderstanding

putting knowledge to use in an unfamiliar or novel situation.

a variable which is defined descriptively, e.g. shape, colour, type of material.

substantive concepts the facts, laws, theories and principles of science (e.g.gravity, photosynthesis, solubility). Also known as declarative concepts.

concepts of eviden,:e the concepts associated with procedural understanding.They include for instance the concept of the fair test, identification of variablesas independent and dependent. validity and reliability.

the understanding of substantive concepts (see above).

Continuous variable a variable which is defined numerically and which can take any value, e.g. height.

Control variable the variable(s) which must be kept constant while the independent variable ischanized in order to keep the test 'fair'.

Dependent variable the variable which changes, and is measured or judged, each time the indepen-dent variable is changed.

Derived variable a variable which is derived from more than one measurement. e.g. speed or rate.

Discrete variable a variable which is defined numerically but which takes only integer values,e.g. number of layers of insulation material (cf. continuous variable).

Epistemology the methodology or basis for knowledge.

Exploration a very open kind of task. where pupils are given the context only and allowedto raise their own questions and then follow these up to find the answer. Thiscould involve practical work, surveys or a library search. It will include prob-lem-solving (and investigations).

1


Heuristic method a system of education in which the pupil is 'taught' to find out things forhim/herself.

Independent variable the value chosen and manipulated by the investigator.

Investigation a specific type of problem-solving defined as 'a task for which a pupil cannotimmediately see an answer or recall a routine method for finding it.'

Key Stages the four age ranges encompassed by the UK National Curriculum: Key Stage I,5 -7-year-olds; Key Stage 2, 7-11-year-olds; Key Stage 3, 11-14-year-olds; andKey Stage 4, 14-16-year-olds.

Problem-solving the solving of problems, i.e. any activity that requires pupils to apply their under-standing in a new situation. Investigations are one type of problem-solving.

Procedural the understanding required in knowing how to do science. It is defined here asunderstanding the understanding and application of (skills and) concepts of evidence.

Procedural understanding is complementary to conceptual understanding.

Piocesses the cognii ,ve processes associated with any intellectual activity including thesolving of scientific problems. They include hypothesising. interpreting, pre-dicting. etc.

Programme of study in the UK National Curriculum, the required teaching for each key stage.

Skills those activities which are necessary but not sufficient in themselves to the carry-ing out of most practical work, e.g. the mechanics of the use of measuringinstruments, how to construct a graph.

Substantive structure the structure of science concerned with its declarative concepts (see also sub-stantive concepts).

Syntactic structure

Synthesis

Understanding

Variables

the structure of skills and concepts of evidence (see above) describing howskills and concepts of' evidence are put together to solvc a problem (cf. syntaxin language).

the putting together of separate bits of knowledge into a connected whole sothat a new structure emerges.

the ability to explain and interpret information within a given context (cf.app'ication).

any observation which can be described by different values, e.g. the colour of abird, length or weight.

Abbreviations

APU Assessment of Performance Unit

APW1S Assessment of Practical Work in Science

AT Auainment Targets in the UK National Curriculum 'things which pupilsshould know, understand or be able to do'

NCC

OPENS

Scl

SEAC

National Curriculum Council

the Open-ended work in Science project funded by the Department of Educationand Science

the attainment target in the UK National Curriculum which focuses oninvestigations

Schools Examination and Assessment Council

CHAPTER 1

Changing views of practical science

The emergence of practical science

Early beginnings

To discover why investigations have conic tooccupy the position that they now hold withinpractical science, we must first go back in time andconsider the emerizence of practical work withinthe science curriculum. In other words, before wecan answer the question 'Why do investigations?'.we must ask 'Why do practical work at all?'

The historical origins of practical work in sci-ence have been well documented by Gee andClackson (1992). Briefly, practical work began toemerge in England soon after the Great Exhibitionof 1851 when state support grants were made avail-able and the Department of Science and Art wasestablished in 1854. Money was provided to set upand equip school science laboratories, although theemphasis in these early days was very much ondemonstrations by the teacher with a clear focus onthe illustration of particular concepts. The increas-ing recognition that practical work was essential inscience education stemmed primarily from twoschools of thought: the recognition of the socialand economic importance of science, and the philo-sophical arguments which had arisen from theworks of Huxley and Spencer and were now beingdeveloped by Armstrong. Armstrong advocatedthe 'heuristic' approach. in which the pupil istrained to find out things for him/herself. It isbased on a belief in the effectiveness of learningthrough action as opposed to the passive assimila-

tion of knowledge. Armstrong (1896:42) believedthat: 'Knowledge alone is not power: but theknowledge how to use knowledge is.'

Armstrong held what would now be consideredto be somewhat extreme views in which he cobsid-ered lectures and textbooks in a rather neeativelight to say the least, urging his students:

Don't look at a text-book: avoid most of them asyou would poison. Their methods are as a ruledetestable and destructive of all honest efforttowards development of powers of sellhelpful-ness...you must never be satisfied with lecturesalone if you wish to do more than spend your timepleasantly...the student of any branch of naturalscience must go to the bench and work hard there.

(Armstrong. 1896:43.50)

This precipitated a move towards encouragingindividual practical work in schools, in the waythat we now take for granted.

However, by the turn of the century, the heuris-tic movement was beginning to fall out of favourbecause it had assumed that scientific conceptscould be discovered by 'common sense'. As scien-tific knowledge increased, this idea was no longertenable. At the same time, there was evidence fromexperimental psychologists suggesting that thetransfer of training in scientific method from oneproblem to another was not, as Armstrong hadassumed, a common occurrence (see, for instance,Wellington, 1989). As a result, the pendulum swungtowards 'content' and away from the emphasis on'method'. Wellington describes how the humanising

INVESTIGATIVE WORK IN ME SCIENCE CURRICULUM

influence of science was promoted at this firm,since it was felt that it had previously been ignoredand that 'science was to be taught for the benefit ofthe learner and not for the benefit of science itself. .

We shall return to these ideas later when weconsider the notion of 'audience' in the next chap-ter and its influence on what children perceive asthe purpose of collecting data from experimentalwork. However, the consequence of this pupil-cen-tred approach to science education was that thestatus of practical work per se was lower than ithad previously been and subservient to theory andcontent. Most of the practical work at that timeconsisted of following 'recipes' to verify theory orto illustrate concepts and, towards the end of thisperiod, there was growing concern that much ofthis practical work was routine and repetitive.Nevertheless, practical work had already estab-lished itself as a vital part of the science curriculumin that there was no longer any great argument asto its desirability, albeit in a supportinil role, in theclassroom or laboratory.

The Nuffield schemes

While practical work in science was and continuesto be regarded as essential. its precise purpose ismuch less certain. In 1959, the Kerr inquiry wascommissioned to inquire into the nature and par-po-, of practical work in school science teachingin England and Wales. The report found that:

When science students were asked about the influ-ence of practical work done at school. few of themthought the finding-out element had been givenan important place nor were they particularlyaware of being led towards a scientific way ofthinking and behaving. The members of theInquiry Team think a much more direct and delib-erate attempt should be made to teach for theseends through practical work. No aspect of scienceeducation is more urgently in need of attention.

(Kerr, 1964:95)

Thus the pendulum swung back again towards apractical science which not only included illustra-tive practicals, but which placed an emphasis on

those where pupils could 'find out' by discoveryand those which enabled students to practise 'sci-entific method'. Hence the influential Nuffieldschemes for secondary pupils (years 7-13) wereborn. It almost seemed as if Armstrong's heurismwas back. Again the intention was that pupilsshould be encouraged to discover science for them-selves. The focus was on scientific method andobjectivity with an underlying assumption that thepupil had no preconceptions (the inductiviststance), so that all observations were perceived asneutral. The essence of the Nuffield philosophywas 'to awaken the spirit of investigation and todevelop disciplined imaginative thinking' (NuffieldFoundation, 1966).

This explicit recognition of the importance ofcognitive processes is perhaps one of the most sig-nificant elements of the underlying philosophy ofthe Nuffield schemes. However, while the renewedemphasis on 'finding out' and scientific methodol-ogy was welcomed by educationists generally,once the Nuffield schemes were implemented. cer-tain fundamental problems emerged. The mostsignificant proble:m was that because the activitieswere relatively tightly controlled and the 'rightanswer' often apparent. there was little scope for'discovery' in the true sense of the word. Hence.Nuffield practicals were often contrived and thespirit of 'imaeinative thinking' was lost. The detailof the practical was not considered to be impor-tant because it was so carefully controlled andindeed much of the equipment was designed sothat little could go wrong in theory at least. Itwas only at the higher levels, such as in investiga-tional projects of the Nuffield A level physicscourse, that there was an element which allowedpupils a less structured format. Here, there wereinvestigations where the outcome in terms ofpupils' performance was less certain. However.for the majority, the reality was that Nuffieldpracticals were about illustrating or refining con-cepts rather than 'finding out', despite a clearlystated philosophy in the teachers' guides whichadvocated a very open approach.

The schemes were well supported financially,but this in itself had disadvantages because itmeant that some schools jumped on the Nuffield

13

CHANGING VIEWS OF PRACTICAL SCIENCE

bandwagon without really understanding itsethos. More importantly, because the 0 and Alevel science schemes received priority, theNuffield schemes were originally designed for themost able grammar school pupils. With theadvent of the comprehensive system, problemsarose. When the Nuffield schemes were used incomprehensive schools. the conceptual nature ofthe whole course was simply too much for themajority. Many of us, the authors included, whohad worked with the Nuffield schemes in a gram-mar school environment, slowly too slowlyprobably passed from unease to the realisationthat such a course was not, and could not be, forthe many. And since, in the comprehensive, thefew for whom the course was designed were nowvery few indeed, alternatives had to be found.There were eventually attempts to address thisproblem (for example, with the publication of theNuffield Secondary Science Teaching Project in1971), but they were not entirely successful.Alexander (1974), in evaluating this project,found that: 'the lessons showed a high degree ofstructure and relatively little work of a completelyopen ended nature...pupils' attitudes to scienceand to the relevance of science in society had notimproved'. This last point, that Nuffield sciencehad no obvious link to the world outside theschool laboratory. was also a general criticism ofall the Nuffield schemes.

In summary, the Nuffield schemes began with aclear intention of teaching scientific method, butin practice their main purpose was still to teachthe concepts of scientific knowledge. The practi-cals were based on the notion that 'seeing isbelieving'. It gradually became clear that these'guided-discovery' practicals were not achievingthe original aims of the Nuffield schemes. Qualteret al. (1990) write:

We have only to look at schemes such as NuffieldCombined Science to realise that a great deal oftime, energy and material resources are devoted to'guiding' the pupil to a predetermined 'discovery'.Critics of such schemes have questioned whethersuch a massive investment of resources is really jus-tified by such narrowly defined learning objectives.

19

The 'processes and skills' movement

In 1967, a scheme arose in America called 'ScienceA Process Approach' (SAPA; see American

Association for the Advancement of Science,1967). It stemmed from a study of what scientistsactually do in their everyday work. SAPA placed afirm emphasis on scientific method or 'process',regarding the 'substantive' concepts of sciencethat is, the facts. laws and principles as being farless important. Other schemes of the same ilk soonemerged and thus the 'processes and skills' move-ment was born.

It should be noted here that there has been, andcontinues to be, much confusion over the terms'processes', 'process skills', 'procedures' and'methods'. In science education. these terms areoften used interchangeably without clear defini-tion, causing considerable problems. In theprocesses and skills movement, 'processes' wereintended to refer not tothe practical skills associ-ated with science, but to the cognitive processessuch as observing, classifying and inferring thatis, the thoughts that go through scientists' mindsas they perform practical science activities. This isthe meaning we shall adopt here. It follows thatthe term 'process skills' which was widely used atthat time, is a contradiction in terms, confusingthe.cognitive process element with the practicalskill element. Some of the problems encounteredby the process movement can be seen to stem fromthis lack of clarity of definition as well as anuncertainty about the place of these processes inschool science generally.

In the UK, Screen (1986) suggested that theswing from content to process was aided by theadvent of' new national criteria for science syllabiand assessment in 1986. Prior to that time, boththe Nuffield schemes and the concern aboutmeeting examination requirements had led teach-ers to concentrate on facts and knowledge. Apolicy document in 1985 stated that: 'The essen-tial characteristic of education in science is that itintroduces pupils to the rnethods of science sothat scientific competence can bc developed tothe full' (DES, 1985). Screen, the author of aprocesses and skills scheme called Warwick


Process Science, further argued that: 'It could besaid that the most valuable elements of a scien-tific education are those that remain after thefacts have been forgotten' (Screen, 1986). Andagain: 'important though the content of sciencemight be, it is not the facts themselves but howthey are arrived at which constitutes an educa-tion in science' (Screen, 1988). There was now afeeling that factual scientific knowledge was soabundant (indeed Screen uses the term 'the explo-sion of knowledge'). that very few students couldpossibly acquire a thorough knowledge of anysubject area. Instead, students needed to knowhow `to access, use and ultimately add to theinformation store when required' (Screen 1986).

The focus of Warwick Process Science wastherefore on transferable 'process skills', whichincluded observing, classifying and interpreting.Screen points to two other advantages of thisprocess-led scheme. First, 'process skills' are notonly useful in practising science but also in quest-ioning the practice of other scientific work. Second.the scheme allows the teacher to observe pupils'conceptions and misconceptions in lessons where,for example, the focus is on predicting, hypothesis-ing or interpreting. The teacher can then use thisinformation to address misconceptions.

Similarly, a new Nuffield scheme emergedaround this time in which 'emphasis has beenplaced on processes...on process rather thanproducts' (Nuffield Foundation, 1987). There arealso a number of assessment schemes, such asThe Assessment of Practical Science (TAPS;Bryce et al., 1983) and the Graded Assessment inScience Project (GASP; Davis, 1989), based onthe assessment of 'process'.

In general, schemes that adopt this processapproach aim to consider, as far as is possible, oneprocess at a time as the focus of a lesson. The prac-tical work then serves to illustrate the process inquestion. Bryce et al. (1983) base their scheme onthe idea that: 'pupils are encouraged to masterbasic skills first and are thereby enabled toprogress to more complex process skills and even-tually practical investigations'. Woolnough (1991)contrasts this 'step-by-step' approach with thequestion: '...do they [the pupils] learn best by a

holistic, experiential approach whereby they areencouraged to do small, complete investigationsfrom the earliest stage, progressing to more diffi-cult investigations later and picking up appropriateskills when necessary?'

With the process approach, problems can arisein trying to incorporate these processes into acoherent scheme of work. Process schemes havealso been criticised for their content becausethe concepts they employ to focus on differentprocesses often lack continuity. There is indeedthe real possibility that learning processes in iso-lation may mean that some pupils will havedifficulty in putting them all together appropri-ately wht-n required to do so. Process-led sciencecannot, of course, be devoid of content, but itsemphasis is clearly on the processes practised inthe context of science: 'Science in Process...isbased on the processes of science on how towork scientifically rather than on a body ofcontent' (Wray et al., 1987).

Investigations

After the processes and skills movement came amove towards the holistic approach referred to byWoolnough (1991) in the form of scientific 'investi-gation'. This move was driven by the work of theAssessment of Performance Unit's (APU) work inscience and subsequently by the inclusion of inves-tigative work in the UK National Curriculum. Inthe next section, we shall provide an example of aninvestigation and in Chapter 3 we shall defineinvestigations in depth by considering differenttypes of investigations. Suffice it to say here thatinvestigations are a specific type of problem-solving which allow pupils a varying degree ofautonomy and which are problems to which thesolution is not obvious. Investigations aim to allowpupils to use and apply both concepts and cognitiveprocesses, as well as practical skills. This book setsout to show that this is a more balanced view of'science, which can be seen perhaps as a response tothe extremes of the past. Various schemes havearisen in recent years in the UK which havefocused on investigations (Foulds et al., 1990,

CHANGING VIEWS OF PRACTICAL SCIENCE

Table 1.1 Summary of types of practical work

L2.11

Type af practic.al

Skills To acquire a particular skill

Observation

Enquiry

Illustration

Investigation

To provide opportunities for pupils to use their conceptual framework in relating realobjects and events to scientific ideas

To discover or acquire a concept. law or principle

To 'prove' or verify a particular concept. law or principle

To provide opportunities for pupils to use concepts. cognitive processes and skills tosolve a problem

1993: Feasey et al. 1991) and their integration intothe practical science curriculum.

Practical science today

What kinds of practical work in science are usedin today's schools? If we adopt a bird's-eye view ofall the school laboratories in action in the UKtoday. we would find a large variety of types ofpractical science being carried out which stemfrom the history we have just described. Therehas been, and still is. a tendency to regard practi-cal work as one amorphous entity. By uncoveringwhat different types of practicals set out to teach.we shall see that this is not the case and that eachtype of practical serves a different purpose.

Various attempts have been made to classifydifferent kinds of practical work in order to definetheir respective roles. We shall consider here aclassification developed by Gott et al. (1988).which consists of five broad types of practicalwork (Table 1.1). The boundaries between thesetypes are not claimed to be watertight: practicalactiNities can clearly include more than one aspect.particularly skills and observation which areimplicit to some degree in the other types.

Enquiry practicals are structured to allow thepupils to discover a particular concept for them-selves. Such experiments have to be carefully setup to enable all pupils to arrive at the same end-point. Many of the Nuffield practicals followed

this format. The main aim of enquiry practicals isconcept acquisition.

Illustrative practicals differ in that they aim todemonstrate or provide a particular concept. lawor principle, which has already been introducedby the teacher, to allow the pupil to 'see the con-cept in action and so relate theory more closelyto reality. Illustrative practicals can take theform of a demonstration by the teacher or apractical where pupils are given detailed instruc-tions or a 'recipe' to follow. The main aim ofillustration is concept consolidation. The spring'sactivity in Fig. 1.1 is a typical practical where thepupils either 'discover' Hooke's Law in anenquiry practical. or if it follows directed teach-ing on the topic. the same activity can be used asan illustrative practical enabling pupils to see theconcept in action.

Skills practicals may involve setting up. readingand using instruments as in the example takenfrom Gott et al. (1988) and reproduced as Fig. 1.2.or they might require pupils to learn and practisethe construction of line graphs or bar charts. Theyare about acquiring the basic skills necessary forcarrying out the rest of practical science. Onceacquired. the skill of constructing a graph from adata set comes as second nature. The collection ofthe data and the interpretation of the graph are ofa different order and arc not within our definitionof skills.

Observation in science has been described as'theory-laden' in that when pupils are asked to


Springs and Hooke's Law

1 Collect a retort, clamp, ruler, spiral spring and a set of 1N slotted weights.2 Measure the length of the spring between points A and B as shown in the diagram.3 Fasten the loop of the spring onto the arm of the retort stand as shown in the diagram.

CM33333333333._)

4 Add the 1N hanger to the bottom of the spring.5 Calculate the extension of the spring caused by

the 1N weight. (This is the amount by which thespring has stretched.)

6 Draw this table in your notebook.

Weight on spring (g) Extension (cm)

10

20

30

etc.

20

c0.(2 to

Weight (g)

7 Put your results into the table.8 Find the extension of the spring for weights of 2N, 3N, 4N and 5N. Add your results to the table each time.9 Use your results to draw a graph using these axes.

10 What conclusions can you draw from this experiment?

Fig. 1.1 An example of an enquiry or illustration practical

observe in science they are expected to apply scien-tific conceptual knowledge to the object or event inquestion. Hence in the example shown in Fig. 1.3.again from Gott ei al (1988), the teacher is askingpupils to apply ideas such as conduction and con-densation to the observed phenomena.

Investigations usually offer several alternativeways of reaching a solution to the problem so thatthe design is much less controlled than in illustra-tive or enquiry work. Investigations use conceptswhich have been introduced by some other means;in an illustrative practical or exposition perhaps.Their main aim is to allow pupils to use concepts.

cognitive processes and skills to solve a problem.An example from Investigations in Science (Fouldset al., 1990) of an investigation called 'On the tiles'sets the scene as follows:

2

It was chaos in the kitchen! The baby had droppedhis bottle and milk was dripping out. Little Rachelknocked her plate off the table the bread landedjam-side down. In from the garden came Dadwith muddy boots on. He was followed by Sparkythe dog, who had just been digging for bones.The floor was really messy. How easy will it be toclean up'?

CHANGING VIEWS OF PRAC FICAL SCIENCE

Slide making

1 Use the pipette to place a medium-sized drop of thepond water from beaker A onto the centre of a cleanslide.

2 Cover with a coverslip taking care to exclude airbubbles.

3 Use a strip of blotting paper to remove excess waterfrom the slide.

4 Place the finished slide onto the microscope stage sothat the water drop is in the centre of the field of view.

5 Focus the microscope at low power.

6 Put up your hand for your teacher to check your slide.

7 Write down what you see in the pond water when youuse the microscope.

Fig. 1.2 An example of a skills practical

Icecan

In front of you is a tin can with a mixture of ice and waterin it.

(a) Look carefully at the outside of the can. What doyou notice?

(b) Say why you think those things happened

Fig. 1.3 An example a an observation practical

The pupils are then asked: 'to find out which is thebest material for covering the kitchen floor', butbefore they begin each group is asked to write onesentence explaining what 'best' means. Foulds el al.(1990) suggest that the different interpretations of'best' could then be discussed by the class to decidewhat course the investigation follows. Some guid-ance is given on structuring the recording:

23

Your results might be easier to read in a table.Before you draw the columns and headings, thinkabout what goes in them think about how manytypes of covering you will use, what you will puton it. what name you will give to your measure-ments. etc.

( Foulds ct (il., 1990)

Summary

In this chapter, we have traced the historicaldevelopment of practical science and seen that ithas offered several different views of its role andpurpose. We have seen that, over the last 150ycars, there have been quite radical swings. Theearliest practical work focused on concepts, whileArmstrong's heurism swune the pendulum firmlytowards the processes of science. The Nuffieldschemes were intended to follow this trend but inpractice focused on concept acquisition, while theprocesses and skills movement swung the pendu-lum back again. The move towards investigationsseems to lie somewhere in the middle.

If we have anything to learn from history in thiscontext, it must surely be that we should avoidthese radical swings that have occurred in the past.In the not too distant past. it often seemed as ifthere were fashions in practical science when onescheme came in, the previous one was thrown out.There has also been a competitive elementbetween schools, suggesting that one scheme was'better' at teaching the practical component of sci-ence than another.

A further related problem is that there has beena tendency for teachers to accept that doing prac-tical work regularly is a 'good thing', with littlethought as to its purpose or learning outcomes. Itis as if all practical work is one amorphous entityserving a kind of 'hands-on' purpose in the cur-riculum. Hodson (1992) suggests that a lot of whatgoes under the name of practical science is 'mud-dled and without real educational value'. citesevidence (Hodson, 1990) to show that the reasonsdifferent teachers give for doing practical work arediverse. Hodson argues that the teacher must beprecise about the required learning outcome of thelesson and then decide whether practical work is

9 0


the best way of achieving that goal. This leads himto question the whole purpose of practical work.We have used one classification here to unravelthe purposes of different kinds of practical work.Hodson suggests that computer-assisted learning(CAL) can be used to learn concepts and theoriesand to learn about the nature of science. but whenit comes to 'doing science', he writes:

A major goal of practical work should be theengagement of students in holistic investigationsin which they use the processes of science both toexplore and develop their conceptual understand-ing and to acquire a deeper understanding of (andincreased expertise in) scientific practice.

(Hodson. 1992)

To conclude, we have painted a picture of thecurrent state of practical work, which suggeststhat there is a lack of clear thinking behind its pre-sent rather disjointed and inconsistent structure.Perhaps we need to stand back and try to developa framework to enable us to locate and pull allthese ideas about the purpose and nature of prac-tical work together. That framework should beginthe task of deciding what different sorts of practi-cal work are for and when, how and why theyshould be deployed, rather than either letting themco-exist in a muddle or, worse, assuming that thelatest type is a competitor which is to supplant orbe defeated by other types. We shall then be ableto locate investigations within that framework sothat their role in the science curriculum becomesclear and so that we can then move on to exploreinvestigative work itself.

References

Alexander. D.J. (1974). Nuffield Secondary Science: AnEvaluation. London. Macmillan.

American Association for the Advancement of Science(1967). Science' .4 Process Approach. Washington,DC, Ginn & Co.

Armstrong, H.E. (1896). How science must be studiedto be useful. From 'The Technical Wot ld', in II E.Armstrong and Science Education (1973: G. VanPraagh. ed.). London. John Murray. .

Bryce, T.G.K., McCall, J.. MacGregor, J.. Robertson.I.J. and Weston. R.A.J. (1983). Techniques _for the.4ssessment of Practical Skills in Foundation Science.London. Heinemann.

Davis. B.C. (1989). GASP: Graded Assessment inScience Project. London, Hutchinson.

Department of Education and Science (1985). Science5 /6: .4 Statenumt of Policy. London, HMSO.

Feasey. R.. Foulds, K., Gott, R. and Pryke. T. (1991).Science in Action 5 to 16: Key Stage 3, Book 1.Glasgow. Nelson Blackie.

Foulds. K.. Mashiter, J. and Gott. R. (1990). Investigationsin Science. Glasgow, Blackie.

Foulds, K.. Gott. R.. Pryke, T. and Borrows. P. (1993).Science in Action 5 to /6: Key Stage 3, Book 2.Glasgow. Nelson Blackic.

Gee, B. and Clackson. S.G. (1992). The origin of practi-cal work in the English school science curriculum.School Science Review. 73(265): 79-83.

Gott. R.. Welford, G and Foulds. K. (1988). TheAssessment of Practical (Fork in Science. Oxford,Blackwell.

Hodson. D. (1990). A critical look at practical work inschool science. School Science Review, 70(256): 33-40.

Hodson, D. (1992). Redefining and reorienting practicalwork in school science. School Science Review,73(264) 65 78.

Kerr. J.F. (1964). Practical Work in School Science.Leicester, Leicester University Press.

Nuffield Foundation (1966). Nuffield Chemistry:Introduction and Guide. Harlow. Longman/Penguin.

Nuffield Foundation (1987). Nuffield 11 to 13 Science.Teacher's Guide 2. How Science is Used. Harlow,Longman.

Qualter. A.. Strang. J., Swatton. P. and Taylor. R.(1990). Exploration: A Way of Learning Science.Oxford, Blackwell.

Screen. P.A. (1986). The Warwick Process science pro-ject. School Science Review, 72(260): 17-24.

Screen. P.A. (1988). A case for a process approach: TheWarwick experience. Physics Education, 23: 146-9.

Wellington, J. (ed.) (1989). Skills and Processes inScience Education. London, Routledge.

Woolnough, B.E. (1991). Setting the scene. In: PracticalScience (B.E. Woolnough, ed.). Buckingham. OpenUniversity Press.

Wray. J.. Freeman, J.. Campbell, I... Hoyle. P..Nimenko. G., Smyth, S. and Whiston. L. (1987).

ir oce in Process. Teachers' Resource Pack. London,Heinemann Educational Books.

2 4

CHAPTER 2

Alternative perspectives

In this chapter, we shall attempt to gather theideas behind the various types of practical workdescribed in Chapter 1 into a coherent frameworkwhieh allows them to be seen as complementaryrather than competing views of science. Such aframework, incorporating ideas about the natureand purpose of practical work, must relate toschool science as a whole. Practical work shouldbe an integral part of the science curriculumwhich mirrors, reinforces and augments the restof the course.

In the first part of this chapter. we present andthen develop a model and a taxonomy (or classi-fication) which are applicable to the whole of thescience curriculum as well as to prauical science.They are based on an epistemological perspec-tive, which seeks to define that which is to betaught and learned, rather than how that is tooccur. This perspective underlies much of the fol-lowing chapters. In the second part of thechapter, we consider the movement towardsteaching thinking skills, which regards educationfrom a psychological perspective. We will look inparticular at the Cognitive Acceleration throughScience Education (CASE) project, which teachesthinking skills through the science curriculum.

A 'content' perspective (an epistemological view)

A model of science

A model based on that advanced by Gott andMashiter (1991) will serve as a starting point in our

search for a coherent framework (Fig. 2.1). Itshould be noted that this model was a basic, anddeliberately simplified, model for science and notsolely for practical science. Here the cognitiveprocesses (which we defined in the last chapter)needed to solve all kinds of problems are seen asinvolving an interaction of 'conceptual' and 'proce-dural' understanding. It should be noted that themodel does not imply that these two types of under-standing are mutually exclusive. Furthermore.other factors such as motivation, context or thepupil's perceived expectation of what should be

Solveproblems

Cognitiveprocesses

Conceptue,understanding

Facts

Proceduralunders anding

Skills

Fig. 2.1 A model for science (based on Gott andMashiter, 1991)


done in a science lesson, each of which can have asignificant effect on performance, are omitted.

For the implications of the model to be clear,we need first to define the terms used within it.

Solving problems

We indicated at the beginning of ihis section thatthe model is not confined exclusively to practicalscience. To retain this position, we need to adopt avery catholic definition of:problem' to includeany activity that requires a pupil to apply his orher understanding in a new situation. This willinclude explanation of phenomena, applied sci-ence problems. theoretical problems as well aswhat we will define as investigations. What differsbetween these is the relative emphasis on concep-tual and procedural understanding.

This book, however, focuses primarily on inves-tigative work which is one type of problem-solvingand which is usually, but not exclusively, practicalin nature.

Conceptual understanding and Idcts

Williams and Haladyna (1982) define facts as'associations between names, other symbols.objects and locations'. and concepts as 'classes ofobjects or events that are grouped together byvirtue of sharing common defining attributes'. Inthe model, conceptual understanding refers to theunderstanding of the ideas in science which arcbased on facts. laws and principles and which aresometimes referred to as 'substantive' or 'declara-tive' concepts. We shall refer to these concepts assubstantive concepts in the chapters that follow.Examples include energy, the laws of motion.heredity. solubility. photosynthesis and so on.

Procedural mulct-standing and skills

Skills here refer to actk ities such as the use ofmeasuring instruments and the construction oftables and graphs. which are necessary but not

sufficient in themselves to the carrying out of mostpractical work. We will restrict the term to thosemechanical aspects of these activities which wehope pupils will acquire and store in the back-ground, so to speak, ready to be extracted as atool, complete with in-built instructions for use.

Procedural understanding is the understancEngof a set of ideas which is complementary to con-ceptual understanding but related to the 'knowinghow' of science and concerned with the under-standing needed to put science into practice. It isthe thinking behind the doing. For example, inmeasurement in a plant growth study, proceduralunderstanding does not refer to the measuringitself, but to the decisions that have to be madeabout what to measure. how often and over whattime period. It also includes the understanding ofthe notion of the fair test as well as the nature of aline graph, how it differs from a bar chart or howit illustrates patterns between variables.

The content of procedural understanding is notwell documented. Too often it is regarded solely asa means of acquiring a concept. Althouah proced-ural understanding can be a means of learning orlearning about a concept, it is also a kind ofunderstanding in its men right. The significance ofprocedural understanding underlies much of theargument in this book.

The term 'procedural knowledge' (or proceduralunderstanding) is used in both maths and Englishbut in a somewhat different sense. In maths, forinstance, it relates to the use of problem-solvingstrategies: in English. to the construction of prose.In both these subjects, the procedural knowledgerequired is basically the recall and use of a set ofrules (in maths. formulae or theories). In science.there is the additional problem of not only know-ing the 'rules' (of the fair test for instance), but ofrelating these rules and the concepts of science toobjective evidence. Similarly in history. reasons foror explanations of change need to be supported bysome kind of evidence from primary or secondarysources. It is this collection and verification of datawhich can be seen as one distinguishing factorbetween procedural knowledge in science and his-tory. and in other subjects.

ALTERNATIVE PERSPECTIVES

Cognitive processes

Conceptual and prr edural understanding cannotbe totally independent of one another: some under-standing of substantive concepts is necessary tocarry out most procedural aspects of science, andsimilarly procedural understanding is necessary toput substantive concepts into practice. The cogni-tive processes in Fig. 2.1 refer to this interactioninvolving the selection and application of facts,skills, conceptual and procedural understanding.These cognitive processes are the means of obtain-ing or processing the information needed to tacklea problem successfully. Although procedural andconceptual understanding are intertwined in thisway, the distinction between them is a useful one inunravelling the complexities of school science.

The emphasis of a particular task may be moreon one side of the model than the other. Hencewhen the conceptual understanding required tosolve a problem is very hard and the proceduralunderstanding easy, then the balance is to the leftand vice versa.

An example will serve to illustrate how themodel we have put forward can be applied.Suppose that the problem is to find out how theaverage or final speed of a toy car travelling downa ramp is related to its weight. The pupil firstneeds to understand the concept of speed andknow that it involves distance and time (concep-tual understanding). He or she will need to havethe skills to be able to measur e distance, time andweight. Then the pupil must decide how to con-struct a fair test and what distance and time tomeasure (procedural understanding). All this infor-mation has to be processed in designing theinvestigation, examining the resulting data anddrawing appropriate inferences, provided that thepupil considers that the data are 'believable' (pro-cedural io/derstanding).

An analogy may be useful here. The facts, skillsand understandings can be envisaged as informa-tion or patterns in the brain's memory bank. Whenfaced with a problem of any sort, but in the senseof some new experience which requires resolution,the brain can bc imagined to scan its do fa banks

27

for facts or previous experiences that may helpwith the new problem. In the above example, those'hard disk stores' will contain ideas about speed,measurement of distance and time, skill routinesabout using instruments, notions of a fair test andhow it relates to the validity of any resulting dataand so on. The central processing unit will thenexamine the problem and look on the hard disk forhelp; this may be in the form of particular ideas, orpast experiences in similar 'circumstances. Thesewill be pulled into the working memory. Then theymust be 'processed', via a series of thought pat-terns that we label hypothesising, or predicting orwhatever, into a solution consonant with, andevaluated against, the demands of the originalproblem. The 'processes' discussed in the previouschapter can now, by and large, be identified withthese patterns of thought. Hypothesising, or pre-dicting, or interpreting, or explaining are diffewtoperations of this central processing unit; differentguides to the scanning of the memory banks ofconceptual and procedural understanding.

We will argue that procedural understandinghas to be taught in order that there is something inthe relevant data stores for the central processorto access and structure.

Classifving practical ivork

We can use this model as a way of locating thetrends and themes in science which we described inthe last chapter. Armstrong's heurism, for instance,would lie firmly on the right-hand side of themodel, while the reality of the Nuffield schemesemphasised the left-hand side. Duggan and Gott(1993) provide a useful summary which relates thefive broad types of practical work put forward inthe last chapter to their principal learning outcomesin terms of conceptual and procedural understand-ing (Table 2.1).

Table 2.1 demonstrates that each type of practi-cal work has a significant role to play in practicalscience. While most types of practical work involvesome elements of procedural and conceptualunderstanding, the table refers only to the principallearning outcome of each type. Hence, it could be

2


Table 2.1 Classification of types of practical work by their learning outcome (Duggan and Gott, 1994)

Type ofpractical

Principal learning outcome

Conceptual understanding Procedural understanding

Skills

Observation

Enquiry

Illustration

Investigation

application

acquisition

consolidation

application

(skill) acquisition

application and synthesis

argued that enquiry practicals could embody bothconceptual and procedural understanding. That iscertainly true in that there will be skills or datainterpretation involved, but it is not the main pur-pose of the exercise. It could also be argued that ifthe aim is to acquire a concept, then asking pupilsto deploy skills and procedural understanding asit'd/ may be inefficient. Experience of Nuffield andin the classroom suggests that such activities over-load most pupils to the extent that none of thelearning outcomes may be achieved.

Table 2.1 also shows the place of investigationsin practical science in that they are the only typeof practical work whose principal learning out-come is to provide pupils with the opportunity toachieve a thorough grasp of procedural under-standing, while at the same time allowing pupils touse and refine their conceptual understanding. Weshall return to this point again later.

Developing the model

In the last section, we used the model in Fig. 2.1 todescribe two kinds of understanding that underpinscience conceptual and procedural understand-ing. In this section, we shall try to identify ingreater detail the types and levels of conceptualand, particularly, procedural understanding thatchildren need in science. In order to do this, wehave used a taxonomy based on work by Bloom eta/. (1956), which was originally intended for use inthe American Grade System.

2

Bloom's taxonomy as a descriptor of science

Bloom's taxonomy of 'educational objectives forthe cognitive domain' has been used, reused andmodified to suit a wide variety of purposes. It wasoriginally intended for assessment purposes toenable teachers to relate educational objectives interms of subject content to the thinking processesinvolved. Williams and Haladyna (1982) writc:'The taxonomy is undoubtedly one of the mostimportant contributions to educational practice inrecent times. It was enthusiastically received andhas been widely used ever since its publication.'

Kempa (1986) suggests that Nuffield 0 levelscience was based largely on the work of' Bloom etal. He describes how Bloom's original taxonomycontains six dicferent levels of cognitive abilitywhich are hierarchical, in the sense that higherlevels subsume lower levels. He continues: 'but formost science examinations three of these tend tobc combined, giving a four level classification'.Kempa defines these four levels as:

Knowledge and recall of scientific facts, hypothe-ses, theories and concepts, as well as terminologyand convention.Understanding of scientific knowledge and rela-tionships, which manifests itself in the student'sability to explain and interpret informationpresented, and to express it in alternative com-munication modes.Application of scientific knowledge and under-standing to unfamiliar (i.e. novel) situations.


The ability to apply knowledge implies thatthe student is able to select from his or herknowledge reservoir those items of knowledgeand relationships that ate relevant to the novelsituation.Analysissynthesis and evaluation of scientificinformation, which involves the breaking down ofinformation into its constituent parts (analysis)and reorganising it so that a new structureemerges (synthesis). Additionally, the informationmay have to be evaluated in terms of its validityor underlying assumptions, and cons( quences.

The difference therefore between understandingand application is that understanding is manifestedwithin one context, whereas application is the abil-ity to apply that understanding to other contexts.

The hierarchical structure of these taxonomiesderived from Bloom has caused much distrustbecause it is difficult to relate to real situations.For instance, it is often difficult to distinguishbetween recall and understanding. There are alsoinstances where evidence appears not to supportthe taxonomy. For example, it is relatively easyto create a practical situation in which a pupilwill apply an idea that he or she is quite unableto recall, or see the relevance of, when faced withan apparently simple question. The 'simple' ques-tion may introduce a different context or stylewhich causes the pupil difficulty. Nevertheless,the taxonomy can be used as a way of describingwhat science might contain, without getting intothe complexities of whether or not it is possibleto use the hierarchical structure to predict levelsof difficulty.

We can apply this taxonomy to both types ofunderstanding which we have defined above andused in the model in Fig. 2.1. To begin with, weshall describe briefly its application to the familiarconcepts of science so that we can then demon-strate that a parallel structure for proceduralunderstanding can be defined. We shall look atthis less familiar structure in greater detail.

Bloom's taxotunny and conceptual understanding

Wc have simplified the classification above andapplied it specifically to conceptual understanding:

29

Conceptual taxonomy

Knowledge and recall of facts

Understanding of concepts

Application of concepts (in unfamiliar situations)

Synthesis of concepts (in problem-solving).

As we noted above, we intend to use this taxon-omy as a description, with no implications that itis a hierarchical structure. As a description of thescience curriculum one of many possible onesof course it suggests that there will be a placefor the simple recall of some parsimonious selec-tion of facts. There must also be room for anunderstanding of the ideas, how the facts inter-relate in the context in which they were taught, aswell as the application of those ideas in novelsituations. And, finally, in the context of solvingproblems, pupils must synthesise knowledge andconceptual understanding.

One of the weaknesses of the model here is thatconcepts can Dme more like facts with increas-ing age and experience. For instance, distance canbe a concept in one situation to a young child butwill be recalled and used much like a fact inanother situation by an older pupil. So, as with allmodels, if it is pursued too far, it is in danger ofcollapse. That does not mean that it cannot helpus as a descriptive tool in making some overallsense of the complexities inherent in the cognitiveabilities required in science.

Bloom's taxonomy and procedural understanding

We can also use the taxonomy to explore thenotion that the 'content' of the procedural aspectsof science merits a similar differentiated treat-ment (Fig. 2.2).

We have based the procedural taxonomy onskills and 'concepts of evidence' which arc comple-mentary to the facts and substantive conceptsoccurring in the conceptual taxonomy of the pre-vious section. Before we look at the structure ofthe taxonomy, however, it is important to clarifywhat wc mean by 'concepts of evidence'.

30 INVESTIGATIVE WORK IN l'HE SCIENCE CURRICULUM

Solveprob ems

Cognitiveprocesses

I domeptualundenttanding

Datinst tcar t:pt.ta'il Factitaxonomy

1

Fig. 2.2 Developing the model

Concepts of evidence

Wc have coined the phrase 'concepts of evidence'.which we have used elsewhere (Duggan and Gott.1994) to refer to the concepts which are associatedwith procedural understanding. In passing, weadmit to having some doubt about this phrase. Wehave considered alternatives such as 'procedures','working methods' or 'procedural concepts', butnone of these terms conveys the desired meaning:procedures or working methods tend to imply alow or algorithmic level of cognitive skill; proce-dural concepts again restricts the meaning toindividual procedures and is somewhat confusing.Our term, concepts of evidence, draws attention tothe importance of this understanding and the con-cepts underlying the doing of science in relation tothc evidence as a whole.

Figure 2.3 shows how we have structured theseconcepts of evidence around the four main stagesof investigative work: namely, those conceptsassociated with the design of the task, measure-ment. data handling and, finally but crucially. theevaluation of the complete task in terms of thereliability and validity of the ensuing evidence. Bystages, we do not mean stages in time, since thesestages are often revisited. For instance, at the data

I: ProosdwalUndt,orstahling

similarta*Onotry bedevelaPod

here

handling stage, a decision may be made to takemore measurements. The evaluation of the taskrequires an understanding of all three stagesdesign, measurement and data handling and thisunderstanding of evaluation is needed as much atthe beginning as at the end of the task. The kindof understanding associated with each of thesemain stages is defined in detail in Table 2.2.

Proceduralunderstanding

A

Evaluation

DesignMeasurementData handling

Skills

Fig. 2.3 Procedural understanding and concepts ofevidence


Table 2.2 Concepts of evidence and their definition

31


Associated withdesign

Definition

Variableidentification

Fair test

Sample size

Variable types

Understanding the idea of a variable and identifying the relevant variable tochange (the independent variable) and to measure, or assess if qualitative (thedependent variable)

Understanding the structure of the fair test in terms of controlling the necessaryvariables and its importance in relation to the validity of any resulting evidence

Understanding the significance of an appropriate sample size to allow, forinstance, for probability or biological variation

Understanding the distinction between categoric, discrete, continuous andderived variables and how they link to different graph types

Associated withmeasurement

Relative scale

Range andinterval

Choice ofinstrument

Repeatability

Accuracy

Understanding the need to choose sensible values for quantities so thatresulting measurments will be meaningful. For instance, a large quantity ofchemical in a small quantity of water causing saturation, will lead to difficultyin differentiating the dissolving times of different chemicals

Understanding the need to select a sensible range of values of the variableswithin the task so that thc resulting line graph consists of values which arespread sufficiently widely and reasonably spaced out so that the 'whole'pattern can be seen. A suitable number of readings is therefore also subsumedin this concept

Understanding the relationship between the choice of instrument and the re-quired scale, range of readings required, and their interval (spread) and accuracy

Understanding that the inherent variability in any physical measurement re-quires a consideration of the need for repeats, if necessary, to give reliable data

Understanding the appropriate degree of accuracy that is required to providereliable data which will allow a meaningful interpretation

Associated with Tablesdata handling

Graph type

Patterns

ulti% a ria tedata

Understanding that tables are more than ways of presenting data after theyhave been collect-.i. They can be used as ways of organising the design andsubsequent :.ata collection and analysis in advance of the whole experiment.

Understanding that there is a close link between graphical representationsand the type of variable they are to represent. For example. a categoricindependent variable such as type of surface, cannot be displayed sensiblyin a line graph. The behaviour of' a continuous variable, on the other hand, isbest shown in a line graph

Understanding that patterns represent the behaviour of variables and thatthey can be seen in tables and graphs

Understanding the nature of multivariate data and how particular variableswithin those data can be held constant to discover the effect of one variableon another

Associated with Reliabilitythe evaluation ofthe complete task Validity

Understanding the implications of the measurement strategy for the reliabilityof the resulting data: can the data he believed?

!nderstanding the implications of the design for the validity of the resultingdata: an overall view of the task to check that it can answer the question


Car Distance I Distance 2 Distance 3 Average

Blue

Red

Green

We have used the term 'variable' in Table 2.2to refer to any observation which can bedescribed by different values for example, tem-perature. length or time. Variables can beclassified in terms of their roles and functions inthe structure of the activity as Independent','dependent' 'control' variables.

The values for the 'independent' variable arechosen and manipulated by the investigator. Thevalue of the 'dependent' variable is then measuredfor each change in value of the independent vari-able. 'Control' variables are those which must bekept constant while the independent variable ischanged to make the test 'fair'. Many scientifictasks can be defined in this way according to their'variable structure'. For example. supposing thetask is to find the effect of car colour on frequencyof accidents, the car colour is the independent vari-able, the frequency of accidents is the dependentvariable and the age of the driver is one of severalcontrol variables. These terms are used in theNational Curriculum of the UK. We should notehere that we are conscious of the limitations ofdefining procedural understanding in terms of thevariable structure of a task, a point to which weshall return.

It is important to note that concepts of meas-urement are to do with the decisions that have tobe made about measurement rather than to dowith the skill of measurement itself. For instance,in a task about the effect of temperature on thedissolving time of sugar, it is not the ability of thepupil to use a thermometer (which we define as askill), but rather the decisions that he or she has

3,2

taken about, for example, the range and intervalof temperatures and the number of repeats whichreflect the understanding that the pupil has aboutthese particular concepts of measurement.

Concepts associated with data handling includethe understanding of the use of a table as a way oforganising data rather than the construction oftables themselves. Hence before even beginningmeasurement, a pupil may construct a table ofvalues of the variable which he or she is going tochange. For instance, in the dissolving example towhich we have already referred, the pupil mayconstruct a table with temperatures of 25, 50, 75and 100°C before beginning the task. Or for thetask of finding out the distance travelled by dif-ferent coloured toy cars, the pupil may constructa table such as the one above.

A further aspect of data handling is the isola-tion of the required variable from multivariatedata. For example, in the multivariate table below,to consider the effect of temperature it is necessaryto compare the times in the left-hand column withthose in the right-hand column. Alternatively, toconsider the relative dissolving times of caster andbrown sugar, the times in the top row have to becompared with those in the bottom row.

Hot Cold

Caster sugar (sec) (sec)

Brown sugar (sec) (sec)


The final evaluation stage subsumes all theother concepts of evidence because reliability andvalidity can only be considered in the context ofthe strategy of the whole task.

There are other ideas that could, and perhapsshould, be included in this list. For example, thenotion of ratio and proportionality is much usedin science. On what grounds are they excludedwhen the concept of 'patterns' is included? Theobvious but not wholly convincing answer, is thatpatterns, which are in the list, must include ratioand proportionality. The other argument is thatproportionality and ratio are mathematical ideaswithin the basic (or axiomatic) construction ofmathematics. We have attempted to restrict ourdefinition of concepts of evidence to ideas thatrelate data to reality, which is a crucial distinctionbetween mathematics (which only needs to be self-consistent) and science (which has to satisfy therequirements of the behaviour of objects in thereal world).

The significance of eridetwe wul the notion ofaudience

We shall see later, it' we take the evaluation of atask as an indicator, that the notion of data as evi-dence would appear to be understood by very fewpupils. The model we have developed is based onthe assumption that evidence is an importantnotion in science education. It might be usefulhere to stand back for a moment and ask why webelieve evidence to be important?

If we adopt the view that the aim of science issimply to arrive at a set of concepts which canexplain real-life behaviour, we can then test thistheory against some actual examples. If we con-sider universal gravitation. for instance, then thesubstantive concept of gravity can explain this phe-nomenon successfully. Evidence here takes asecondary role in the real world of school sciencein that it serves only to validate what is nowaccepted theory, although at the time the theorywas being established, evidence was crucial for itsvalidation. If we consider the phenomenon of elas-

33

ticity, then we can use Hooke's Law to explainstretching reasonably well, although the relation-ship varies with different materials and the law isonly applicable within certain limits. In this ex-ample, sound evidence is crucial because data arerequired to make the law usable a spring con-stant needs to be calculated for specific instances.Finally, if we consider the rate of water flow in ariver, this is extremely difficult to explain withoutempirical evidence (perhaps aided by computermodelling). Here, the data are the basis for themodel and are paramount. This is not to say, ofcourse, that substantive concepts do not guide themodel: for instance, the concept of friction willsuggest that the flow may be less at particular loca-tions. In these three examples, evidence and thedata which supports it are important but they takeon different roles from validating theory in the firstexample to being the key component in the last.Similarly, the conceptual element is crucial in thefirst example but of less importance in the last.

In all three examples above, the evidence isneeded to justify the outcome. In the first ex-ample. however, the theory of universalgravitation is so well accepted that evidence whichsupports the theory is unlikely to be closely scruti-nised. particularly in the classroom. It could ofcourse be argued that this should not be the case:theoretically, evidence should always be impor-tant. In reality, however, if the lesson is aboutillustrating Boyle's Law, then as long as the evi-dence broadly supports the theory, it is likely tobe accepted without question. The chances ofrefuting a long-established substantive concept inthe school laboratory are indeed low. However, inthe third example the evidence determines themodel, so it is open to interpretation and indeedto misinterpretation. Because the 'solution' is notobvious, the evidence is much more crucial. If themodel is to be believed, it must be supported bysound evidence. In science, it is often necessary tomake that evidence public or available for othersto see. so that they can evaluate for themselveshow much weight to give to the explanation,interpretation or solution.

If we can gct across this sense of the public

INVESTIGA1 IVE WORK IN THE SCIENCE CURRICULUM

nature of evidence or, to put it another way, theidea of evidence for an audience, then pupils aremore likely to understand the notion of data asevidence. We shall return to the notion of audiencein Chapter 6 when we consider strategies for intro-ducing these ideas in the classroom or laboratory.

We have introduced these ideas here because thesignificance of evidence (and of concepts of evi-dence) in science underlies our views of proceduralunderstanding. to which we shall now return.

procedural taxmlomy

We have developed a taxonomy for proceduralunderstanding which is shown below using skillsand concepts of evidence:

Procedural taxonomy

a Knowledge and recall or skills

Understanding of concepts of evidence

Application of concepts of evidence (in unfamiliarsituations)

Synthesis of skills and concepts of evidence (inproblem-solving)

Some examples may help to show how the tax-onomy might be applied. As with the conceptualversion discussed earlier, we are suggestinu herethat a curriculum should contain something of theknowledge and recall of skills, such as the use of athermometer. It should also encompass the under-standing of concepts of evidence, as in theunderstanding of the role of the fair test within afamiliar context, or the range and number of read-ings required in measurements of temperature. Theconcepts must also be applied to novel situations(transferred). The ability to apply the notion of thefair test should be available in a whole range of cir-cumstances: in all experimental work of whatevertype, or in criticism of other people's experimentalaccounts. Finally, we have the ability to synthesiseskills and concepts of evidence into the solution toa problem where, for instance, the links between a

3

fair test and the validity of any resulting data, orbetween the accuracy of a set of readings and thereliability of the data, are taken into account ingenerating 'believable' data.

We can see, therefore, that the descriptivemodel which has been so influential in structuringand assessing the conceptual component of the sci-ence curriculum can also be applied in a similarway to the procedural component. The import-ance of this is not in some arbitrary mirroring ofan existing structure or imposing a needless levelof complexity on an already complex enough situ-ation. Rather, it is to do with recognising thatthere is a content to the procedural side ofwhich can be described and which must be rcL,.,.-nised and planned for in curriculum design andassessment.

If we reconsider the principal learning outcomeof the main categories of practical work which weput forward earlier in the light of these two tax-onomies, we can see that observation practicalsare mainly about the application and synthesis ofconceptual understanding. Illustrative practicalsusually concern the understanding of substantiveconcepts. Investigations, however, provide theopportunity for pupils to synthesise conceptualand procedural understanding.

An alternative perspective of the relationshipbetween concepts of evidence and the science cur-riculum that deserves mention is the psychologicalperspective, which is exemplified by the thinkingskills movement.

The thinking skills approach and CASE

There is a growing movement towards the teachingof 'higher-order' thinking skills in the UK (Young,1993) and a recognition that such teaching has avaluable contribution to make to the curriculum(see, for example, Coles and Robinson, 1989).'Higher order here' refers to the thinking which isnot tied to specific subjects and includes, for ex-ample, the ability to sort out common features orpatterns in a series of pictures or texts or to gener-alise from them. It also includes the ability toevaluate conflicting evidence. There are several

ALTERNATIVE PERSPECTI ES

courses, all of which are based on the notion of flex-ible and articulate thinking. Children are encouragedto think about their own thinking (ntetacognition)and to reflect and share their learning experiences.The teacher is seen as a mediator and facilitator.

These courses stem from a psychological per-spective and are based largely on the work of Piagetand Vygotsky. Many are not specifically science-based. The Somerset Thinking Skills course (Blagget al.. 1988), for instance, is usually set withinEnglish or Personal and Social Education (PSE)lessons. It uses visual activities and focuses on cat-egories of thinking which include recognisingpatterns, dealing in probabilities, drawing analo-gies. evaluation, analysis and synthesis. Anothercourse. Feuerstein's Instrumental Enrichmentprogramme. is based on Vygotsky's theory thatevery human being has the potential to become aneffective learner. The teacher by mediation can pro-mote the child's cognitive development.

The Cognitive Acceleration through ScienceEducation (CASE: Adey et al.. 1989) interventionis, however, set within science lessons, and like theSomerset course, emphasises the importance oftransfer and of metacognition. The CASE projectis based on the argument that the science curricu-lum makes high cognitive demands on averagesecondar school pupils which are not adequatelydealt with in 'normal' teaching. This 'mismatch' isaddressed by an intervention which is aimed atimproving or accelerating the child's reasoningprocesses. These same reasoning processes are par-ticularly relevant in practical science. The theorybehind the method stems from Piagetian psycho-logy, which is outlined briefly below.

Piaget's developmental model

In The Growth of Logical Thinking .front Childhoodto Adolescence. Inhelder and Piaget (1958) reportthe findings of detailed studies of the growth ofmathematical and scientific concepts in children.This book was a landmark in the history of' psycho-logical investigation into thinking and reasoningprocesses. It had (and to some extent still has) aprofound influence on psychology and education.It has to be said, however. that Piaget was first and

35

foremost a psychologist and epistemologist. Hisdevelopmental model and his experiments wereintended to be diagnostic tools for classifying chil-dren's thinking processes and were not intended tobe transferred directly to the classroom.

Piaget's developmental model is based on theidea that the child's thinking progresses through anumber of stages (Fig. 2.4), each of which followson from the successful acquisition of the previousstage. There are three fundamental stages thesensori-motor stage (approximately the first 18months of life), the concrete operational stage (upto about 12 years) and the formal operationalstage (12-15 years) the last two being most rele-vant to primary and secondary education. Each of'these stages has sub-periods and sub-stages. Theconcrete operational stage, for instance, has twosub-periods: the pre-operational sub-period (18months to about 7 years) and the concrete opera-tional sub-period (approximately 7 to 12 years).The kind of thinking which characterises each ofthese stages and more importantly the limitationsof the thinking and reasoning at each stage,applies to the handling of all sorts of conceptsright across the curriculum. The child's progressfrom one stage to another occurs through aprocess of equilibration an interaction of cogni-tive growth and environmental input.

Stage I: Sensori-motor stage (birth-18months)

Stage II: Concrete operational stage

Sub-penod lia: Pre-operational(18 months-7 years)

Sub-period Ilb: Concrete operational(approx. 7-12 years)

Stage Ill: Formal operational stage (12-15 years)

Fig. 2.4 Piaget's developmental model

t ( )


Piaget's evidence on formal reasoning camelargely from fifteen experiments which wereundertaken by children between the ages of 4 and16. Each of these experiments consisted of a prob-lem-solving task not unlike investigations, duringwhich the subject was encouraged to experimentso that he or she could then explain certain phe-nomena to the observer. These tasks were used toexplore the kind of thinking that the children wereusing to tackle the problem. The fourth of Piaget'sexperiments - 'the oscillation of a pendulum'will serve to illustrate Piaget's theory and the rele-vance of this model both for the CASE projectand for the framework we are developing.

In the pendulum experiment, the children weregiven string which could be shortened or length-ened and a set of varying weights and asked to findout what factor(s) affect the oscillation of the pen-dulum. The pre-operational child's thinking ischaracteristically egocentric, in that she regards theworld from her own point of view, unaware thatthere are other points of view or that she is limitedby her own. For instance, the pre-operational childwill confirm her own theory regardless of the evi-dence and will often go on to contradict her ownprevious theories. In the case of the pendulum, shewill be unable to isolate a particular variable suchas weight. The concrete operational child will,however, be able to order length, weight. etc.. andbe able to make objective judgements. but willhave difficulty isolating one variable in a multivari-ate situation she may. for instance, alter stringlength and weight simultaneously. A child at thisstage. then, has difficulty with the idea of controlsand the concept of fair testing. The older child atthe third stage (formal operations) will, accordingto Piaget's theory, manipulate one variable at atime and be able to look at several combinations ofvariables and then arrive at a valid conclusion.

Piaget's theory has been the subject of muchcriticism. The idea, for instance, that the stages arerelated to chronological age was soon overtakenby the view that they are more likely to be relatedto mental age (e.g. Dodwell. 1961). Similarly, thenotion of the development of concepts across thecumculum being at the same cognitive stage simul-taneously has been disputed (see, for example,

3(3

Annett, 1959). More fundamentally, the relation-ship of Piaget's theory to education has been thesubject of much criticism (see Brown andDesforges, 1977; Rowell, 1984). His theory hasbeen interpreted in a variety of ways, most ofwhich have caused controversy. Rowell (1984) car-ries out a thorough and comprehensive review ofthe arguments for and against the import of psy-chological theory in education and points out someof the assumptions which underpin Piaget's theory.We shall not enter the debate, but in passing drawthe reader's attention to the fact that the basis ofthe CASE project is not uncontroversial. Weshould note, however, that Piaget used the tasks asindicators of levels of thinking, as assessment tools,and as such they are not necessarily central to theissue any appropriate tasks would do.

The CASE project

The CASE project (Adey, 1988. 1992; Adey andShayer. 1990) used Piaget's theory to develop anintervention strategy in the early years of secondaryschool which is designed to accelerate the develop-ment of formal operational thinking. Theintervention strategy focuses very specifically onactivities designed to promote types of reasoningwhich are characteristic of the formal operationalstage (Table 2.3). The CASE project uses a series ofscience activities, some of which are investigative innature, allowing pupils an opportunity to 'test ten-tative theories against reality' because 'It is only byinteraction with reality that a learner can test his/hermodels of the nature of reality' (Adey. 1992).

Table 2.3 The focus of the CASE activities

Control of variablesProportionalityCompensationProbabilityCombinationsCorrelationClassificationFormal modelsCompound variablesEquilibrium


The intervention consists of thirty activitieswhich are designed to be used alongside the'normal' science curriculum at the rate of aboutone a fortnight over a two-year period. The firstactivities focus on the relevant vocabulary (vari-ables. etc.). Then each of the reasoning patternswhich is thought to underlie formal operationalthinking is taken in turn and a lesson built aroundit. The underlying premise is that science requireshigher levels of cognitive thinking (see. forinstance, the early work by Shayer (1972. 1974) onthe Nut Yield 0 level syllabus). So if Piagetian taskscan be used in reverse, so to speak, to acceleratecognitive development, then pupils will be betterable to cope with science. Again the emphasis isnot on the particular tasks so much as on the typeof task that articulates with the pupil's developinglogical structures.

Activity 3. for ..:xample. which follows twoactivities designed to introduce the idea of vari-ables, is built around the idea of the fair test.Children are given a variety of lengths and widthsof tubing in different materials and asked to inves-tigate- the effects of the variables (length andwidth) on the note produced when the tubes areblown across. The children arc instructed to trythe tubes in pairs and asked whether particularpairs of tubes provide fair tests. The method isprescribed, although the control of variables is notspecifically stated. The pupils are given work-sheets. which by providing a table with thevariables as headings, direct the pupils towardsrepeating their readings and the method of record-ing and presenting their results. Pupils are thenasked specific questions such as:

Wkit is the effect of hwgrh?Which experiments tell ou this?What is the effect of materia?Which experiments tell you this?

(Ade al. 1989 )

In this way, they arc directed towards makinga generalisation or conclusion. The worksheetprovided therefore quite tightly controls theactivity in terms of the method, the questions tobe answered, the format of the results and the

37

conclusions drawn, and in this sense the activitiesare closed. A series of written fair test problems isprovided as a follow-up to the practical.

We can compare this activity with a typicalinvestigation (Fig. 2.5). The similarities are clear.Each task is concerned with the identification ofthe effects of more than one independent variable(material, length and width of the tubes: lengthand width of the beams). However, in the investi-gation, the problem is defined but the method isonly limited in so far as the equipment is provided.The pupils can then decide how to use the equip-ment. what measurements to take, how manymeasurements to take and so on. so that .themethod and means of arriving at a solution areopen. We shall return to the difference betweenthe CASE activities and investigations later in thenext section.

A decoratingproblem

Sally and Sam watched Mum and Dad decorating.To reach a high place, Dad put a plank of wood betweentwo chairs.

Mum said. 'The plank will bend too much. You're too heavy.'Sally said, 'If you use a different plank, it won't bendas much.'Sam said, 'If you use a wider plank, it won't bend as much.'

Who was right?

1-F-I-nd out whether the amount the wood bendsdepends on(a) the type of material which is used, or(b) the width of the material, or.(c) both of these

Write a clear report saying what you did and whatyou found out. Don't forget to show your results.

Fig. 2.5 Example of an imestigation (from Foulds ei al..19921


As well as the individual activities, there arethree essential features in the teaching of CASEwhich we shall consider in turn: cognitive conflict,metacognition and bridging.

Cognitive conflict

Cognitive conflict refers to the situation in which apupil is confronted by results which do not tit hisor her existing expectations. The conflict meansthat the pupil may be forced to 'equilibrate' orreconstruct his or her thinking in order to accom-modate the ncw evidence. Since conflict stimulatescognitive development, pupils are encouraged toacknowledge and consider conflict when it occurs.It is of course possible to avoid cognitive conflictby ignoring the results or by accommodating con-flicting evidence.

Aletacognition

Metacognition here refers to the process wherebythe teacher encourages pupils to reflect on theirown thinking processes. For instance, the classmight discuss what aspects of an activity theyfound difficult and why.

Bridging

Bridging, the third feature of CASE teaching. isusually in the last part of the lesson when theteacher draws the pupils' attention to the use ofthe relevant reasoning pattern in completely dif-ferent contexts, in science or elsewhere. Clearly,this feature is designed to promote transfer; thatis, the transfer of learning from one situation toanother, which we defined as application in anearlier section.

The CASE intervention strategy was first triedfrom 1984 to 1987 with pupils in years 7 and 8.Cognitive development was measured using thePiagetian Science Reasoning Test both before andafter thc intervention. The results suggested thatimmediately after the intervention, the experimen-tal group of pupils showed significantly betterlevels of cognitive development than the controlgroup, but no better performance in science in

each school's end of year science tests. Two andthree years after thc interventIon, however, theexperimental group performed significantly betterat GCSE in science as well as in mathematics andEnglish. These early results are encouraging andhave the potential to have a profound influence oneducation. There is. however, a need for large-scale replication, particularly since the size of the1987 sample was limited. The subsequent analysis,while not denying the potential importance of theapproach. has come in for some criticism (Preece.1993). There is clearly a need for further empiricalwork on the effects of the CASE approach.

Concepts of evidence and CASE

What is the relationship between the psychologicalperspective of the CASE intervention and the tax-onomic approach put forward in the first half ofthis chapter? Clearly. the CASE intervention has amuch wider goal in that it aims to improve think-ing skills in general which will be applicable inother subject areas. Our taxononlic approach istargeted much more specifically on science.

There are, however, similarities between thefocus of some of the CASE activities and some con-cepts of evidence such as variables and probability.There are also areas which are not covered by theCASE approach and vice versa. The three essentialfeatures of CASE teaching (namely, cognitive con-flict. bridging and metacognition), however, havemuch in common with well-taught investigativework, as we shall see in Chapters 6 and 7.

We have already noted that the CASE activitiestend to be relatively closely controlled, whereasinvestigative tasks, which allow children theopportunity to apply and synthesise conceptualand procedural understanding. are 'open' in thesense of. for instance, allowing children to choosetheir own methods. The more fundamental differ-ences between these two approaches are:

1 The CASE approach comes at the problem viaan assumption that the intervention will accel-erate cognitive development so that the pupilwill be better equipped to do all of science as

3 3


well as other subjects. The taxonomic approachassumes that concepts of evidence are 'thingsthat can be taught' as part q. science, which willincrease individuals' capacity to cope withproblems simply because they know moreabout how to set about it as well as the abilityto evaluate other people's evidence. Ultimately.the approach aims to improve proceduralunderstanding.Following the CASE intervention, pupils areexpected to return to the teacher's 'normal'scheme of work, which may not include anyfurther tasks which focus on concepts of evi-dence. The taxonomic approach suggest.; thatconcepts of evidence be gradually taught anddeveloped through progressively more difficulttasks and then regularly practised and rein-forced within investigations as part of theteacher's scheme of work.

3 The CASE approach focuses on individual con-cepts of evidence. whereas the taxonomicapproach to procedural understanding aims toimprove the understanding. application and syn-thesis of similar concepts by enabling pupils tocarry out whole investigations and put theirunderstanding into practice. Investigations.therefore, allow opportunities for pupils todemonstrate the highest level of cognitive ability.

We would argue that defining the science cur-riculum to include concepts of evidence is likely tobe the more productive way of looking at theissue: its inclusion in the list of things to be taughtand learnt would ensure its presence in the class-room so that procedural understanding is taughtin its own right.

Summary

To arrive at a framework we have used a modeland a classification based on Bloomian taxonomy.which have enabled us to locate the developmentsin practical science outlined in the last chapter andto describe in some detail the cognitive abilitieswhich are inherent in sound practical science.While the taxonomy has its weaknesses which we

39

have discussed, it has been used here to help usunravel some of the complexities of 'doing' sci-ence. It has also enabled us to demonstrate thatthe levels of cognitive ability commonly used inrelation to the conceptual component of sciencecan be applied in a similar way to the proceduralcomponent. We have introduced the idea of con-cepts of evidence which will be used to structurethe discussion of the data in forthcomine chapters.We hope that we have persuaded the reader thatprocedural understanding is more than a matter ofrecalling and using skills and procedures. butrather that it is a set of understandinQs which areimportant in their own right. This understandingis important not only in practical science but alsoas a means of effectively examining evidence fromother sources.

The CASE intervention which adopts a psycho-logical perspectiw has been considered as analternative way' of teaching concepts of evidence.While both approaches teach concepts of evi-dence. one of the key differences is that thepsychological approach of' the CASE interventionis aimed at increasing the scientific reasoning abil-ity throuith a series of tasks which then. inprinciple at least. become redundant. The taxo-nomic approach to procedural understanding, onthe other hand, aims at promoting the applicationand synthesis of concepts of evidence which areseen as a `content' of the curriculum in their ownright. Investigations provide the opportunity forpupils to put such ability into practice. In the nextchapter. we shall look more closely at the defini-tion of investigations.

References

Adey, P. (1988). Cognitk e acceleration: Revitm andprospects. International Journal of Science Education,10(2): 121 34.

Adey. P. (1992). The CASE results: !mplications forscience teaching. International Journal of ScienceEducation. 14(2): 137 46.

Adey, P. and Shayer, NI. (1990). Accelerating the devel-opment of formal thinking in middle and high schoolpupils. Journal of Research in Science Teaching,27(6): 553 74.

40 INVFSTIGATIVE WORK IN THE SCIENCE CURRICULUM

Adey, P.S.. Shayer, M. and Yates, C. (1989). ThinkingScience: The Materials of the CASE Project.Teacher's Pack. London: MacMillian.

Adey, P.S.. Shayer, M. and Yates. C. (1989). ThinkingScience: The Materials of the CASE ProjectTeacher's Pack. London, Macmillan.

Annett, M. (1959). The classification of four commonclass concepts by children and adults. British Journalof Educational Psychology, 29: 223-35.

Blagg, N.R.. Ballinger. M.P.. Gardner, R.J., Petty, M.and Williams. G. (1988). The Somerset ThinkingSkills Course: Foundations lOr Problem-solving.Oxford, Blackwell.

Bloom. B.S.. Engelhart. M.D.. Furst, E.J., Hill, W.H. andKrathwohl, D.R. (1956). Ta.vonomy of EducationalObjectives: The Cognitive Domain. New York,Longmans. Green.

Brown, G. and Desforges, C. (1977). Piagetian psycho-logy and education: Time for revision. BritishJournal of Educational Psychology. 47: 7-17.

Coles. M.J. and Robinson, W.D. (1989). TeachingThinking: A Survey of Programmes in Education.Bristol. Classical Press.

Dodwell, P.C. (1961). Children's understanding of numberconcepts: Characteristics of an individual and a grouptest. Canadian Journal of Psychology, 15: 29-36.

Duggan, S. and Gott, R. (in press). The place of investi-gations in practical work in the UK NationalCurriculum foi Science. International Journal ofScience Educotion.

Foulds, K., Gott. R. and Feasey. R. (1992). 'InvestigativeWork in Science'. Unpublished research report.University of Durham.

Gott, R. and Mashiter, J. (1991). Practical work in sciencea task-based approach? In: Practical Science (B.E.

Woolnough, ed.). Buckingham, Open University Press.Inhelder, B. and Piaget, J. (1958). The Growth of Logical

Thinking from Childhood to Adolescence: An Essay onthe Construction of Formal Operational Structures.London. Routledge and Kegan Paul.

Kempa, R. (1986). Assessment in Science. CambridgeScience Education Series. Cambridge, CambridgeUniversity Press.

Preece. P.F.W. (1993). Comment: Cogniti% accelera-tion and science achievement. Journal of Research inScience Teaching. 30(8): 1005-6.

Rowell, J.A. (1984). Many paths to knowledge:Piaget and science education. Studies in ScienceEducation. II: 1-25.

Shayer. M. (1972). Conceptual demands in Nuffield 0level physics. School Science Review, 186(54): 26-34.

Shayer, M. (1974). Conceptual demands in the Nuffield0 level biology course. School Science Review,56(195): 381-8.

Williams, R.G. and Haladyna. T.M. (1982). Logicaloperations for generating intended questions(LOGIQ): A typology for higher level test items. In:.4 Technology.for Test-item Writing (G.H. Roid andT.M. Haladyna. eds). London, Academic Press.

Young, S. (1993). Notching the brain cells up a gear.Times Educational Supplement, 12 February.

CHAPTER 3

Investigations: What are they?

In the last chapter, we developed a frameworkfor practical science wit hin which we located therole of investigations. But what qualifies as aninvestigation, and are there different types ofinvestigations? Lack of clarity of definition hasbedevilled (and still does bedevil) education ingeneral, and science education in particular. Inthis chapter, we shall consider some majorresearch projects which have concerned them-.selves with investigations and which havedeveloped ways of classifying different types ofinvestigations.

Problem-solving and investigations

There has been some confusion about therelationship between problem-solving and investi-gations which we shall consider here briefly.Problem-solving is a general term which has beenapplied to many subject areas. In maths and sci-ence, for instance, it has frequently been appliedto cognitive, written problems. Its place in scienceeducation has been reviewed by Garrett (1986),who ends his paper: 'As Sham (1976) has pointedout, the whole field of endeavour in problent-solving is particularly vast and largely disorgan-ised and this has been shown to be true even inthe limited arca of science education.'

Watts and Gilbert (1989) have attempted toclassify problem-solving in science into two kindsof tasks: first, the paper and pencil tasks which arewell defined and have little or no redundant infor-

mation. Watts and Gilbert call these PSI tasks.PSI tasks were particularly prevalent in the 1970sand mid-1980s. Second, there are the wide varietyof 'ill-defined' problem-solving tasks whichemerged in the late 1980s, called PS2 tasks. By ill-defined they mean here tasks 'where only outlinerelevant information and materials are supplied'(Watts and Gilbert, 1989). Watts and Gilbert sug-gest that PS2 tasks have grown out of the searchfor a means of making science relevant and ofallowing pupils to apply scientific principles. Thesetasks can be either written or practical but have astrong emphasis on skills and methods and includepuzzles, design-and-make activities, and extendedproject work. Investigations can be seen as onetype of problem-solving, in science, whose defini-tion we shall now consider.

The Assessment of Performance Unit

The Assessment of Performance Unit (APU) wasset up by the Department of Education andScience (DES) following political debates in the1970s which expressed concern about standardsin education. Its brief was wide-ranging and con-siderable funds were made available for aninnovative approach to assessment. The intentionwas that a number of subjects, including maths,science and English, would be assessed using a'light sampling' process. The light samplingprocess allows tests to be created which are muchlonger than could be taken by any individual

4 i


pupils. These extended tests, in which severalpapers were produced and given to different, butparallel, samples of pupils, the results then beingaggregated, provided the opportunity for a wide-ranging review of what constituted science inschools at that time. Hence, although no pupil wastested for more than an hour, the compositeresults were equivalent to a test lasting as long as19 hours. Pupils in both primary and secondaryschools in England and Wales at the ages of 11, 13and 15 were sampled. The assessment was basedon a framework comprising six science activitycategories (SACs). It should be noted that theAPU used the term science 'activity' rather than'process'. which suggests mental process. Theirwork focused, it claimed, on the assessment of the'doing' of science rather than on cognitiveprocesses. The six activity categories were:

1 t!sing symbolic representation.2 Using apparatus and measuring instruments.3 Observation tasks.4 Interpretation and application.5 Planning of investigations.6 Performing investigations. (API.% 1985)

It is with the last category, the performance ofinvestigations, that we arc concerned here.

The APU defined an investigation as 'a task forwhich the pupil cannot immediately see an answeror recall a routine method for finding it'. TheAPU recognised that this type of practical workwas different from other types of practical work.Hence, a report in 1989 stated that: Performinginvestigations" enjoyed a unique status in that it

was perceived and justified as the embodiment ofan important aim of science teaching whichencompasses more than the separate elements rep-resented in the other Categories, all of which areinvolved in it' (APU, 1989).

The types of task defined as investigations bythe APU (1987) and an example of each are shownin Table 3.1. In practice, the constraints of thenational assessment (e.g. time and cost of equip-ment) limited the number of types of tasks thatcould be used. The last thrce types in Table 3.1were developed and trialled, but not used in large-scale surveys. The initial aim was to selectinvestigations within these two types which didnot rely heavily on concepts. In that way, it waspossible to assess a single facet of pupil perform-ance procedural understanding. Some of theinvestigations used are shown in Table 3.2.

The first four of these investigations are of the'decide which...' type and the rest are of the 'find theeffect of...' type. It will be seen that the investiga-tions span a wide range of 'contexts'. Context hererefers to the wording in which the investigation isembedded. In the main, the investigations are set inan 'everyday' (familiar) as opposed to a 'scientific'context, and they vary in the degree to which thequestion is defined. 'Swingboard', for instance,defines the variables to be tested (i.e. length andwidth), while 'Flooring' is entirely open since thepupil has to decide what 'suitable' means. The chil-dren were given equipment from which to choose,but other than this their choicc of method wasentirely open. Children were observed individuallywhile performing investigations. The emphasis on

Table 3.1 The types or questions defined by the APU as investigations

Problem type Example

Decide which...

Find the effect of...

Find a way to...

Find the cause of...

Make a structure/machine to...

...kind of paper towel will hold the most water.

...the water level in a container on the rate at which the water runs out of a holein the bottom.

...adapt weighing scales that won't measure up to the baggage allowance.

...the failure of a light bulb to light a circuit.

...support a brick using one newspaper and sellotape.

4 2

INVESTIGATIONS: WHAT ARE THEY?

Table 3.2 The APU (1985) investigations

43 1

Question title Investigation

Survival Which fabric would keep you v;irmer? (given 2 fabrics)

Cars If all the cars are given the same chance, which one will travel furthest? (given 3 cars)

Paper towel Which kind of paper will hold the most water? (given 3 kinds)

Flooring Which one of the floor coverings do you think is the most suitable for a kitchen floor? (given 4 types)

Woodlice If woodlice are given a choice of the four placer below, which one do they choose to lie in? A placewhich is: damp and dark, dry and dark, damp and light, or dry and light?

Hotwash Does this washing powder wash a dirty cloth as clean in cold water as it does in hot water?

Candle How does the angle of a wax taper affect its rate of burning?

Swingboard What difference does changing the length and width of the board make to how quickly it swings?

procedures was accompanied by a growing aware-ness of the importance of variables in investigationsin defining their difficulty.

A descriptive model (Fig. 3.1) was developed bythe APU teams primarily to consider those aspectsof performance which it was thought appropriateto assess. The model provides a more detaileddescription of what is going on when pupils per-form investigations. This model is not supposed torepresent the mental processes that pupils must gothrough in order to carry out an investigation, butrather it is a list of things that can be done, notnecessarily in that order and not necessarily doingall of them. The intention behind the model was todescribe an iterative approach, with the investiga-tor continually evaluating decisions and adjustingas necessary. The first cycle around the loopmight, for instance, be nothing more than a trialrun to get the 'feel' of the quantities involved.

The APU data had considerable impact inschools. While at first it was seen as a threat toschool autonomy, it was later seen to have devel-oped innovative techniques which worked well inthe classroom as curriculum material rather thanmerely as assessment items. In particular, theinvestigations were found to be of considerableinterest to pupils in terms of enjoyment.

We shall discuss the findings of the APUresearch in detail in the next chapter.

Skills, concepts of evidence and the A PU problem-solving model

The problem-solving model used by the APU hasbeen used widely in science and indeed in othersubjects. The factors itemised in the model encom-pass the skills and concepts of evidence defined inChapter 2. The understanding which guides theongoing evaluation, the iterative loops. are. interms of concepts of evidence, equivalent to thenotions of validity and reliability. This continuousreflection on the design and implementation in thelight of the problem as set and the requirement ofthe data to answer it is, as we shall see later, thesingle most important factor missing in pupils'work, in all investigations and at all ages.

The Open-ended work in Science (OPENS) project

The OPENS project was a three-year research pro-ject also set up by the Department of Educationand Science, to explore how open work can bestbe incorporated into science curricula and how itcan be assessed. It was divided into two phases.The first phase reviewed the understanding andpractice of open work among teachers of years 7-1 I in secondary schools (Simon and Jones, 1992)while the second phase used this knowledge to testout the development of open work in schools(Jones et al., 1992). The project defined 'open

4 ,)

Problemidentification


Solution

Furtherreformulation

Evaluation ofmethods and

results

Reformulaticn ofa testable

question, decidingwhat to measure Change in

design

Interpreting thedata and drawing

conclusions

Planning theinvestigation

Change intechnique

Recording thedata in tables and

graphs

Carrying out theinvestigation,

makingobservations andmeasurements

Fig. 3.1 A model for problem-solving activity (Gott and Murphy. 1987)

work' as 'activities which give the initiative to stu-dents for finding the solution to problems. Theseactivities place an emphasis on autonomy inmaking decisions and on the integration of knowl-edge and skills' (Simon and Jones. 1992).

The project reviewed what teachers of years 711 mean by open activities and found that theterm is used to refer to a wide variety of tasks1)c.luding investigations, but also extending to

project work, model-making and surveys. In orderto classify these tasks according to their degree ofopenness, Simon and Jones suggest that threestages in the 'doing' of the activity need to be con-sidered: defining the problem, choosing themethod and arriving at a solution (Fig. 3.2). Thefirst two stages in any task can be positioned on a

4 4

continuum ranging from 'closely defined' to 'notdefined'. The last stage, arriving at a solution, canbe positioned on a similar continuum rangingfrom activities where there is only one solution tothose where there are many possible solutions (forexample. a survey). Clearly. tasks considered to bezTen lic to the right of the continuum in one ormore of their stages.

This framework was developed to enable teach-ers to see how they could manipulate the degreeof openness in tasks, which in turn depends onwhat they want their pupils to learn. Simon andJones point out that moving the stage of a taskfrom left to right on the continuum generallyresults in moving the initiative from the teacher tothe pupils.


Defining the problem

Closely defined Notdefined

Choosing the methods

Closely defined )0- Notdefined

Arriving at solutions

One Many

Fig. 3.2 The OPENS continua (based on Simon andJones. 1992)

Returning to the definition of investigationsarrived at in the previous chapter. we wouldsuggest that activities that lie to the extremeclosed end of each of the three continua, cannotqualify as investigations since they severely limitthe opportunity for pupils to use and apply pro-cedural understanding. On the classificationdeveloped in the previous chapters, activities ofthis kind might fall into the categories ofenquiry or illustrative practicals. Some examplesof tasks and how they might be crudely classi-fied on the extremes of the continua are given inTable 3.3.

Task 1 in Table 3.3 is a closed task on all thecontinua and qualifies as an enquiry-based prac-tical in that it is carefully structured to enable all

Table 3.3 Classifying tasks on the OPENS continua

45

pupils to reach the same endpoint. Task 2 is aninvestigation, since although the task defines thedependent variable (the number of seeds that ger-minate), it is otherwise open in that the method ischosen by the pupil and the solution could be oneof many. Task 3 is clearly an open task on allthree continua. If the investigations used in theAPU survey and those used in the research to bereported here are classified using the OPENSframework, it becomes clear that in the 'definingthe problem' stage, the investigations fall largelyat the 'closely defined' end. The tasks given to thepupils were determined by the teacher/researcher,while choosing methods and arriving at solutionswere open. The 'defining the problem' stage wasclosed for pragmatic reasons. To compare perfor-mance in research terms, pupils had to be giventhe same tasks, otherwise there would have beenno effective control of the task they actually car-ried out.

The examples overleaf drawn from our researchshow how somewhat diffcrent methods canemerge when groups of pupils are presented withthe same task.

Here we can see that Ross and Kevin designedtheir investigation so that the weight required tomake the bridge sag by 10 cm was the dependentvariable, while in Simon d Gavin's design. thesag of' the bridge with 150 g weight was the depen-dent variable.

I Put 20 seeds on the windowsill and 20 seeds in a dark cupboard. Leave them for 10 days. Count how many seedsgerminate in cad- batch.

2 How does the amount of light affect the number of seeds that germinate?

3 What factors affect germination?

Task DOning flu, problem Choosing tlw nwthod .4mving at solutions

closed closed closed

2 closed Open open

3 open open open


Salida *ger bridge

The hridge'scross the river sags when ears go across it.

If you were going to build a new bridge across the river which material would sag the most?

Ross and Kevin (age I /)We placed the Nescafe tins 30 cm (1 ft) apart and put the plastic over the top making sure we didn't move thetin. Each tin was 12.5 cm high. We piled weights on top of the material until it sagged to 10 cm in height.

Material Length between Amount of weight Height we allowed ittwo tins the material stood to sag to

Plastic 30 cm 800 g 10 cm

Wood 30 cm 1900 g 10 cm

Cardboard 30 cm 400 g 10 cm

The wood is the strongest material that we used and is therefore the most reliable for standing on between 2 objects.

Simon and GavinMethodWe took 2 tripods and placed 4 different materials across them. We then put 50 g and 100 g weights in themiddle. When the materials bent, we measured how far down it dipped.

Blue White Wood

We found that the softer the material the farther it bent down.

Card


The Assessment of Practical Work in Science(APWIS)

In Chapter 2, \k'e. saw how APWIS (Gott et al..1988) classified practical work into different types.but what is relevant here is that the authors alsodeveloped the definition of investigations furtherby using a classification of the variables within theinvestigation. The classification aimed to describeand then to assess the way in which pupils inter-preted the nature of a variable.

APWIS used the terms 'categoric'. 'discrete'and 'continuous'. A variable which is categoric isdefined descriptively. For example, the type ofinsulation such as polystyrene or fibreelass is acateeoric variable. A variable which is definednumerically but which takes only inteeer values isdefined as a discrete variable. An example wouldbe the number of layers of insulation around aheating tank. Finally:a continuous variable is onewhich is defined numerically and which can takeany value, such as the thickness of an insulatinglayer of polystyrene beads.

Later. Foulds and Gott (1988). while acknowl-edging the influence of concepts and context(among other factors). developed a typology oftasks based lamely on the variable structure of theinvestigation. They used the typology to suggestlevels of difficulty associated with proceduralunderstanding. They proposed grouping investiga-tions into four main types, each with a differentvariable structure (Table 3.4).

From classroom experience, it was suggestedthat a type 1 question is likely to be easier for most

47

pupils than a type 2 question. Many pupils opt todefine a continuous variable as categoric: tempera-ture may' be defined in the categories hot, warmand cold. Type 3 is likely to be more difficult thaneither types 1 or 2. since it involves multivariatedesigns which many pupils find difficult. A varia-tion of the multivariate design is the biological-typecontrol experiment, where one value of the inde-pendent variable forms the control or standard.For example. if the investieation is to compare theeffect of different fertilisers on growth. then oneplz..nt would be given none. Type 4 is more to dowith technological problem-solving.

Imestigative work in science

The project which gave rise to this book set out toresearch pupils' performance on all aspects ofScience 1 (Scl) in the National Curriculum, withparticular reference to progression in children'sunderstandine. We shall refer to this research asthe 'NCC project' (Foulds et al.. 1992). since itwas prepared for the National CurriculumCouncil (NCC), the research being jointly fundedby the NCC and the DES.

The project was carried out between April 1990and September 1991. with the specific aims of(a) documenting how Sc I was being implemented.(b) identifying teaching methods which promotethe effective integration of Scl with knowledge andunderstanding, and (c) identifying elements of pro-gression in the levels of Scl. The sample consistedof over 3500 children undertaking investigationscovering Key Stages I. 2 and 3. Of these. over 2000

Table 3.4 A typology of investigations (Foulds and Gott, I98S)

QIWAl ion Example

1 A single categoric variable Which is the best type of insulation for a hot water tank?

2 A single continuous variable

3 More than one independent variable

Find out how the rate at which the water cools is dependent on theamount of water in the tank.

Is it the type of insulation material or its thickness which is keeping thewater hot?

4 Constructional activities Make the best insulated hot water tank


Table 3.5 Types of investigation used in the NCC project and their variable structure

Ty Pe Independent Independent Dependent

variable I variable 2 variable

1 categoric continuous

2 continuous continuous

3 categoric categoric continuous

4 continuous continuous continuous

Example

Find out which fruit gives the biggestvoltage

Find out how the voltage depends on thedistance between the strips of metal

Find out whether the voltage of the celldepends on the type of metal, the type offruit or both of these things

Find out how the voltage of the celldepends on the distance between the metalstrips and the amount of metal strip underthe surface

children were in secondary education, theseschools beim!, in five education authorities in thenorth-east. Two hundred and ninety primary andsecondary teachers from sixteen LEAs completedquestionnaires regarding teachers' perceptions ofSc 1. Fifty of these teachers were subsequentlyinterviewed for validation purposes.

The NCC project defined investigations, fol-lowing the APU line, as 'tasks which revolvearound a practical problem for which there is aminimum of instructions'. The investigations inthe project were selected on the basis of beingaccessible to the majority of pupils. Proceduralcomplexity was manipulated through the variablestructure of the task. Tasks were classified intofour types according to the nature of the indepen-dent variable(s). The structures of these task types.with an example of each, are given in Table 3.5.

It will be seen that the NCC types I and 2 are thesame as the APWIS types 1 and 2, but the multi-variate designs are now divided into types 3 and 4.Discrete variables were not included. For clarity,the examples in Table 3.5 are drawn from one con-text, which is that of a 'fruit battery' where twodifferent metals inserted into a fruit produce a volt-age. The voltage depends on the area and type ofthe metals, the distance thcy are apart and the typeof fruit. This example also serves to demonstrate

how a single context can be structured in a varietyof ways to provide the basis for a range of complex-ity of investigations. These types of investigationswill be referred to frequently in discussing the dataarising from the project in the next two chapters.

A note on the limitations of focusing on the variablestructure of investigations

If we take a restricted view of investigations asbeing solely to do with variables and numericaldata, then large swathes of science, particularlychemistry and those elements of science borderingon technology, can become neglected. This hasproved a problem with the National Curriculum inthe UK. A broader viewpoint would consider notsimply variable-based tasks,but also other types ofinvestigative work summarised in Table 3.6.

Recalling our arguments of the last chapter,we can see that all of thesc types have incommon the requirement, to a greater or lesserextent, that pupils synthesise skills and conceptsof evidence in arriving at thcir solution. Anexample of a logical reasoning task will help toillustrate this point: in the context of a forensic-type detective activity with indicators or otherqualitative tests, pupils arc required to carry out

4 0


Table 3.6 Types of investigative work

49

Type gif Mvestigative work Example

Variable-based

Logical reasoning tasks

Measurement-focused

Constructional/engineering tasks

Constructional/technological

Explorations

Type 1 (involving a categoric independent variable (see table 3.5)Type 2 (involving a continuous independent variable)Type 3 (involving more than one categoric independent variable)Type 4 (involving more than one continuous independent variable)

Tasks involving the carrying out of a sequence of (often qualitative) tasks, the datafrom each of which structures subsequent tasks leading to a solution of the problem(forensic science, electrical fault-finding)

Often in science, engineering and technology the key problem is finding a way to measurea variable quantity which cannot be measured directly with available instrumentation

'Engineering' tasks aim to produce a solution to a problem and then test itseffectiveness, rather than investigate the underlying factors. Making the best-insulatedhot water tank, for instance, requires that the design be optimised

This type of task would encompass such activities as the construction of an electricalcircuit where the criterion of success is concerned with whether or not the circuit doesthe job required

The most open of the types involving pupils in raising the question and defining thetask prior to developing a method of solution which may involve any ot' the typesabove (and other resource-based, non-practical work)

a relatiely simple task, the pH of some solutionperhaps, which then forms the basis for the selec-tion of another test. Hence there is a progressivenarrowing down of options for example, if thepH is 4 then it couldn't be x or y, but it could bep or q. To distinguish p and q, pupils would needto do a flame test and so on. This is akin to the'Find the cause of...' tasks of the APU. Here thesynthesis of skills and concepts of evidence suchas repeatability, reliability and validity are neces-sary to reach a solution.

Summary

The projects described in this chapter have led usto formulate a definition of investigations and toclarify what types of problems constitute investi-gations, while acknowledging that wc are onlyfocusing on one type of investigative work (vari-able-based investigations). Most of the research

that has been done in investigative work has beenconcerned with tasks which can be defined by thevariable structure we have described, the variable-based types in Table 3.6. The reasons for this maybe that such tasks are quantitative and are easierto define in terms of assessment criteria. There isalso the fact that Piagetian theory has had aninfluence on science education for many years andPiaget's work includes experiments which arelargely variable-based.

The omission of other kinds of investigativework was recognised, but within the constraints ofthe APU and for the purposes of APWIS andNCC, the number and structural complexity ofinvestigations had of necessity to be limited. Thenotion of concepts of evidence which we devel-oped in the last chapter was also derived fromvariable-based tasks. ln reality, research into pro-cedural understanding is in its infancy and weshould regard focusing on variable-based tasks asbeing no more than a start.

50

References


Assessment of Performance Unit (1985). Science inSchools: Ages 13 and 15. Research Report No. 3.London, HMSO.

Assessment of Performance Unit (1987). AssessingInvestigation at ages 13 and 15. Science Report forTeachers: 9. London, HMSO.

Assessment of Performance Unit (1989). National Assess-ment: The APU Science Approach. London, HMSO.

Foulds, K. and Gott, R. (1988). Structuring investiga-tions in the science curriculum. Physics Education,23: 347-51.

Foulds, K., Gott, R. and Feasey, R. (1992). InvestigativeWork in Science. Durham, University of Durham.

Garrett, R.M. (1986). Problem-solving in science educa-tion. Studies in Science Education, 13: 70-95.

t-siu

Gott, R. and Murphy, P. (1987). Assessing Investigationsat Ages 13 and 15. APU Science Report for TeachersNo. 9. London, DES.

Gott, R., Welford, G. and Foulds, K. (1988). TheAssessment of Practical Work in Science. Oxford,Blackwell.

Jones, A.T., Simon, S.A., Black, P.J., Fairbrother,R.W. and Watson, J.R. (1992). Open Work inScience: Development of Investigations in Schools.Hatfield, Association for Science Education.

Simon, S.A. and Jones, A.T. (1992). Open Work inScience: A Review of Existing Practice. London,King's College London.

Watts. D.M. and Gilbert, J.K. (1989). The 'new learn-ing': Research, development and the reform of schoolscience education. Studies in Science Education,16: 75-121.

CHAPTER 4

Pupils' performance of investigations insecondary schools: An overview

In the preceding chapter, we considered what kindsof practical work can be classified as investigationsand, within that definition, how investigationsmight be further categorised into different typesaccording to their structure.

In this chapter, we shall consider the mainfactors that are likely to influence overall per-formance. Chapter 5, will then take a detailed lookat how children perform investigations and howthey deploy their procedural understanding. Thiswill enable us to find out which parts of an investi-gation most children do reasonably well and atwhich points they experience difficulty. Clearly,both aspects of research on children's performancehave implications for teaching, progression and theissue of assessment, which we shall consider in sub-sequent chapters.

Factors affecting performance

What makes one investigation more difficult thananother at a more general level? In this chapter,we shall try to answer this complex question withreference to existing research findings. Some of thepossible factors that may influence the perform-ance of children doing investigations are shown inFig. 4.1

The An, research pointed to three of these keyfactors which influence the level of difficulty inany investigation:

the difficulty of the substantive conceptsinvolved (1 in Fig. 4.1);the context within which the investigation is set(4);the procedural complexity of the investigation(2), in terms of its variable structure.

The NCC project was designed to explore thesesame factors, together with two others: the 'open-ness' of the question and the age of the pupils. Theresearch started from the hypothesis that, of thesefive factors, the major factors influencing per-formance would be:

the substantive concepts which underpinned thetask and the level of difficulty of the conceptswithin that subject area (I in Fig. 4.1);the procedural complexity as defined by thetask types 1-4 defined in Table 3.5 in the lastchapter (2);the age of the pupils (3).

The secondary factors considered were:

context (4); andopenness (5).

The brief for the project was to investigate therelative effect, in so far as that is possible, of thesevarious factors with a view to outlining the major

-4 4 Context7 Teachers'perceptions


1 Concept 2 Proceduralarea complexity 3 Age

CHILDREN'SPERFORMANCE

OF INVESTIGATIONS

6 Pupil factorsmotivation, expectations,

perceptions, gender,culture

Fig. 4.1 Factors affecting pupils' performance of investigations

factors to be considered when building progres-sion into the design of a curriculum and itsassessment. The brief predetermines the research&sign to the extent that large samples are neces-sary, both of pupils and tasks, if any sort ofgeneralisable statement is to be made. Before con-sidering the research evidence about proceduralunderstarkling, we shall need to outline the sampleand methodology of the NCC project in broadterms, to give the reader some idea of the weightwhich can be placed on the evidence.

Table 4.1 Sample size by age for each type of investigation

5 Openness

The NCC sample, design, methodology and analysis

Sample

The breakdown of the sample into the numbers ofchildren performing each type of investigation isshown in Table 4.1. In all, twenty-three differentinvestigations were used. The data which were col-lected for type 4 investigations are restricted totwo concept areas only and a smaller sample: gen-eralisations should therefore be treated withparticular caution.

Type ofinvestigation"

Number ofinvestigations ofeach type

Sample size

Year 7(age 11years )

Year(age 12years)

Year 9

(age 13years)

Total

10 271 395 256 922

2 7 158 410 69 637

3 4 76 298 101 475

4 2 84 47 43 174

Total 23 589 1150 469 2208

"See Table 3.5 for definition.

PUPILS' PERFORMANCE OF INVESTIGATIONS IN SECONDARY SCHOOLS: AN OVERVIEW

The design

The investigations were chosen to give details onthe influence of the predicted major and secondaryfactors on performance as well as the interactionof some of these factors, as will be shown below.

Conceptual demand or difficulty

The investigations were chosen so that they covereda range of substantive concepts (see Table 4.2).Some groups of investigations were designed totest the effect of increasing conceptual difficultywithin one concept area. An example of a group ofinvestigations which all concern forces and motionis given in Table 4.3.

Table 4.2 Concept areas of the NCC investigations

53

Procedural complexity

Procedural complexity was defined by the variablestructure of the task (for definitions of types ofinvestigations, see Table 3.5). To recap briefly here,types I and 2 involve a single independent variable,while types 3 and 4 both have two independent vari-ables. Types 1 and 3 involve categoric independentvariables, while types 2 and 4 involve continuousindependent variables The tasks in the concept areaof forces and motion are outlined in Table 4.4.

The interaction of concepts and proceduralcomplexity

This interaction was tested by designing investi-gations in a range of concept areas spread across

Concept area Context and concepts underlying the tasks

Electricity from A fruit battery, in which pupils are asked to investigate the effects of factors such aschemical reactions separation of the electrodes, their depth and the types of fruit on voltage (see Table 3.5)

Forces the flexibility Bridges (the effects of factors such as type of material, width or length of material, orof materials the weight on the bridge on the amount the `bridge' sags)

Dissolving Sugar (or a 'chemical') dissolving in tea or coffee (the effects of factors such as the typeof sugar or the temperature of the water on the dissolving time)

Forces and motion A model car fired from an elastic band launcher (the effects of factors such as theamount of energy in the elastic band or how much it is stretched, on the distance travelledor the speed)

Heat transfer Keeping drinks warm (the effect of the material of a cup on the rate of heat loss)

Energy transfer Fuels (find out which is the best fuel)

Table 4.3 An example of varying the concept difficulty in a forces and motion investigation (type 2 tasks)

The concept( sunderlying the variables

Investigation

Distance only

Speed and length

Speed and energy

Find out how the amount of stretch (of an elastic band) affects the distance travelled(by a model car in practice, a margarine tub)

Find out how the speed of your model depends on how much you stretch the elastic band

Find out how the speed of your model depends on the amount of energy stored in theelastic band


Table 4.4 An example of varying procedural complexity within one concept area

FORCES the flexibility of materials

Type I

Type 2

Type 3

Type 4

Find out whether the amount by which the bridge sags depends on the type of material used

Find out how the length of the bridge affects how much it sags

Find out whether the amount by which the bridge sags depends on:th, type of the material, orthe width of the material, orboth of these things

Find out how the sag of the bridge depends on:the weight of the person. andthe length of the plank

Table 4.5 Varying the procedural complexity across different concept areas

Proceduralconzplexity

Concept area

Electricity.fromchemicalreactions

Forces theflexibilityof materials

Dissolving Forces andmotion

Heat transkr Energytranskr

Type IType 2Type 3Type 4

the four different types of investigations. Table4.5 shows how the investigations were spreadacross type and concept area.

Age

The design of the sample (Table 4.1) covered pupilsin years 7, 8 and 9 (ages 11 -13). The total sample ofsome 2200 pupils comprised some 700 groups.Although investigations were carried out by groups.individuals within these groups were asked to writeup their investigation independently.

Context

Three pairs of investigations in the concept areasof dissolving and heat transfer were varied in thatone of the pair was set in an 'everyday' contextand the other in a 'scientific' context. Some ex-amples are shown in Table 4.6.

Table 4.6 Varying the context (types I and 2 only)

Scientific 'Everyday'

Find out which chemicaldissolves fastest. (type I)

Find out how thetemperature of the wateraffects how quicklysodium hydrogencarbonate dissolves.(type 2)

Find out which type ofsugar dissolves fastest.

How does the temperatureof the water affect howquickly the sugardissolves?

Openness

Finally, two pairs of investigations were varied inthe degree of 'openness' (Table 4.7). Openness hererefers to the way in which the task is presented.


Table 4.7 The openness of tasks (type I only)

( More) open ( More) closed

Which fuel is best?

Which cup is best?

Which fuel gives out the mostheat?

Which cup would keep a drinkhot longest?

Methodology

Each school taking part in the NCC research wasasked to carry out the investigations duringnormal science lessons. Following a half-daytraining session, teachers were asked to completeparts of an observation checklist while the pupilswere carrying out the investigations, some ele-ments of which are given in Table 4.8.

Basic equipment for each task was provided cen-trally but schools were asked to provide otherequipment, if requested by the pupils. The final dataconsisted of pupils' individual accounts of theirinvestigation together with an observation checklistfor each group completed by the class teacher.

55

The analysis

The pupils' scripts were compared with the teach-ers' checklists and any suspect data in terms ofinternal consistency discarded (approximately 5per cent). Scores which were derived from theteachers' checklist data were validated by theresearcher against pupil scripts which had beenanalysed independently. Additional data derivedfrom pupil scripts were added to the computerdata set.

The data were analysed using a 'task score',which is a summative score based or. all the indi-vidual elements of the investigation reflecting thepupils' overall performance. It was calculated bysimply adding up the ticks in the boxes on thechecklist. For the purposes of the analysis, the taskscore was broken down into three parts: the vari-able score, the data score and the interpretationscore. The variable score was based on the check-points relating to the identification of the variablesin the task. The data score also consisted of severalcheckpoints relating to measurement and represen-tation nf the data, while the interpretation score (a0-1 sco- e) was based on an overview of the whole

Table 4.8 Examples of items in the teacher's observation checklist

Variable types NI or x Comment

Independent variable I defined as categoricIndependent variable I defined as continuousDependent variable defined as categoricDependent variable defined as continuous

Putting variables into practiceIndependent variable put into practice effectivelyAppropriate variables controlled

Measurement

Scale adequate

Accuracy appropriateRecording of results

Data handlingData recorded in tableData recorded in bar chartData recorded in line graph


task and reflected whether or not the pupil wasable to make an appropriate interpretation andgeneralisation from his or her data.

During the analysis, it became clear that theoverall task score behaved in a very similar way totwo of the part scores, the variable and datascores, but frequently quite differently to the inter-pretation score. The results that are presentedhere, therefore, will only refer to the task andinterpretation scores. For a more detailed discus-sion, the reader is referred to the c riginal researchreport (Foulds et al., 1992). Suffice it to say herethat the scores were analysed using analysis ofvariance and that differences are significant at the0.001 level unless otherwise stated.

The resulting data will be used extensively inthis and the following chapter, together with datafrom other research findings. We shall continue byconsidering the findings concerning each of thefactors in Fig. 4.1 in turn.

The effect of substantive concepts

What effects do substantive concepts have on per-formance? Both the APU and the NCC projectssuggest, not surprisingly perhaps, that substantiveconcepts strongly affect performance. The APUteam distinguishes between 'everyday' concepts(that is, commonly known concepts) and 'taughtscience concepts'. Although the investigationsused in their research were those requiring little inthe way of taught science concepts, they still notedtheir strong influence:

The results have shown that it is hazardous toattempt to generalise about children's performancein the various investigations. The particular sub-ject matter of a problem has a very stronginfluence on performance, introducing a numberof variables whose influence cannot easily be dis-entangled. This is borne out both by the detailedaccounts of performance in an investigation as awhole,...and the analysis of results in terms of' vari-ous component parts of the investigations.

(Russell et al., 1988.)

J

In the NCC project, the data were analysed toexamine the effect of concept area alone by averag-ing performance across all tasks within each of thefour principal concept areas in which the majorityof the data were collected. Figure 4.2 shows howperformance, in terms of the task and interpreta-tion scores, varied between the four concept areas.

Both the task and interpretation scoresrevealed significant differences in performancebetween the different concept areas, suggestingthat the underlying concept has a strong influenceon performance. Of these four concept areas, onewould expect electricity to be the most difficultbut neither of the scores showed this to be true.However, if we look more closely at the investiga-tions themselves, we can distinguish between theconcepts that are embedded in the investigationand the concepts which are actually essential toperforming a particular investigation successfully.For example, the investigations concerning elec-tricity (which are described in Table 3.5) do notnecessitate the critical application of any concept

70

60

50

40

30

20

10 r111 Task score El Interpretation score

CC 0CO V.

Fig. 4.2 Task and interpretation scores for four conceptareas


of electricity for successful performance. Childrenin the sample interpreted their results adequatelywithout reference to the underlying concept. If anexplanation had been required, then it is likelythat the children would have found it more diffi-cult. Looking at the interpretation score, bothtasks involving forces orov d to be more difficultto interpret than the other two concept areas(electricity and dissolving). We suggest a reasonfor the differences between the two forces andmotion tasks in the next chapter. What we mightmention here is that, rather than the conceptsthemselves, it is the associated context the appa-ratus and its familiarity perhaps which is themore significant factor.

The effect of the level of difficulty of the concept

The example in Table 4.9 shows that as the diffi-culty within a concept area increases, in this casefrom 'distance' to 'speed and energy', so the taskscore decreases slightly. In contrast, the interpreta-tion score rose slightly, from a low value of some30 per cent.

This seems to make no sense at all. Energy andspeed are clearly more difficult ideas than simplymeasuring how far the car travelled. So whyshould there be so little difference, a differencewhich barely reaches statistical significance? Whatthe data point to is that the effects of concepts onperformance are by no means as straightforwardas perhaps might be imagined. One part of theexplanation may lie in the constraints of the taskitself. Given the apparatus available, there arevery few alternative sets of variables to measure,

Table 43 Task scores for increasing conceptualdifficulty in the forces and motion investigations

The concept( s ) underlying Thsk scorethe variables (%)

Distance

Speed and length

Speed and energy

68

66

62

57

although there are different ways of proceedingfrom there on. Pupils then, presumably, identifythat the distance the elastic band launcher ispulled back will be related to energy, no matterhow tenuous their grip on the concept of energymight be. A similar argument applies to speedwith the exception that very few pupils indeedactually calculated a speed, usually being contentwith the component parts of distance and time.

The effect of procedural complexity

The APU found that where an investigationinvolves one independent and one dependentvariable, most pupils werr able to design theinvestigation successfully. However, when twoindependent variables are involved, as in thewoodlice investigation (Table 3.2) where damp-ness and light are the independent variables, thepercentage of pupils able to handle the interac-tion fell markedly, in this particular case to 43 percent (Archenhold et al., 1988). Of these pupils,21 per cent manipulated all four environmentstogether, while 22 per cent set up one combina-tion at a time. In some cases, four separate littleenvironments were created, separated by a con-siderable distance equivalent to many 'woodlicelengths'. A number of woodlice would then beplaced in the middle of the four environmentsand expected to decide from afar, without beingallowed time to wander from one to the other,where they would like to go. When the four en-vironments were tested cne at a time, the pupilswere forced to rely on some independent measureof woodlice contentment for example, one pupildecided that happiness in woodlice was indicatedwhen 'they lay on their backs and wriggled'.

In this case, the problem of two independentvariables was compounded by problems of decid-ing how to measure 'happiness' as well as notionsof animal variation. In the swingboard investiga-tion (Gott and Murphy, 1987) where there was nosuch complication, of the pupils who were askedto investigate the effect of length and width on therate of the swing, 44 per cent still failed to test ade-quately both independent variables.


We saw in the last chapter how the NCC pro-ject divided procedural complexity into four typesof investigations (see Table 3.5) in terms of thenumber and type of independent variablesinvolved. When the data are analysed by investi-gation type disregarding age, concept area andcontext, the pattern that emerges is different forthe task scores and interpretation scores (Fig. 4.3).

If we consider the task scores first, there aretwo underlying trends. First, there is the failure ofsome pupils to identify independent variables ascontinuous, which makes type 2 and type 4 inves-tigations appear to be more difficult than types 1and 3. Superimposed on this pattern is a gradualdeterioration in performance, as measured by thetask score, with task type. This deterioration canbe attributed to a more general factor, the overallcomplexity of the task, represented by the num-bers of independent variables involved.

Turning to the interpretation score, the patternhere shows a gradual decline in percentage scorewith task type. The difference in the behaviour ofthese two scores can be explained by consideringwhat the two scores represent. The task score is asummation of a disparate set of actions. If apupil makes an error of task definition, for

70

60

50

40

30

20

10

2 3

Investigation type

111 Task score Ei Interpretation score

Fig. 4.3 Scores by investigation type

4

53

instance, then this will reduce his or her overallscore somewhat. But since the majority of theindividual elements are to do with the carryingout of the investigation and recording the data,which still apply even if the task has beenwrongly defined, then the fall in the overall scoreis small. The task score, then, represents the abil-ity of pupils to perform elements of the tasksuccessfully. The interpretation score, on theother hand, relies more on the ability to synthe-sise the key elements of the task and to presentfindings which relate to the original task as set,rather than as subsequently defined by the pupil.We will return to this issue in the context ofassessment in a later chapter.

The effect of age

At the time of the APU survey, overall pupil per-formance on investigations between ages 13 and15 was not noticeably different. Among the differ-ences that were observed was a greater tendencyamong 13-year-olds to control all possible controlvariables whether or not they were relevant. More15-year-olds used tables to record their data andthey also tended to revise the design of their inves-tigation more often than the younger pupils. Inthe 'survival' investigation (Which fabric wouldkeep you warmer?), more 15-year-olds (56 percent) than 13-year-olds (44 per cent) measuredboth the initial and the final temperatures,but this sort of improvement was attributed toincreased conceptual understanding rather thaninvestigatory skill (Archenhold et W., 1988). Inthe longitudinal study, where the same pupilswere tested at ages 12 and 14, there was evidenceof progression (Strang et al., 1991), although onlytwo investigations were observed.

The NCC project also found that there was pro-gression in the performance of investigations withage (years 7, 8 and 9) in terms of both the task andinterpretation scores (Fig. 4.4). The changes in theoverall task score were relatively small, certainlyless than the effect of different concept areas. Thechange in the ability to interpret and generalisefrom the data was more marked, however.


70

60

50

20

10

Y7 Y8

Age

Y9

IITask score Interpretation score

Fig. 4.4 Task and interpretation scores by age

The effect of context

Gott and Murphy (1987), in considering the effectof the everyday context of most of their investiga-tions, wrote: 'The evidence so far suggests that theproblem context does influence pupils' perform-ance but that the effect is not a simple one.' Theynoted that a scientific context can inhibit somepupils' performance and that this effect seems tobe linked to specific concept areas. For example. ifa pupil perceives electricity as a difficult topic,then he or she may transfer this perception to anyinvestigation set in this context, which in turn canaffect performance. On the other hand, the APUfound that an everyday context can lead sorricpupils to the idea that an everyday answer is allthat is required and the notion that 'we shouldonly behave scientifically when the task looks sci-entific'.

On two investigations, one set in an everydayand onc in a scientific context, pupils were morcsystematic and quantitative when performing theinvestigation set in a scientific context. Stranget at (1991). in reporting the work of the APU.suggested that the scientific context 'cues pupils

59

into working in a scientific way by remindingthem of investigations they have done previously...The importance of the context here is in providinga frame of reference for pupils which allows themto access previous experience.'

The NCC project used three investigations whichwere presented in both everyday and scientific con-texts. The performance of the children in terms ofthe task scores and interpretation stores and dis-regarding other factors are shown in Table 4.10.Both scores show that performance was betterwhen the context was scientific as oppose.; toeveryday. It may be that if children are asked tofind out about types of 'chemicals' rather thansugar in an everyday setting, it focuses them in thedirection of solubility and of 'being scientific'. Theeveryday setting almost seems to distract themfrom the science of the task and leads them tothink that non-scientific answers will suffice.Perhaps the most significant message here is thatchildren do not see that it is necessary to 'be scien-tific' in everyday situations.

It is also important to remember that the influ-ence of both the concept area and the context islinked to the gender and culture of the pupils.Johnson and Murphy (1986), in discussing thefindings of the APU, pointed to the gender differ-ences in the type of experiences pupils haveoutside school. We cannot assume that particularconcepts or contexts are familiar to all pupils.

The effect of openness

The APU tasks did not include open questionssuch as those used in the NCC project. In thelatter, two pairs of investigations were varied inthe degree of openness (see Table 4.11). The effectof openness was not significantly different in terms

Table 4.10 Everyday vs scientific contexts

Everyday Scientific

Task score 58 66

Interprvtation score 46 62

r

60

Table 4.11 Open versus closed tasks. Note that thedifferences between task scores are not significant


More open More closed

Task score

Interpretation score

61 65

44 79

of the task score from that for the more directed,closed tasks. It is encouraging to note that all ofthe children who were presented with an open taskchose appropriate variables in defining the task.Another facet of the more open tasks was that anumber of children attempted multiple tests onseveral different dependent variables, in anattempt to form some overview of suitable proper-ties. This approach may be modelled on realitysince, for example, the consumer would want toknow about more than one property to decidewhich was the 'best' fuel.

However, Table 4.11 shows that the interpreta-tion scores of the open and closed tasks weresignificantly different. Pupils were better at inter-preting the data and producing an appropriategenet alisation (which as we have argued earlier.represents their ability to synthesise various ele-ments) in closed tasks than in open ones. In opentasks, they tended to regress to qualitative com-parisons. as indeed they did in the case ofeveryday versus scientific contexts.

The relative significance of these factors

The important issue here is not that concepts, con-text or task complexity influence performance ofcourse they do. The issue in planning curricula and.particularly, assessment, is how great are theseeffects one relative to another? Examining the datain the barcharts in some detail, we can suggest ten-tatively that procedural complexity and the conceptarea have a major effect on both the task and inter-pretation scores. But openness and age also have asignificant influence on the interpretation score.

We have argued that the interpretation scorebetter represents the ability to synthesise all theelements of the task. As the complexity of the task

increases (whether it is due to more or more com-plex variables, or more difficult or unfamiliarconcepts and contexts, or having to grapple withdefining a more open task), so does the abilityneeded to hold it all together. These factors do notaffect the task score so much because, even if apupil has lost the thread, he or she can still go onand do something, even if that something has lostits direction. The message for teaching is that wemust place greater emphasis on this ability to keepthe whole task in view.

The data further suggest that relatively closedtype I tasks, set in concept areas where the ideasrevolve.around familiar ideas such as length and,furthermore, in scientific contexts, are likely to be agood starting point for curriculum planning. Fromthere on, progression in the type of tasks presentedto pupils will need to be tightly monitored toensure that too many factors are not changed tooquickly, so halting pupils' development.

If we analyse the same data using multiplelinear regression, then we find that concept area,age and procedural complexity explain only about10 per cent of the variation. Other factors such asmotivation or pupil expectations will have a majorinfluence. On the other hand, given that the pupilsor schools were not matched in any way and thatinvestigations are a very complex activity, it is notsurprising that there is a lot of 'noise' (or randomeffects that mask trends) in the data. But it is stilltrue that the effect of the underlying concept is avery significant one. on both scores, and this find-ing has considerable implications when makingdecisions about the concept demands of tasksintended for the assessment of procedural under-standing and the number of different conceptareas that must be covered, if we are to get a reli-able and valid 'handle' on pupil ability.

.60

Motivation, expectations and perceptions

The first and most important point to note is thatpupils do remarkably well in investigations. Theyvery rarely fail to carry out the task in some way.It is this very success, we suggest. that is respons-ible for the 'high motivation among pupils doing

nr-


investigations, that was reported by many schoolsengaged in the NCC research. Watts (1991) hasalso commented on the sense of empowermentand ownership that this kind of problem-solvinggenerates.

In 1981, the APU team (Harlen et al., 1981)asked their assessors to rate the motivation of11-year-old pupils as they undertook investiga-tions. The results across six investigations areshown in Table 4.12. Apparently, even underassessment conditions and at a time when mostpupils had had no experience of investigativework, almost half the pupils found investigationsto be both interesting and enjoyable. When anassessor was asked whether pupils cooperatedand were interested in the investigations, thereply was: 'Without exception. Once embarkedupon tasks they became really involved and mostwere determined to solve the problem no matterhow long it took' (Archenhold et al., 1988). Thedegree to which pupils feel in control of their ownlearning is a significant factor here, but one inwhich there has been little research, particularlyin practical science.

Simon and Jones (1992) discuss several factorsthat can affect pupil motivation and so influenceperformance. These factors include 'learningexpectations'. that is, what the pupil expects tolearn in a science practical. For instance, if theemphasis in previous practicals has always beenon facts and concepts, pupils may miss the pointof a practical which is designed to improve proce-dural understanding and may even consider it tobe a pointless exercise and therefore nct performwell. 'Expectation of completion of the task' refersto what the pupil thinks the teacher wants. Forinstance, he or she may associate satisfactory corn-

Table 4.12 Motivation (Har len et al.. 1981)

Category Percentage of pupils

Evidence of real interest 47

Willing but no great enthusiasm 48

Uninterested 4

61

pletion of the task with being busy 'doing' or writ-ing copiously rather than reflecting and thinkingabout the nature and purpose of the task.Performance may also be affected if pupils believethat they know 'the right answer' and see this as away of obtaining good marks. They may. thenwrite a convincing report based on previous ideasignoring their own data, whether or not the dataagree with their prediction of what the rightanswer should be and regardless of the teacher'sguidance. Again we have recently seen evidence inthe UK that some pupils are purposely gearingtheir work to achieve particular assessment goals.

A pupil's perception of his or her own ability tolearn in science practicals is also likely to influenceperformance. If he or she has experienced failurein the past, then subsequent practicals are likely tobe approached negatively, expecting to fail again.Such children will avoid any challenge and tend togive up easily. This motivational style is known as'learned helplessness'. At the other extreme arechildren who perceive difficult tasks as challengingrather than threatening.

Perhaps the most important point here, and oneeasily overlooked in the search for complex reasonsfor high motivation, is that we are all motivated bysuccess. No matter what it is we are asked to do, ifwe succeed, and continue to do so especially whenthe going gets tough, then we tend to persist. If theopposite is the case, then confidence plummets.There is nothing motivating about being told thatsomething you cannot do would be good for you ifyou could do it. Investigations, because of their veryopenness, allow all pupils to feel that they are suc-cessful to some degree, as indeed they are. So in themidst of our search for progression, let us notforget that making things too hard, too soon, is inno one's interest.

Finally, the peer interaction that occurs ingroup work influences performance. Collaborauvcskills can enhance learning, providing support forthe less confident pupils. Occasionally, of course,group work can 'go wrong' and it is here that therole of the teacher is again crucial. While werecognise the importance of all these factors, in themain they are outside the control of the teacher.

Ci.

INVESTIGATIVE NN ORK IN THE SCIENCE CURRICULUM

Teachers' perceptions

When we consider the factors that influence pupilperformance in investigations, it seems sensible toassume that both the teacher and teaching have aneffect. But does the way investigations are taught,which depends on the teacher's views as to whatthey are for, make any difference? Sharp v.nd Green(1975) noted the significance of teachers' beliefs asto the role(s) that they adopt and how these beliefsinfluence their practice. For instance, a teacher whobelieves that investigations are about 'allowing chil-dren to discover things for themselves', may adoptan extreme non-interventionist role acting only as amanager and provider of resources. What do weknow about teachers' views?

The research data are sparse in this regard. TheNCC project included a questionnaire which wasdesigned to probe teachers' understanding of ATIin the general context of practical work in science.The questionnaire was completed by 290 teachers,123 of whom taught at the secondary level, 85 atthe junior level and 82 at the infant level. Theteachers were also asked to submit sample mater-ial of the type they use for an investigation. Asample of 50 teachers was interviewed shortly afterthe questionnaire had been sent out and the result-ing data used to validate the questionnaire.

Teachers were asked to order five possible aimsfor practical work (Table 4.13). The data showedthat, at the infant teacher level, the emphasis wasquite clearly on the idea of observation as the key

aim of science. Junior teachers see a move towardsconcept understanding, whether through illustra-tion or enquiry, a move which is furtheremphasised at secondary level when observationhas fallen very much lower down the list.However, when teachers were asked about thepurpose of investigative work, the results showed asomewhat confused picture (Table 4.14). Therewas a general move towards 'processes and skills',that ubiquitous and ill-defined phrase that wecame across in an earlier chapter.

The most telling information came from thesamples of investigations that had been submitted.The samples were analysed by a process of group-ing the material into categories with obviouscharacteristics in common. Only then was adescriptive title given to each of the groupings, thetitles being based on the definitions we haveadopted in this book. The major categories areshown in Table 4.15.

What we see now is that infant teachers, and toa lesser extent teachers of juniors. based much oftheir work on simple observation activities despitea claim for increased emphasis on 'processes andskills'. At secondary level, the samples weremainly 'guided investigations' or investigationswith a 'recipe' provided. What is certain is thatthere was little obvious correlation between theprofessed aims and the samples. The interpreta-tion of this questionnaire was problematic due tothe difficulty of ensuring that all involved wereusing the terms in the same way. As a conse-

Table 4.13 Teachers' perceptions of the aims ofpractical work

Table 4.14 Teachers' perceptions of the purpose ofinvestigative work

Major stated purpose ol Year I Year 3 Year 7Most important aim of Year I Year 3 Year 7practical work (%) (%) investigative work t".."4 ("i,)

Concept illustration/consolidation 23 29 39 Skills or processes 19 29 29

Raising questions and devising Concept discovery (enquiry) 22 20 18

solutions 27 35 31 Use or refine concepts 9 12 18

Observation 52 29 10 Raise questions or test ideas 15 16 14

Concept discovery (enquiry) 9 Enjoyment or motivation 10 11 7

Skills I () 11 Observation 3 2 1

6


Table 4.15 Types of activity used by teachers as'investigations'

Type of activity Year I Year 3 Year 7(%) (%) r!-;,

Skills development(measuring, manipulating)

Design and build

Illustrative experiment

Guided investigation':teacher directed

Observation and recording

Investigation including: freeaccess to apparatus, externalreference, minimal instructions.choice of recording format 1 1

1 1 8

11 11

1 1 7

3 12 19

53 36 5

15

quence. it is probably only safe to say that theword investigation' was being used to describe arange of practical work. There seemed little in theway of a common philosophy as to how teachersperceived 'investieations'.

Evidence from HMI surveys

In the UK. there have been annual inspections of asample of primary and secondary schools whichhave examined science in schools since the intro-duction of the National Curriculum in 1989. TheUK National Curriculum is based on an assess-ment system defined by ten levels within fourattainment targets (ATs). The.attainment targetsare centred loosely around investigations (Scl), lifl .and living processes (Sc2). materials and their prop-erties (Sc3) and physical processes (Sc4). These fourattainment targets are assessed individually. Eachof the ten levels is in turn defined by more preciseobjectives or Statements of Attainment.

The first annual inspection during the schoolyear 1989-90 found that: 'Too much of the Year7 work either insufficiently linked Scl to theother Attainment Targets or was narrowlyfocused on individual Statements of Attainment'(HMI, 1991). The second report (during 1990-1)

63

found that in secondary schools investigative workwas being blocked at a low level: 'In much of thework pupils were given insufficient opportunity todevelop higher skills of hypothesising, designinginvestigations or interpreting evidence. Pupilachievement in ATI was thus being blocked atlevel 4 in many schools' (HMI, 1992). The thirdand most recent report (during 1991-2) found thesituation relatively unchanged in middle and sec-ondary schools:

Investigative work had increased and pupils werebecoming more adept in planning and carryingout simple investigations. However their ability totackle more complex investigations, to hypothe-sise. identify variables and evaluate the outcomesof investigations. remained weak. As a result.work above level 4 in Scl continued to be rare.

For pupils of all abilities there was a lack ofchallenge and under-expectation in work relatedto Scl.

(HMI. 1993)

It would seem, therefore, that there has been agradual increase in the proportion of investigativework being carried out in secondary schools butthat teaching of the more complex investigations isinfrequent. It is also apparent that opportunitiesfor pupils to focus on the nature of evidence ininvestigations are few. The impression from read-ing these reports is one of teachers engaging withindividual skills and concepts of evidence, but notwith their synthesis.

The above is consonant with the NCC findingsthat many teachers were confused as to the role andpurpose of investigations. We should note, how-ever. that the questionnaire was carried out shortlyafter the formal introduction of investigations intothe curriculum in the UK. The HMI reports sug-gest that the situation may be improving.

A note on the effect of teaching about investigationsand concepts of evidence

Of the schools which took part in the NCCresearch, a small number were actively working inthe area of investigations, developing some of the


teaching ideas and schemes of work which we putforward in Chapters 6 and 7. The remainder werenot very far along the road at all. It is of interest,then, to compare how such schools fared. Becausethe investigations were spread out. with no schooldoing all of them, it is not possible to make adireacomparison based on any one investigation.What can be done, however, is to reduce the taskscores (the overall pattern for interpretation scoresbeing very similar) to *z-scores'. This techniqueturns a percentage score on a particular task intoone based on the mean for that task over all thegroups which carried it out. It is now possible tocompare one investigation with another, in rela-tive terms if not absolutely.

So which schools did well? The school whichwas furthest ahead in its development of inves-tigative work came out top of the list. Otherssimilarly advanced, were all near the top. Butwhat about the differences in the ability level ofthe schools' intake? It might be that the schoolsat the top of the list are the ones with the mostable pupils anyway. To test this, the best that canbe done with the data as they stand, is to referback to league tabies of GCSE results (which areexaminations taken at the end of compulsoryschool). We can use these as some sort of, admit-tedly doubtful, measure of the overall ability levelof a school's catchment.

If we aggregate investigation data for a schooland compare it with its GCSE sco:-es, the patternin Fig. 4.5 emerges. Clearly, there is little, if any,

3

2.5

2

8 1.5

N 1

0.5

0o

School effect

sip

20 40

% pupils with 5 or more passes at GCSE

Fig. 4.5 Performance and GCSE passes for individualschools

60

correlation between performance in investiga-tions and this crude measure of ability. Theschool with the highest score on investigations isalmost at the bottom of the league table ofGCSEs, while the school at the top of the GCSEleague is almost at the bottom of the investiga-tion score. What this hints at, and it can be littlemore than a hint, is that performance on investi-gations does not rise and fall with overallattainment by educational osmosis. Rather. itsuggests that investigative ideas can be taughtand taught successfully. And, equally import-antly from the point of view of motivation. pupilswho are not likely to do very well at GCSE, areable to succeed in investigative work. Of course.it could be the Hawthorne effect operating - theschool developing investigations may simply bemore enthusiastic, which is transmitted to thepupils. We do not know. Only more carefulresearch planned with that question in mind willtell us.

G4

Summary

To recap, the research we have examined showsthat:

Of the factors tested in the NCC project, sub-stantive concepts had the strongest influence onperformance. We have discussed how this influ-ence is sometimes unexpected.Progression within investigations which arebased on variables is influenced by both thetype and number of variables.Pupil performance improves with age. The abil-ity to interpret and generalise improves morethan the overall task performance which is rela-tively little affected.Pupils do less well in investigations set in'everyday' contexts than in t:cientific contexts.Open contexts are more difficult than closedcontexts.Motivation appears to be high, although thc evi-dence to support this statement is of ananecdotal nature. Expectations and pupil per-ceptions may have a significant influence on


performance, but there is a paucity of researchin these areas.Teachers' views of investigations at the time ofthe NCC study appear to have been both diverseand confused. Although there has been a gradualincrease in the proportion of investigative workin schools, it tends to be at the lower levels.Older pupils are not being allowed the opportu-nity to tackle the more complex investigationsand their associated concepts of evidence.The effect of focusing teaching on concepts ofevidence is that pupils' performance improves.The data also suggest that children who per-form well generally in their GCSE examinationsdo not necessarily perform well in investiga-tions and vice versa.

References

Archenhold, F., Bell, J., Donnelly, J., Johnson, S. andWelford, G. (1988). Science at Age 15: A Review ofAPU Survey Findings, 1980-134. London, HMSO.

Foulds, K., Gott, R. and Feasey, R. (1992). 'Investigativework in science'. Unpublished research report.University of Durham.

Gott, R. and Murphy, P. (1987). Assessing Investigationsat Ages 13 and 15. APU Science Report for TeachersNo. 9. London, HMSO.

65

Harlen, W., Black, P. and Johnson, S. (1981). Science inSchools: Age 11. APU Science Report for TeachersNo.1. London, HMSO.

Her Majesty's Inspectorate (1991). Science Key StagesI, 2 and 3: A Report by HMI on the First Year,1989-90. London, HMSO.

Her Majesty's Inspectorate (1992). Science Key StagesI, 2 and 3: A Report by IIMI on the Second Year,1990-91. London, HMSO.

Her Majesty's Inspectorate (1993). Science Key StagesI, 2 and 3: A Report by HMI on the Third Year,1991-2. London, HMSO.

Johnson, S. and Murphy, P. (1986). Girls and Physics:Refleaions on APU Survey Findings. APU OccasionalPaper No. 4. London, HMSO.

Russell, T., Black, P., Harlen, W., Johnson, S. andPalacio, D. (1988). Science at Age 11: A Review ofAPU Survey Findings, 1980-84. London, HMSO.

Sharp, R. and Green, A. (1975). Education and SocialControl: A Study in Progressive Primary Education.London, Rout ledge and Kegan Paul.

Simon, S.A. and Jones, A.T. (1992). Open Work inScience: A Review of Existing Practice. London,King's College London.

Strang, J., Daniels, S. and Bell, J. (1991). Planning andCarrying Out Investigations. Assessment Matters No.6. London, SEAC/EMU.

Watts, M. (1991). The Science of Problem-solving: APractical Guide for Teachers. London, Heinemann/Cassell Educational.

C5

CHAPTER 5

The performance of investigations in secondaryschools: A detailed look

In this chapter, we shall take a detailed look athow procedural and conceptual understandinginfluence children's performance as they carryout an investigation. We shall start by consider-ing the skills that children need in order to doinvestigations. Then we shall identify the con-cepts of evidence that children find difficult andthe extent to which the substantive concepts ofscience influence performance at various stages.This kind of diagnostic information will allowteachers to focus their teaching on specific pointsof difficulty.

Skills

We discussed earlier how basic skills can be seento underpin procedural understanding (sec Fig.2.1). It is clear that children should have the skillsnecessary to carry out investigations, otherwisepoor performance can simply be a reflection of thelack of a particular skill (or skills). What does theresearch tell us about children's basic skills?

The APU looked at pupils' skills of measure-ment and in a small progression study carried outbetween 1987 and 1989 (Archenhold et al., 1991)found, not unexpectedly, that there were notice-able trends in making and using measurementswith age. Children were asked to:

read pre-set instruments;use measuring instruments: andestimate measurements.

The APU found that children improved with agefrom making qualitative to quantitative measure-ments, from using non-standard measures to usingstandard measuring instruments and from usingsimple to more complex instruments. Mostprogress was made in reading instruments ratherthan using them or estimating measurements.Overall, the study found that reading instrumentswas heavily dependent on the naturc of the instru-ment being used, although children often mademistakes with minor divisions in instrument scalesregardless of the instrument.

An example will serve to illustrate this point. Inthe APU survey (Welford et al., 1985), pupils wereasked to read voltmeters and ammeters which hadbeen set up in a 'circus' practical. The reading ofthe scales presented very considerable problems,problems that are probably independent of thequantity that is being measured. For the voltmeter,the scale was in single whole numbers with divi-sions of 0.1 on a range of 0 to 5 V. The scale of theammeter read from 0 to 1 A with major divisionsof 0.2 A and minor divisions of 0.02 A. Not unex-pectedly, reading the voltmeter was found to bemuch easier: 61 per cent of 15-ycar-olds read thescale on the voltmeter accurately, whereas only 11per cent did so for the ammeter. The emphasis in atask such as this is on the skills (as we have definedthem in Chapter 2) of reading the instruments. Butin this particular case, the key factor is the under-standing or lack of it of decimals in maths.

Lack of understanding concerning the substan-tive concepts of current and voltage was high-

G


lighted in another APU task where pupils weregiven a circuit which was already connected andasked to insert an ammeter and then a voltmeterto measure current and voltage, respectively. Only9 per cent of 15-year-old pupils connected bothinstruments correctly. Thirty-nine per cent chosethe voltmeter for measuring current and 17 percent chose the ammeter to measure voltage.Written questions also revealed that, 'for themajority of pupils the units for current couldalmost have been a random selection from "volts"or "amps" (Gott, 1984). In this case, the 'skill'was demonstrated: they did connect the meters.But the understanding needed to know 'which toconnect where' was missing.

The APU progression study also consideredgraphing skills by means of written tests (Figs 5.1.and 5.2). They found that the majority of pupils, bythe age of 14. had grasped the basic skills needed toconstruct a line graph. although less than half wereable to draw a line of best fit. These findings suggestthat pupils had grasped the skill of graph construc-tion but not the underlying concept of patterns.

Only 20 per cent of 12-year-olds and 71 per centof 14-year-olds actually connected the points of thegraph at all despite having plotted them accurately.

0 50 100

Frequency of response

o Drawina line of 111 Selectingbest fit conventional axes

El Plotting Selectingaccurately scales

and of those who did, the majority did so usingstraight lines. Archenhold et al. (1991) concludethat: 'It would appear that, despite some progres-sion in drawing a line of best fit, a considerablepropertion of pupils at both ages did not appreciatethe continuity of line graphs.' There was also someconfusion between bar charts or 'stick graphs' andline graphs, with many children drawing bar chartsor 'stick graphs' when they had been asked to drawa line graph. Archenhold et al. (1991) write: 'formany pupils at age 12, the required curricularprogress from "bar chart" to "line graph" construc-tion is likely to be a substantial "step forward.

We see, then, that the basic skills needed to con-struct graphs are not such a major problem,although there are indications here that while chil-dren can recall the skill of constructing a graphthey may not understand its purpose. We mustnote, however, that we are considering isolatedwrittcn tests. We shall see in the next section howchildren use graphs in whole investigations.

The overall pattern emerging is that, in sec-ondary schools, it is either substantive concepts inmaths or science, or concepts of evidence underly-ing the skills that are the major problem, ratherthan the mechanical aspects of investigative work.

0

50

Frequency of response

Labelling axes withnames of units

1111 Labelling axes withnames of variables

100

Fig. 5.1 Do ekipment or graph construction skill. Fig. 5.2 Do elopment or graph commtmication skills

6 7

THE PERFORMANCE OF INVESTIGATIONS IN SECONDARY SCHOOLS: A DETAILED LOOK


What does research tell us about how childrenunderstand the concepts of evidence as defined inChapter 2, which lie at the heart of proceduralunderstanding and which enable pupils to useskills effectively?

We can use data from the NCC project to seethe underlying pattern of children's understandingof concepts of evidence in terms of the differentstages of an investigation. The NCC projectanalysed checklist data on children's performanceusing 'criterion scores'. Briefly, the scores arecumulative in the sense that they are a measure ofhow far the pupil succeeded along successivestages of the investigation. Type 2 investigations(a single continuous independent variable), forexample, were divided into eleven discrete steps.The first few are concerned with the identificationof the correct variables and the design of a fairtest. The middle section reflects the stages in themaking of the appropriate measurements, whilethe end is concerned with the presentation andinterpretation of the data.

When these scores are plotted for all type 2investigations (Fig. 5.3). it is noticeable that there

100

90

80

70

60

50

40

30

20

10

0 Ii

Identifyingvariables

Interpretingdata

1 2 3 4 5 6 7 8 9 10 11

Criterion score

Fig. 5.3 Performance as measured h) criterion scores ont)pc 2 investigations

69

are two points where the percentage scores fallmarkedly, which in turn reflect the points at whichpupils experience the most difficulty. The firstpoint occurs at the design stage (steps 2-3), wherepupils are required to identify the type and com-plexity of the variables. Once the variables havebeen identified, pupils are then able to carry outthe investigation with some success until thesecond steep decline in the bar chart, which is atthe data handling stage (steps 9-10). where pupilsare required to represent the evidence in the formof a bar chart or line graph. It is evident that veryfew pupils are able to go on to the last stage of theinvestigation where they are required to generaliseand evaluate appropriately.

The distributions for the other task types followbroadly the same pattern, except that in thosetasks with categoric independent variables, the firstpoint of decline does not occur. With this overallpattern in mind, we shall consider performanceand procedural understanding under the four mainheadings of design. measurement. data handlingand evaluation.

Design

l'ariable identification and the fair test

Both the APU and the NCC project found thatmost children, regardless of age. were able todesign investigations so that the effect of the rele-vant independent variable could be investigatedsomehow. It is in the detail of how they did theinvestigation that the points of failure identified inthe bar chart in Fig. 5.3 can be located.

The APU found that if there was more than oneindependent variable, then performance declined.In terms of control variables, the more 'obvious'the variables, thc more likely they are to be con-trolled, especially when they arc few in number.They also found that the effect of the substantiveconcept was particularly noticeable at the begin-ning of the investigation where the problem isdefined by identifying the relevant variables. Lackof conceptual understanding caused sonie pupilsto go astray.


For example, in the NCC project, in an investi-gation where speed is the dependent and also aderived variable, identifying the correct com-ponents not unexpectedly caused problems. Pupilstended to pick one or other of the components ofspeed and even when they did measure both, theysometimes proceeded to use only one set of data,although of course, distance travelled is not a badsurrogate for speed if the variable is being used ina qualitative sense as 'faster than'. In the followingexample, Anna's group measured only the dis-tance travelled.

Anna's.group (Ail*

Does theiiieed-of_the model depend on theamount Of energy stored in the elastic band?

How far pulled hack How far tlw buttertub travelled

2 cm

4 cm

6 cm

8 cm

10 cm

54 cm

83.cm

86 cm

106.5 cm

140 cm

We found out the speed of the model depends onthe energy stored in the elastic band.

Table 5.1 shows the NCC data for the percent-age of pupils who successfully identified therelevant variables for ,ach of the types of tasks. (Itshould be noted that these data are drawn fromindividual checkpoints on the observation check-lists rather than the criterion score.)

The data show that the vast majority of pupilsunderstood the purpose of the task in that theyrecognised the appropriate independent anddependent variables. Where they did not, theeffects of 'interesting' (for some reason) bits ofapparatus sometimes caused distraction: In theexample opposite, Steven's group was distractedby the presence of slotted weights in the class-room. They decided to vary the weights in themodel, rather than the distance the elastic bandwas pulled back a completely different investiga-tion to the one they were asked to do.

The overall pattern (Table 5.1) is of a progress-ive decline in performance as the complexity of thetask increases from type 1 to type 4. A task whichis limited to one independent variable and onedependent variable causes few difficulties. Thedata suggest that pupils do not have difficultydeciding what the relevant variables are. This isperhaps not unexpected given that the tasks werelargely well defined, or closed, in terms of definingthe problem. Nevertheless, in type 4 tasks wherethere are two independent variables, pupils clearlystill have sonie difficulty in their identification, thecomplexity of the design having defeated them.This trend is continued in the data associated withthe control of variables, where 75 per cent ofpupils controlled at least two variables in type Itasks, 51 per cent in type 3 tasks and 23 per cent in

Table 5.1 Identifying independent, dependent and control variables for types of tasks

Identification Type / Tipe 2 Type 3 Type 4

Independent variable 1 identified correctly 98 94 83 71

Independent variable 2 identified correctly n/a n/a 87 67

Dependent variable identified correctly 98 93 89 92

At least 2 variables controlleo 75 73 51 23

THE PERFORMANCE OF INVESTIGATIONS IN SECONDARY SCHOOLS: A DETAILED LOOK 71

What to do.

Attach an elastic band between the stool legs. Pull back and let go.

How far pulled back How far it went . Weight inside

15 cm

15 cm

15 cm

3 m 30 cm

3 m 68 cm

3 m 29 cm

10 g

20 g

30 g

The results show how far it went and how many centimetres and metres.

Table 5.2 The identification of continuous variables

Identification Type I Type 2

Independent variable defined as continuous

Dependent variable defined as continuous

n/a 54

98 93

Type 3 Type 4

n/a 43

89 92

type 4 tasks. The complexity of the design in types3 and 4 frequently seems to mean that the relevantcontrol variables are often ignored or overlooked.This is consistent with a general overload effect.

Variable types

We saw in Fig. 5.3 that the fi' it point of difficultychildren experience in type 2 investigations is inidentifying the independent variable as continuous.The definition of 'continuous' in the data analysiscentred on the use of at least three values of theindependent variable. For the dependent variable,the definition need only rely on evidence thatpupils had made a quantitative measurement. On

UV COPY AVAILABLE

these definitions, the percentage of pupils identify-ing both the independent and dependent variablesas continuous are shown in Table 5.2.

Table 5.2 shows that a significant number failedto identify the independent variable as continuouswhen appropriate in type 2 and type 4 investiga-tions, although they nearly all did so for thedependent variables. An example of this failurecan bc illustrated by considering the task: 'Findout how the temperature of the water affects howquickly ti.e sugar dissolves.' Pupils would oftenchoose to use 'hot' and 'cold' water, sometimeswithout measuring the temperature at all. Clearly,this categoric interpretation of the independentvariable limits the extent to which the relationship


between the independent and dependent variablescan be explored.

To examine this point more closely, the percent-age of pupils who identified the independentvariable as continuous was broken down into spe-cific tasks so that the concept associatedspecifically with the independent variable could beidentified (Table 5.3).

Two issues emerge from these data. First, Table5.3 shows that distance is an 'easier' substantiveconcept to identify as a continuous variable thantemperature. This might be expected in view of thefact that children are more familiar with the skillsof using rulers and tape measures and the lan-guage of length and height than they are withusing thermometers and degrees. Another contrib-utory pragmatic factor may be that length orheight is simply easier and quicker to measure,while measuring temperature takes longer. Pupilsmay therefore decide to take the easy option andidentify temperature as categoric hot and coldrather than measuring, if they think this is suffi-cient to answer the question.

Second, while the concept of distance was usedin several tasks, it was particularly noticeable thatwhere it referred to the pulling back of an elasticband to launch a model car (the forces andmotion task), children found defining this sort oflength as a continuous variable much easier thandistance in the other investigations (Table 5.3).This may be explained by the fact that the otherinvestigations involved a more 'static' kind oflength, such as the length of a plank. By contrast,in the forces and motion task, pupils physically

pulled back the elastic band so that they were incontrol of its change and could see it 'in action' asa continuous variable.

If we turn to the nature of the dependent vari-able, however, the picture changes. Table 5.4shows how the percentage of pupils identifyingthe nature of the dependent variable as continu-ous is high for all the substantive concepts in thesample. Temperature is noticeably easier forpupils to define as continuous than it was as anindependent variable, although it is still lowerthan the other quantities. The reason for thischanged picture could be related to the intrinsicrole of the dependent variable. Since the depen-dent variable responds 'automatically' to thevalue of the independent variable, there is noneed or the pupil to actively choose values.Pupils therefore do not need positively to applythe idea of a continuous variable.

The 'fair test'

We noted in Table 5.1 that the ability to controlvariables decreased as the complexity of the taskrose. But there are other factors which influencethe control element of' a fair test. The more 'obvi-ous' the variable, the more likely it is to becontrolled, even to the extent that everything insight is controlled as a ritual, rather than in athoughtful way. Some variables to be controlledare not at all visible. For instance, in the heat task,

Table 5.4 The effect of different substantive conceptson identifying a dependent variable as continuous

Table 5.3 The effect of different substantive concepts Dependent Percentage of*pupilson identifying an independent variable as continuous variable identifying variable

as continuousIndependent Percentage qf pupils

Distance (in all investigation) 77variahle identifying variableas continuous

Temperature 78

Temperature 41 Time 81

Distance (in all investigations) 60 Voltage 88

Distance (in forces and motion Distance (in forces and motiontasks only) 74 tasks only) 91

.4 AI 1


pupils have to measure the rate of loss of heat inthe two containers whose insulation they are com-paring. That comparison should, theoretically, bedone at exactly the same temperature, since therate of loss of heat depends on the instantaneoustemperature difference between the contents of thecontainer and its surroundings. Very few pupilsseem to have controlled this temperature as a con-scious act.

The evidence suggests that the notion of a 'fairtest' is well established by the time children reachsecondary school; it certainly merits a good dealof attention in primary science. But it is doubtfulif the connection between a fair test and the valid-ity of the resulting data is well understood. It is injust such instances that the importance of the syn-thesis of concepts of evidence becomes apparent.

Sample size and variation

The APU survey included an investigation aboutwoodlice, described in Table 3.2. In designing theinvestigation, pupils have to decide how manywoodlice to use in each trial or over a number oftrials. Over half the pupils at age 15 ust.d morethan five woodlice in each trial, about a third usedbetween two and five and less than a tenth usedone woodlouse at a time (Driver et al., 1984). Itappears, therefore, that at the age of 15, choosingan appropriate sample size is still a problem for asignificant number of pupils.

In summary, the research points to the conclu-sion that pupils experience little difficulty in themost basic part of the design of investigations, butthat they do have difficulty in recognising theadvantages of interpreting the independent vari-able as continuous and in understanding theconsequences of this decision for the investigationas a whole. It is here that the sudden drop in per-formance (on the criterion score bar chart: Fig.5.3) manifests itself. Choosing an appropriatesample size is also a problem for many pupils.

It seems that a number of individual conceptsof evidence, particularly the notion of a continu-ous variable, are causing problems which relate inthis example to the difficulties noted in the earliersection about pupils' use of measuring instruments

73

and graphs. If the very idea that a quantity cantake on any value and that this value is connectedto how that quantity behaves in the real world isnot well understood, then the advantages of quan-titative data will not be apparent. Over and abovethese individual factors, any increase in complex-ity of an investigation lowers performance aspupils start to lose track of the whole task.

Substantive concepts intrude most obviouslyand directly in the ability to define the appropriatedependent variable and to recognise that a vari-able must be controlled.

Measurement

Concepts of measurement refer not to the skill ofmeasurement itself but to the decisions that have tobe made concerning measurement. These includedecisions about what instrument to choose (themost appropriate forcemeter for the task in hand,for instance), over what range and interval, whenand how often to measure, and consideration of theneed to repeat measurements.

The results of the NCC research tell us thatchildren in primary schools (Key Stages 1 and 2)seem generally reluctant to measure, even thoughthey are capable of using measuring instruments.At these key stages, only 30 per cent of childrenused measurement in investigations (Foulds et al.,1992). Even at secondary school level, a significantnumber of pupils continued to judge changes inqualitative terms (34 per cent at age 11 and 13 percent at age 13). In discussing the issue of measure-ment, Foulds et al. write (in relation to a sampleof primary school children):

It would seem that the use of measurement exceptwhere the question might explicitly specify theneed is arbitrary at best, non-existent at worst....We can only conclude that this a: pect of inves-tigative work is one which has not teen grasped,for whatever reasons...

In an APU 'circus' practical (where pupils visitdifferent stations to perform a variety of tasks),children were asked to use measurement whenmaking observations, relationships and predictions.

4.0


Time (mM). Bronze tin Tin can Polystyrene Plastic Plastic beaker Glass B

88°C 88 88 88 88 88

78 78 86 84 82 82

75 76 84 80.5 80 79

3 74 74.5 81.5 77.5 76.5 76

4 71 71 79 75 75 73

5 69 67.5 77.5 73 73.5 71

Temp. wentdown by 19°C 20.5 10.5 15 14.5 17

At age 12, only one-fifth used measurement volun-tarily except in a task (toy cars) in which they wereheavily cued to do so (Archenhold et al., 1989). TheAPU research also suggests that children's abilityto handle measurements of length and temperaturediffer (Dickson et al., 1984; Archenhold et al., 1991;Foulds et aL, 1992) temperature being the moredifficult, as we noted above.

The example above shows an inappropriate use ofthe measurement of time and temperature in a heattransfer investigation. Investigations concerning heatloss seem to trigger a particular response in somepupils, who associate this kind of investigation withcooling curves regardless of whether or not this isappropriate. Perhaps they have been impressed bythe unusual boredom of waiting for something tocool down, tl7e laboratory equivalent of watchingpaint dry. Suzanne's table above illustrates the prob-lem. Only the first and last measurements oftemperature (shown in bold print in the table) arenecessary to answer the question. The notion of scaleis also relevant here in that five minutes is clearly notlong enough to give any sort of reliable results.

Foulds et al. (1992) analysed the investigations

in the NCC project in terms of whether childrenhad used an appropriate scale. They defined scaleas the carrying out of investigations 'in a mannerwhich approximates sensible conditions and quan-tities'. The results show a clear progression withage and experience (Table 5.5).

In terms of accuracy, Foxman et al. (1990),reporting on the APU findings in mathematics,concluded that in performing measuring tasksaccuracy may be associated in part with the pupil'sperception of the 'expected' accuracy. There wa3 aconsiderabie change between the ages of 11 and 15years in the accuracy with which pupils measured astraight line. At age 11, 61 per cent measured a line

Table 5.5 Percentage of pupils using scale and accuracyappropriately (Foulds et at, 1992)

Scale Accuracy

Age 11 (Year 7) 58 76

Age 12 (Year 8) 69 73

Age 13 (Year 9) 79 78


of 130 mm to within 1 mm. At age_15, when pre-sented with a similar task, 94 per cent measuredthe line to within 1 mm. In terms of science, theconcept of appropriate accuracy is fundamental.There is, for instance, rarely a need to measuretime to three decimal places, although pupils withdigital watches often do so (constrained by thenumbers on the watch or not familiar with the ideaof rounding decimals perhaps?).

In terms of the range of readings, Strang (1990),in reporting on an APU task in which childrenwere asked to find out how quickly the watercoming out of the spout ef a tea urn depends onthe level of the water in the urn, found that almosthalf (42 per cent) of the 13-year-olds who did thetask did not appreciate the importance of using anadequate range of levels.

Only a very small percentage of children repeatmeasurements in investigations. In the forces andmotion investigation in the NCC project (seeTable 4.3), where repeating measurements couldvery easily have been carried out and where it wasclearly necessary, only six out of a sample of morethan 250 children did so.

All these points rely, ultimately, on the pupils'ability to appreciate the idea of 'believable' data.It is only with this goal in mind, that the need forappropriate accuracy and range and appropriaterepeats of measurements have any meaning overand above the algorithms of 'as accurate as poss-ible', and 'always repeat things three times andaverage'. Unless and until pupils gain an under-standing of validity and reliability of the evidence,they will, quite understandably, regard these issuesas just 'one of those strange things that you do inscience. Indeed, it is only in the context of thewhole task that these particular concepts of evi-dence can be seen to have meaning.

Given those reservations, most pupils, as wesaw in the criterion score bar chart, make a rea-sonable attempt at taking measurements. It is onlywhen we stand back and look at those measure-ments in the light of how 'believable' they arewhen interpreted, that we see that there is a realproblem underneath the apparent success. Datahandling is the subject of the next section.

Data handling

75

Once children have taken some measurements,how do they then go on to use them in the investi-gation? In terms of recording data, the APU foundthat at the age of 11, children's recording of datawas frequently disorganised and descriptive withvery little use of tables, even though their work inother categories showed that they were able toconstruct tables (Russell et al., 1988). In the NCCproject, an average of 68 per cent of all secondarypupils used tables to record their data.

The example overleaf taken from the NCC datashows a good use of tabulation in the 'fruits'investigation. David's conclusion, however, is avery parsimonious affair. This is a recurring fea-ture of pupils' work; good basic presentation ofdata and inadequate interpretation.

It seems likely that pupils like David, with lim-ited experience of investigative work, may nothave been taught what is expected of them. Theymay well see the pattern in their results but notrecord it. The only clue that a fuller conclusion isrequired is the word 'how' in the question and thiscan easily be missed. David's conclusion may notreflect a lack of ability but a lack of appreciationof what is required.

If we consider the type 1 investigations in theNCC project (those with a single categoric inde-pendent variable), 72 per cent of children doingthese investigations used a table. Although a barchart is the most appropriate form of representa-tion for type 1 investigations, it serves only as aform of display so it is not essential for the pur-pose of interpretation or generalisation. So,viewed in this way, it could he said that type Iinvestigations can be adequately represented by atable. Indeed, Table 5.6 shows that only 11 percent of children drew a bar chart, with 5 per centdrawing a line graph.

Looking back at the criterion scores in Fig. 5.3for type 2 investigations (those with a single con-tinuous independent variable), we noted that thesecond significant point of difficulty is where thechildren who have progressed successfully in theinvestigation, arc required to represent their data


David (Aged 13). .

. FiaalOUthoar the distance tietween the meal electrodes affects the voltage using different fruits.

We took an orange and wired it up to the voltage meter. We done this three time at different places on theorange. We recorded the results. e.g. the metal electrodes were 1 cm apart. the voltage was 6. We done thiswith the lemon and apple.

Fruit cm apart Voltage

apple

apple

apple

3.7 cm

5.5 cm

6.4 cm

4.5

4.0

3.9

Fruit cm apart Voltage

lemon

lemon

lemon

1.7 cm

3.1 cm

3.8 cm

6.1

5.6

5.4

Fruit

orange

orange

orange

Mr apart

1.7 cm

3.8 cm

3.1 cm

Voltage

1.1

1.8

1 .9

Yes the distance between metal electrodes does affect the voltage.

graphically before proceeding to use their evidenceto support a final conclusion or generalisation. Intype 2 investigations, a line graph is the mostappropriate form of representation. Ideally, suffi-cient data should be collected to show any patternbetween the independent and dependent variables.

The representation here serves a much moreimportant role than in type 1 investigations, inthat while it is still a form or display, it can alsoreveal the nature of the pattern or the relationshipbetween the variables. This is particularly the casewith complex data or data where the line 'flattens


Table 5.6 Percentage scores for each investigation type

Type 1 Type 2 Type 3 Type 4

Bar chart 11 18 4 4

Line graph 5 4 3 9

off', but even with linear relationships a line graphcan serve a predictive function.

The APU found that very few 11-year-olds andonly 10 per cent of 13-year-olds represented theirdata graphically, despite the fact that these childrenperformed well on constructing graphs when thiswas assessed as an isolated activity (Strang, 1990).The choice of graph type was also often a problem.

It can be seen from Table 5.6 that in the NCCproject, very few children actually chose to use aline graph for continuous data. Those that did, didnot always do so in the most appropriate way. Forinstance, more pupils used a line graph for type 1investigations, where it is not appropriate, than intype 2, where it clearly is appropriate.

An example of failure to represent data appro-priately is shown on page 78. Paul and Craigcollected data eminently suitable for a line graph.But their inappropriate choice of a bar chart wasexacerbated by a decision to treat the x-axis as aset of labels, as if the values were of no greater sig-nificance than categories such as 'red' or 'big'.They also plotted the data inaccurately. Their con-clusion was a repetition of their results with noawareness of a pattern or any attempt to gener-alise. They seem to have no understanding of thepattern representing the reality of the movement oftheir model.

Data collection and representation can takc onan air of desperation on occasions. Reasons ofspace only dictate this choice of a relativelyrestrained approach to the cooling cups problem.In the investigation which Zoey recorded (p. 79).only the temperature readings in bold type werenecessary. Zoey's group appears to have lost sightor the task they seem to have carried out meas-urements in a ritualistic way. The drawing of threeseparate graphs also seems unnecessary and posi-tively unhelpful in view of' the fact that the

77

conclusion that Zoey arrives at is incorrect andnot informed by the data.

A similar effect was found with the bendingbeams task, to the extent that numerous tables andgraphs of weight and sag appeared for nine ormore 'planks', leading to a conclusion whichreflected a failure of memory as to what the objec-tives of the investigation had been in the first place.

When it comes to interpreting the patterns ingraphs. the APU reported that most childrenfound it difficult to describe and use (read off, pre-dict) patterns in graphs presented in written tests,though this improved with age (Taylor andSwatton, 1990). Austin et a/ (1991), again report-ing on the APU findings, found that children of 12and 14 years tended to bypass the data and drawon preconceived ideas. There was a tendency toimpose patterns on ambiguous data, though it wasnoted that suspension of judgement is probablynot encouraged in classroom science.

The NCC report also found that the inability touse data persists even when children collect theirown data:

...no pupils referred to, or made use of theirdisplay work at any stage. Patterns which wereapparent in the display were not recognised:irregularities in the sequenced results were notrecognised...The overriding impression obtainedwhilst.reviewing the reports was that pupils seethe production of some form of graphical displayas little other than a ritualistic exercise (something'you do' after practical work) without recognisingany purpose or significance in what they do.

(Foulds et al., 1992)

Austin et al. (1991) suggest that ordering data isa necessary first step towards generalisation, sinceunordered data are difficult to handle. By 14 yearsof age, more children ordered the data, but thiswas no guarantee they would then make a general-isation. Austin et al. also suggest that numericaldata can be distracting in that children do notalways link the data to the reality of the phenome-non that they represent, and so they may justrepeat the numbers in their conclusions when theyare asked what they have found out (as in the Pauland Craig example. see p. 78).

114, -t )


Paal and Craig (Aged 11)

Find out hemi the distance moved by the model depends on the amount the elastic band is stretched.

We had to find out the stretch of the elastic band made the butter tub went the longest. We put the elasticband on two chair legs and pulled it back and we let it go the best it was the one we pull back 30 cm it wentfour metres twenty six cm the one that went the least was the one we pulled back 5 cm it went 30 cm. It was a

tear test because we used the same tub and the same elastic band.

How.far we pulled it hack How Jar it went

25 cm

20 cm

15 cm

1() cm

30 cm

5 cm

3 m 10 cm

3 m 40 cm

2 in 22 cm

m 19 cm

4 m 26 cm

31 cm

500

400

300

200

100

0

In another example from the NCC project,Paul. Mark and Stephen (see p. 80) produced agood table of data. They had ordered thcirresults and reported a generalised pattern at thestart of their written record. It seems that theythen decided that a bar chart might be a goodthing. They then drew one which, one could beforgiven for thinking, was a deliberate attempt

to confuse the reader. Why they chose to shufflethe order of their x-axis is left to the reader as anexercise in imagination. It appears from the barchart that. to this group. the values of the stretchof the elastic band are little more than labelsbearing no relationship to each other. Their con-clusion was clearly not based on the pattern intheir bar chart.

THE PERFORMANCE OF INVESTIGATIONS IN SECONDARY SCHOOLS: A DETAILED LOOK 79

Ars (Ased 12),

Whith cup kept watca hottest?

Type of cup Start temp. After I min. After 2 min. After 3 mM. After 4 min. After 5 min. How muchtemp. dropped

Polystyrene 88°C 82°C 79°C

Plastic 70°C 64°C 62°C

Paper 75°C 72°C 70°C

76°C 73°C 70°C 18°C

60°C 58°C 58°C 12°C

65°C 62°C 60°C 15°C

90

80

70

60

50

40

30

20

10

0

Polystyrene cup

0 1 2 3

Paper cup

4 5

80

70

60

50

40

30

20 -

10

00 1 2 3 4 5

70

60

50

40

30

20

10

Plastic cup

Our conclusion is that of the three cupsthe paper cup is the best as from its startingtemperature it dropped the least amount

1.4 r


10)1Mingideldivagis on the amount the elastic baud is stretched.

What we found: What we found was the further you pull the elastic band back the tube went further.

Stretched (cm)

2 40.1 cm

4 66 cm

6 133.5 cm

8 250.8 cm

10 237 cm

12 250 cm

14 361.5 cm

16 400 cm

18 537 cm

20 547 cm

600

500

400

300

200

100

2 16 20 12 14 8 10 6 18

Another group produced similar data but choseto order their bar chart:

10 15 20 25 30

Stretch

and then concluded: 'The further the elastic band ispulled the faster the model moves across the floor.'

fr;

This group appears to be moving towards therecognition of a pattern but have not yet realisedthat a line graph would illustrate the pattern better.

Foulds et a/. (1992) wrote that while the major-ity of children attempted a conclusion, theconclusions 'were not in keeping with the data atall, but were, in fact, at odds with it'. And also:'Many of their conclusions and inferences madelittle, if any, use of the data which had been gath-ered.' We have seen a number of instances of thelatter in examples earlier in this chapter.

Does the type of investigation affect children'sability to make sense of the investigation? Table 5.7shows the percentage scores for 'sensible' general-isations for each investigation type. (Our definitionof 'generalisation' was liberally applied; if we hadrequired the generalisation to be both true andfounded on the data, the figures would have beenvery much lower.) The overall trend is that the per-


Table 5.7 Percentage scores for generalisation for eachinvestigation type

Type I Type 2 Type 3 Type 4

Appropriategeneralisation 53 :6 33 25

centage score decreases a: the investigation typeincreases. As would be expected, type 1 investiga-tions are the easiest from which to generalise, withover half the sample generalising appropriately. Asthe complexity of the data increases, so the inter-pretation of the data becomes increasingly difficult.

The example below of a type 2 investigationshows only the graph produced (from a table) by'Group A', who measured speed by measuring thedistance the butter tub travelled in one second.Their generalisation is tied to an understanding ofthe concepts underlying the task or to the abilityto link the variables and the data to the reality

81

that they represent. It could be that such anunderstanding is necessary before the notion ofthe pattern in the data becomes meaningful. Onlythen, perhaps, can we expect to see the transitionfrom a disorganised bar chart (Paul, Mark andStephen, aged 11), through an organised one tothe line graph below.

In type 3 and type 4 investigations, where mul-tivariate data are involved, neither bar charts norline graphs are particularly helpful. Similar per-centages of pupils chose to represent their datagraphically as in the other types of investigations,supporting the above observation that the drawingof any type of graph may be a ritual rather than apurposeful exercise. Incidences of success were fewand far between here. So, rather than list them. wehave included two examples of pupils' work whichshow how well they can do. Simon's final conclu-sion may be open to question, but there can belittle doubt that this is a good piece of work froman I 1-year-old pupil.

Gireep A

Does the speed of the model depend on the amount of energy stored in the CialltiC bead

c

0E

-0 g76 0>-

300

250

200

150

100

50

00 10 20

Distance pulled back (cm)

30

The energy in the elastic band does aliect the speed or the carton. The more energy in elastic band, the fasterthe carton goes.


Simon (age 11)

The great tea problem

I am doing the one to find out whether both things affect how quickly the sugar dissolves

First I filled a beaker with hot water and made sure it was 75 ml, then I put 10 spatulafulls in and see the timergoing simultaneously. I did the same again but put cold water in instead of hot water. After that I put somehot water in and put white sugar in instead of brown, then I put cold water in instead of hot water. I alsostirred them all.

Type of sugar Temperature qf water(°C)

Time for sugar todissolve (min:s )

Brown 20 3:46

White 21 1:46

Brown 72 1:07

White 75 1:00

250

200

150

100

50

111111-mmIII Brown, hot

III Brown, cold

White, hot

jJ White, cold

They both make a difference but sugar doesn't matter as much as the water.

Brent (age 11)

In this experiment I set up the apparatus as shown in the diagram. I then changed the distance between thetwo strips of metal and observed any change in the reading of the voltmeter. I then repeated the experimentthis time keeping the distance apart constant and varying the depths of the strips in the fruit (a lemon) andmeasured the reading. My results are displayed in the table and graphs opposite.

n0 I


Distance between Volts

5 2.6

4 2.7

3 2.8

2 2.9

3.6

83

Amount submerged Volts

5 cm

4 cm

3 cm

2 cm

4.5

3.9

3.7

3.6

My conclusion is that the greater proportion of the plate submerged and the closer together (withouttouching) the higher the voltage will be.

Graph to show varying voltage dueto changing distance between poles

4

3

(i)

'3 2

oo 1

2

Distance

4 6

Graph to show varying voltage dueto changing depth between poles

5

4

3

2

0 1

0 2 4 6

Distance

The overall picture, then, is that in generalchildren experience considerable difficulty ininterpreting graphical information and indeed ininterpreting data overall. Again, we are driven tothe view that it is the underlying understandingof the purpose of gathering data, and the rolethat evidence plays in the work of science, that isconspicuous by its absence. This is not surprising,and indeed it may be asking too much of pupilsof this age to expect them to cope with theseideas. But until we have tried teaching theseideas, we shall not know. What can be said is thatthere arc enough instances of pupils' work which

do show elements of' that understanding to giveus hope that systematic teaching will result inconsiderable improvement.

Evaluation

What do we know about children's understandingof the validity and reliability of an investigation?There is very little direct research evidence con-cerning these concepts. This may he because theyare in practice very difficult to probe.

In an early APU report (Hat- len et al., 1981),11-year-olds were assessed as to their 'willingness

e--


to be critical of procedures used'. This was basedon observation by the assessor and on the pupils'answers to a question after the investigation aboutwhat changes they would make if they did theinvestigation again. Clearly, the question itselfprompts the pupil to evaluate. The resulting data(Table 5.8) arc difficult to interpret given thatHarlen zt al. admit that in some investigationswhere the pupils had performed satisfactorily, theonly alternative to consider was a less satisfactoryprocedure. In these cases, there was no need for apupil to be critical of his or her work.

In the NCC project, less than 1 per cent of thetotal sample attempted any form of evaluation.This may well reflect a different definition of theterm. We are defining evaluation in this book asthat underlying understanding which guides thedesign of the complete task. Other workers haverequired a more explicit indication of the under-standing of evaluation by relying on pupils to spoterrors in their procedures.

This evidence, however, is retrospective.Asking pupils to comment on how they thinkthey could improve their investigation is alreadytoo late. What we are concerned with here is theability to keep the requirements of believable(that is, valid and reliable) evidence in mindthroughout the task from defining variables sothat the question is being answered, to choosingappropriate ranges of instruments-and readingsto spot patterns unambiguously, to seeing thelink between the data and the type of graphicalrepresentation, right through to the critical inter-

Table 5.8 The APU assessment of pupil attitudes

Category Percentageof pupils

Uncritical of procedures used 41

Aware of alternative procedures butdoes not have very good reasons fersuggesting changes 38

Shows awareness of variables rotcontrolled, the need to repeatmeasurements, ineffective proceduresor factors central to investigation 20

pretation of both data and method of data collec-tion. What we have seen is that this notion is atbest patchy and at worst non-existent.

They are difficult ideas. Some would argue thatthey are more appropriate to sixth-form level (17-and 18-year-olds) or higher, and certainly themathematical treatment of errors is notoriouslydifficult for pupils at A level. But lower-le7e1notions of validity and reliability are, we believe,achievable. Work in some local schools certainlysupports that belief, but it also highlights the factthat sorting out the best techniques of teaching isnot going to be easy.

Summary

In this chapter, we have used existing researchdata to apply the notion of concepts of evidenceto children's performance in investigations. Wehave shown that children's procedural under-standing of the design are generally good in thatthey can structure the investigation successfully.but that children do experience difficulty in identi-fying variables as continuous where appropriate.Secondary pupils have a reasonable grasp of themeasurement, although they seldom repeat mea-surements. Ideas about data handling andevaluation are particularly poor. Foulds et al.(1992) wrote: 'During analysis of Children's workit became very clear that this area of children'sworking [the use of data and evaluation] appearedto be severely neglected', and 'The most strikingfeature of pupils' work is their lack of understand-ing of the nature of evidence.'

We would wish to suggest that, harking back toour procedural taxonomy, while the individualskills and concepts of evidence are being dealtwith in teaching to some extent, the ideas of appli-cation and, in particular, synthesis, are less to thefore. We need examples of schemes of work whichintegrate these ideas successfully and whichdevelop a common language of discourse betweenboth teachers and pupils. As we argued above,unless the link between the design and implemen-tation of an investigation and the requirements ofbelievable evidence are dealt with, pupils will tendto 'go through the motions' in practical lessons.

83

THE r. 'ERFORMANCE OF INVESTIGATIONS IN SECONDARYSCHOOLS: A DETAILED LOOK

All of which points inescapably to improvementsin teaching and in schemes of work to take theseideas into account. And it is to issues of teachingthat we now turn.

References

Archenhold, F., Austin. R., Bell, J., Black, P., Braund,M., Daniels, S., Holding, B., Russell, A. andStrang, J. (1991). Profiles and Progression in ScienceExploration. Assessment Matters No. 5. London,SEAC/EMU.

Austin, R., Holding, B., Bell, J. and Daniels, S. (1991).Patterns and Relationship.s in School Science. Assess-ment Matters No. 7. London, SEAC/EMU.

Dickson. L., Brown, M. and Gibson, 0. (1984).Children Learning Mathematics: A Teacher's Guide toRecent Research. Eastbourne, Holt, Rinehart andWinston for the Schools Council.

Driver, R., Child. D., Gott. R., Head, J., Johnson, S..Worsley, C. and Whyte, F. (1984). Science in Schools:

85

Age 15. APU Science Report for Teachers No. 2.London, HMSO.

Foulds, K., Gott, R. and Feasey, R. (1992). InvestigativeWork in Science. Durham, University of Durham.

Foxman, D., Ruddock, G. and McCallum, I. (1990).APU Mathematics Monitoring, 1984--88 ( Phase 2 ).Assessment Matters No. 3. London, SFAC/EMU.

Gott, R. (1984). Electricity at Age 15. APU ScienceReport for Teachers No. 7. London, HMSO.

Harlen, W., Black, P. and Johnson, S. (1981). Science inSchools: Age 11. APU Science Report for TeachersNo. I. London, HMSO.

Russell, T., Black, P., Harlen, W.. Johnson, S. andPalacio, D. (1988). Science at Age I I: A Review ofAPU Findings, 1980-84. London, HMSO.

Strang, J. (1990). Measurement in School Science.Assessment Matters No. 2. London, SEAC/EMU.

Taylor, R.M. and Swatton, P. (1990). Graph Work inSchool Science. Assessment Matters No. I. London,SEAC/EMU.

Welford. G.. Harlen, W. and Schofield. B. (1985).Practical Testing at Age.s 11,13 and 15. APU ScienceReport for Teachers No. 6. London. HMSO.

"u 4

CHAPTER 6

Investigations and teaching

Introduction

In the light of the research presented in the lasttwo chapters, how can ilivestigations best beincorporated into the science curriculum? We havesuggested that one of the ultimate aims of the sci-ence curriculum is to enable as many pupils aspossible to recall, understand, apply and synthe-sise (I) a whole range of skills and concepts ofevidence as well as (2) the 'traditional' or substan-tive concepts of science to solve a range ofproblems. To achieve these aims, teachers needgradually to expose pupils to a range of increas-ingly more complex concepts and procedures.

How can investigations help in achieving thesecurricular goals? We considered the roles of differ-ent types of practical work in Chapters 1 and 2and from this suggested that investigations are thebest type of practical to provide children with theopportunity to syntlwsise procedural understand-ing, which in turn relies on their understanding ofconcepts of evidence.

What we must explore is how investigationscan best be selected to develop that range ofskills and procedural understanding whichunderpins this area of science. In this chapter, weshall consider first how investigations can bedesigned to focus on a specific learning outcome.And, second, we shall put forward some ideasfor reinforcing particular concepts of evidenceand then examine the issues of progression anddifferentiation.

Changing the overall focus of investigations

The procedural and conceptual understandingused in investigations cannot be separated becausethey are inextricably intertwined; procedures inscience cannot be employed without using con-cepts and likewise concepts cannot be usedwithout employing procedures. However, we canemphasise one or the other of these two types ofunderstanding by carefully manipulating the struc-ture of the investigation. Foulds et al. (1992) givesome examples of how this can be done (Table6.1). The examples are restricted to the quantita-tive investigations described in previous chapters.

The investigations in column I might be usedwhen pupils are beginning investigative work.They use relatively simple concepts and the pro-cedural understanding required includes basicmeasuring skills together with ideas of a fair test,range, interval and patterns. By contrast, theinvestigations in the second column, which involvetwo independent variables, require an increasinglysystematic application of procedural understand-ing. The evidence that has to be collected,interpreted and evaluated is more complex, since itinvolves multivariate data. The concepts involved,however, remain relatively straightforward.

The third column gives examples of investiga-tions where the conceptual demand is higher, inthat the identification of the variables requiresmore advanced knowledge and understanding. Thcvariables may also be derived quantities as in thecase of speed. This increases the sophistication of


Table 6.1 Changing the emphasis of investigations (based on Foulds et al , 1992)

Low proceduralunderstanding

Low conceptualunderstanding

High proceduralunderstanding

Low conceptualunderstanding

Low proceduralunderstanding

High conceptualunderstanding

High proceduralunderstanding

High conceptualunderstanding

Find out whether sugardissolves faster in hotwater than in cold water

Find out how the distancetravelled by a toy cardepends on the amountthe elastic is wound up

Find out whether plantsgrow better if they arewatered with fertilisersolution than if they arejust given ordinary water

Find out whether the rateat which sugar dissolvesdepends on:(a) the type of sugar. and(b) the temperature of the

solution

Find out how the distancetravelled by a model cardepends on:(a) its weight, and(b) the force used to get it

moving

Find out how the growthof a plant depends on:(a) the amount of light

which strikes it, and(b) the temperature of the

surroundings

Find out which of thesechemicals cause hardnessin water

Find out whether thespeed of a model dragsterdepends on the amountof energy stored in theelastic band

Find out how the rate atwhich fermentation takesplace depends on thetemperature of the solution

Find out how the speedot' the thiosulphate reactiondepends on:(a) the concentration of the

solutions and(b) the temperature

Find out how the efficiencyof an electric motordepends on:(a) the load being lifted, and(b) the speed at which it is

operated

Find out how the rate atwhich photosynthesis takesplace depends on:(a) the light intensity. and(b) the temperature

the evidence without necessarily increasing itsvolume. The investigations in the fourth columnare examples when both the procedural and con-ceptual demands are high.

Targeting specific concepts of evidence within wholeinvestigations

It is tempting to think that carrying out an investi-gation requires that pupils be taught skills andconcepts of evidence in advance and, as far as ispossible, one by one. But this inductive approachis at least a debatable one. Most concepts of evi-dence are closely tied to whole investigations, sothat it is difficult to teach them individually.

Consider the notion of the relative scale of thevariables in a practical: the choosing of sensibleproportions for the relevant variables in relation

8 6

to the instruments available. Let us consider asimple situation, such as finding out which ofseveral kinds of paper towel is most effective insoaking up water. If pupils attempted to measurethe amount of water soaked up in one square cen-timetre of paper towel, most people would agreethat this would be an inappropriate scale, since therelative amounts of water soaked up by such asmall area would be difficult to measure using themeasuring instruments which are normally avail-able in schools. Clearly, it is difficult to conveythis idea of scale without describing the wholeinvestigation of which it is a part. Again, if weconsider the idea of 'believability' of evidence, it isnecessary, at least until the notion is embedded, toaddress the issue in the context of data that pupilshave collected themselves and, therefore, 'under-stand'. We can also draw on our own experienceand the experience of other teachers in the area.

INVESTIGATIONS AND TEACHING

which suggest that lessons which develop skillsand concepts of evidence within investigativework, with the skills set clearly within the contextof a whole investigation, are seen as more mean-ingful by the pupils.

So the order should be to teach whole butcarefully focused investigations first. Theseshould then be followed up by exercises on spe-cific skills and concepts of evidence to teach orreinforce specific points of difficulty which havebecome apparent in the investigation. This orderof teaching may not be easy to organise, but webelieve it is necessary to avoid 'fragmentation'.

In the following section, we suggest ways oftargeting each of the broad categories of conceptsof evidence, design, measurement, data handlingand evaluation, all within investigations. Theseinvestigations, together with the follow-up activi-ties, will then constitute a menu to be drawn onin the design of schemes of work considered inthe next chapter.

Design

The identification ofvariables (asindependent anddependent).

Example:Which fuel is best?

Any simple investigation where a variable is notspecified can be used to focus on the identificationof variables. The lesson could begin for instancewith a brainstorming session on what 'best'means, either directly with the whole class or aftersmall group discussion. A simple planning exercisebefore carrying out the investigation using boththe tasl: in hand and other examples can be usedto reinforce the concept of independent anddependent variables.

The design of the fairtest and its associatedcontrol variables.

Example:Which paper towel isbest for mopping upwater?

89

The choice of an investigation which involves sev-eral control variables can reinforce the importanceof the fair test. Brainstorming can be used heretoo. A discussion of the effect of not controllingvariables is also useful.

The type of variables(categoric, discreteor continuous).

Examples:Which type ofinsulation is best forlagging a centralheating boiler?(independentcategoric variable)

How is heat lossaffected by numberof layers ofinsulation?(independent discretevariable)

How does heat lossdepend on thevolume of the boiler?(independentcontinuous variable)

Types of variables can be emphasised by using aseries of investigations such as the examplesabove, either sequentially with the whole class orby having different groups in the class doin2 dif-ferent investigations. In the latter case, whole classdiscussion following the investigative work isessential to focus on and compare thc differenttypes of variables.

After the investigation, the relationshipbetween the type of variable and the type of graphcan be discussed. For instance, it may be helpfulto relate the 'sudden jumps' in categoric variablesto the appearance of the bar chart and the gradualchange of continuous variables to the linear orcurvilinear form of the line graph.

Sample sire andariation.

Example:Find out how lightaffects the growth ofcress seeds.

INVESTICATIVE WORK IN THE SCIENCE CURRICULUM

Sample size can be emphasised by any investiga-tion where this concept plays a significant role,often in biological contexts. It may be useful toallow the investigation to proceed and then corn-pare the 'believability' of the results of groups whohave selected different sample sizes.

All these examples are dependent on the teacherfocusing the practical on the relevant idea. Heregroup and/or class discussion at strategic pointsthroughout the investigation are essential so that theaim of the lesson to learn about design is clear toall. The role of the teacher here is paramount, afactor which is discussed further in the next chapter.

Measurement

Concepts of evidenceassociated with scale.

Examples:Which paper towel isbest for mopping upwater?

Find out how the rateof dissolving of (achemical) depends onthe temperature ofthe water.

The importance of relative scale can be a difficultconcept to convey. The first of the examples abovecan be used to draw pupils' attention to the im-portance of matching the scale, of the amount ofwater that can be squeezed from the piece ofpaper. to the instrumentation available. A rela-tively coarse measuring cylinder, for instance,constrains the optimum size of the sample if themeasurement is to be accurate enough. In theother example, thc saturation of a solution is aconcept which influences scale. Clearly, if toomuch chemical is used, the solution will becomesaturated. If this situation arises among groups inthc course of the investigation, then by comparingresults afterwards, the effect of inappropriate scalewill become obvious.

8 3

Concepts of evidenceassOciated with rangeand interval and thechoice of instrument.

Example:Find out how theheight of a slopeaffects the amountof pull needed to pullan object up it.

Investigations which involve continuous variablesalmost inevitably lead to considerations of theconcepts of appropriate range and interval. Theimportance of range often becomes obvious wheninterpreting patterns in line graphs. If the range istoo narrow, only part of the pattern will emerge;in the example above, for instance, the relation-ship is not linear but in fact peaks at an anglesomewhat less than 900 due to the effects of fric-tion. Again, whole class discussion of differentgroups' results is very useful here.

The concept of the choice of instrument can beemphasised by making available a wide variety ofmeasuring instruments from which the pupils haveto select. In the same investigation, a large numberof forcemeters with different ranges should bemade available as a prerequisite, if pupils are tounderstand the effect of their choice on the accu-racy of their measurements. The teacher mightallow pupils to choose inappropriately in order tolearn from their mistakes or he or she may stopthe class after groups have chosen their instrumentand made their first few readings and elicit ideasfrom the children about the pros and cons of' dif-ferent forcemeters for the task.

Repeatability. Example:Find out how thebounciness of thesquash ball dependson temperature.

Th..! need for repeatability can be targeted by usingany investigation in which variability is a promi-nent feature. In the example, the need for repeatsis essential if the evidence is to be reliable and


accurate enough to uncover the trend in theresults. Again, group comparisons can be usefulafter the investigation or, alternatively, the teachercan question groups while they are collecting theirdata in order to encourage them to consider thereliability of their measurements. The teachermight ask, for instance: 'If you did it again wouldyou get the same results?'

Accuracy. Example:Which sugar dissolvesthe quickest?

The appropriateness of a particular level of accu-racy can be brought out by comparing the datafrom different groups in the class. For instance, indissolving investigations pupils often record timewith digital watches to three decimal places. Othergroups may record the dissolving time to the near-est minute. The latter may not be accurate en Jughto reveal a trena but the former may be equallyinappropriate.

Data handling

Concepts of evidenceassociated 1.x.,th datahandling; the use oftables.

91

Examples:Which type of sugardissolves quickest?

How does thetemperature of thewater affect dissolvingtime?

The use of a table as a planner or organiserreflects an understanding of the design of theinvestigation in terms of the identification and thetype of the variables. All too often it is used afterdata collection as a neat way of organising jottingsin the backs of books. A table can show theintended number, range ard interval of the valuesof the dependent variable. Two related investiga-tions, one of which involves a categoric and theother a continuous independent variable, canserve io point to the advantages of a table as anadvance organiser (Fig. 6.1).

It can be seen that the column headings identifythe independent and dependent variables. In tables

Which type of sugar dissolves quickest?

The independentvariable

The dependentvariable

Sugar type Time (sec)

Caster

Granulated

Soft brown

How does the temperature of the water affect dissolving time?

The independentvariable

The dependentvariable

Temperature(°C)

Time (sec) .

10

25

50

75

100

The number of valueschosen reflects theinterval and range

Fig. 6.1 The use of tables to structure and plan investigations

0


of continuous variables, the values chosen also setthe number, range and interval of the readings.Concepts related to the type of variable, range andinterval all come into play here; it is perhaps thisinteraction that causes pupils so many problems,with continuous data in particular. A table such asthis, used as a discussion point on the blackboard,can serve as a very useful teaching aid in discussingthe planning of investigations.

Graphs. Example:Find out how thedistance travelled (bythe toy car) dependson the amount theelastic band is pulledback. When you havefinished, you will begiven a distance andasked to predict howfar back to pull yourelastic band.

The choice of graph type is related to the type ofvariable. For many pupils, graphs are simply adisplay, and the importance of line graphs asreflecting an underlying relationship between vari-ables is missing. One way of approaching the issueis through an investigation where the line graphcan be shown to be a positive advantage ratherthan a chore, as in the example above. This kindof competition which demands interpolation orextrapolation, encourages pupils to see t:te pur-pose of Jin graphs for exploring the relationshipbetween two variables. Without such a purpose,the drawing of a line graph for continuous datacan become a ritual.

The concept of patterns provides another oppor-tunity to look at the pupils' ideas of the relationshipbetween the data and the reality that they repre-sent. Patterns in the data are usually determined byexamining tables or graphical representation.Understanding how those patterns relates to thereality of the ti.sk in hand can be targeted hy askingdifferent groups to discuss what their data mean.Alternatively, if, in the second dissolving investiga-tion (Fig. 6.1), the pupils are asked to predict from

their results what happens if, for instance, the tem-perature is doubled from 30 to 60°C, they mayproceed to double (or halve) the dissolving timethat is, using mathematical patterns which do notrelate to the scientific pattern of the variables.

Multivariate data. Examples:Which cups should themarket sfall holderbuy to keep drinkshot? Should hechoose:

Plastic orpolystyrene?Cups with lids orwith no lids?

How does the sag of aplank depend on theweight on the plankand its length?

In the first example, the most efficient way of con-sidering the effect of the two independent variablesis as shown in Fig. 6.2. The research evidenceshows that many children will tackle such investiga-tions by testing the two independent variablesseparately. In the example given, pupils tend to testthe effect of the type of cup and presence/absence oflid in two separate investigations. This investigation

Startingtemp.

Temp. after10 min.

Temp.drop

Plastic cup(no lid)

Plastic cup(with lid)

Polystyrenecup (no lid)

Polystyrenecup (with lid)

Fig. 6.2 A multivariate table


could be madc mote realistic if the costs of the cupsand lids are given to the pupils so that the relativeeffects of heat loss and cost can be assessed. Wherethe independent variables are continuous, as in thecase of the second example above, two or moregraphs drawn on the same axes can help to illustratethe relative effects.

Evaluation reliability and validity

Children can be gradually introduced to these con-cepts by carrying out investigations where there is atarget audience, perhaps others in the class. Ifgroups are doing different tasks, each group can seethe need for convincing other groups who have not'seen' and therefore are not in a position to 'believe'.Choice of task here is less important, although onein which there is not a self-evidently correct answeris likely to be more fruitful. The slopes investigationmentioned above is one such task.

Techniques such as asking groups of pupils todefend their results against cross-examinationby an advocate from another group, or the class.can be of use here. Such techniques require care-ful introduction and a sense of theatre to drivehome the importance of the audience, which hasto be convinced by the 'objective evidence ofthe scientist. Gradually, pupils begin to realisethat the nature of scientific evidence is notstraightforward.

The reinforcement of concepts of evidence

We shall mention briefly here some ideas for rein-forcemcnt, most profitably used as a follow-upactivities to targeted investigations.

Associated with design

The example of a question from a written exercisefrom CASE (Adey et a/., 1989), which focuses onthe concept of the.fair test and associated controlvariables,is shown here:

93

were timed Qui-mg this race, they both started torun at the same time.Is this a fair test?If not, why not?

CASE uses pictures and diagrams to help to clar-ify the question and also makes these exercisesmore attractive to the pupil.

Associated with measurement

These two follow-up activities were designed tofollow on from a 'slopes' task (`Find out how theangle of the slope affects how easy it is to pull thebrick up'). The first activity focuses on the choiceof measuring instrument for a particular task (seeFig. 6.3). The second (Fig. 6.4) focuses on the ideaof a continuous variable and, more particularly,on the notion of range, interval and number of aset of values for the independent variable.

Measuring the force

0-10N

0-25N

0-50N

Which of these is the best Newtonmeter to useto m..asure a force of 7 Newtons?

Why?

During a school sports afternoon, it was decided Fig. 6.3 An activity emphasising the choice of anto see it' boys ran faster than girls. John and Jane appropriate measuring instrument

The pupils are asked to think about an investigation tofind out how the angle of a slope affects how easyit is to pull a brick up.

Which angles would you use?

Here are the tables from several groups

Group 1 Gro,.ip 2 Group 3

10307090

0306090

Group 4 Group 5

05152040457090

020406080

01020304050

Which group's is best?

Why?

Which group's isworst?

Why?

Fig. 6.4 An activity emphasising the choice of range and interval

Susan,Vic*y and Leanne had made model cars using margarine tubs. To launch the cars, they used elastic bandsstretched across stools. They each measured how far the cars travelled. They put their results in a table.

Amount elastic pulled back Distance travelled

2 cm4 cm6 cm8 cm

10 cm

35 cm55 cm70 cm87 cm

101 cm

120

100

80

60

40

20

00 2 4 6 8 10

120

100

80

60

40

20

0

They each drew a graph to show their results.

120

100

80

60

40

20

Oo2 4 6 8 10

There was going to be a competition to see who could land their car closest to a line drawn on the floorThey could measure how far it was from the elastic band to the line before they started.1 The three graphs show the same results in different ways.Which graph do you think is best? Explain why you chose ths one.2 Discuss with your friends which graphs they think are best.When you have agreed which is the best, draw it onto a Lill sheet of paper. Remember that the labels are important.

Fig. 6.5 An activity concerning the choice of type of gni ph (based on an example in Foulds al_ 1990)


Try irawing sketch graphs for each of these:

1 'atch some sugar dissolving inhot water. Sketch a graph onthese axes:

Amount ofsugar

Time

Then choose your own axes for these:

Your height as you gotolder

95

tRoll a marble or a toy caralong the bench

Speed

Distance along bench

Force needed to push alorry as it is loaded with bricks

Fig. 6.6 Examples of a line graph exercise which can be used to teach the relationship between the behaviour of avariable and a graphical pattern

Associated with data handling

Understanding the reasoning which underlies thechoice of graph type, with its link to the type ofvariable, can be approached using 'second-hand'data such as in Fig. 6.5. Developing the under-standing of patterns in data is, as we have notedfrom the research, one of the more difficult con-cepts. One technique which works well is to usesketch. graphs. Asking pupils to say which graph'feels' like the stretching of a rubber band, forexample, seems to help in establishing that linkwith reality that is so often missing. Some ex-amples for follow-up activities are given in Fig.6.6. After pupils have become familiar with theidea of sketch graphs, the teacher could ask pupilsto predict before doing an investigation what theshape or pattern of the resulting graph might be.

Associated with evaluation

Evaluation is closely linked with a recognition thatinvestigations rely on valid and reliable data. With

older and more able children, recent press cuttingson topics such as the incidence of leukaemiaaround Sellafield or the testing of new drugs couldform the basis of a class discussion aimed at intro-ducing pupils to the idea that data can bepresented which may be invalid, or unreliable intheir accuracy or interpretation. Such exercisesalso help children to relate school science to thereal world. With younger children, using datafrom other similar-aged children can be a goodsource of appropriate second-hand data.

Targeting conceptual understanding

Investigations can be useful at the beginning of atopic as a 'window' into children's implicit. under-standing. Sometimes pupils have an everydayunderstanding of an idea such as heat transfer orfriction. Often those ideas are held implicitly.Sometimes they are wrong. Some examples of openinvestigations based on the research programmeand used in this way are shown in Table 6.2.


Table 6.2 Examples of investigations which can uncover children's misconceptions

Investigation Concept focus

Which fuel is best?

Which materials wouldbe best for a hedgehogto use for its nest?

This investigation can be used as an introduction to the concept of fuels asconcentrated sources of energy. The choic: the dependent variable includes: smoke,smell, time to heat water, amount of fuel to heat water. Each of these responses canbe used to open up ideas concerned with environmental pollution, or the distinctionbetween power output and total energy content.

This investigation can be used as an introduction to the concepts of heat transfer.

An example of a pupil response is of interest here. One low-ability pupil asked for ahard-boiled egg to use as a hedgehog. He asked that it be hard-boiled for a long time.Eventually, it dawned on the teacher that he imagined that an egg could be packedwith heat a caloric theory view.

Table 6.3 Varying procedural demand within a single context

Investigation l'ariable structure

Find out whether the amount of juice depends on the type of applewhich is used

Find out how the amount of fruit juice extracted depends on theamount of pectinase enzyme which is used

Find out whether:(a) the type of enzyme used, or(b) the type of apple which is usedhas the greatest effect on the amount of fruit j lice extractedfrom the pulp

Find out how the amount of fruit juice extracted depends on:(a) the amount of enzyme, and(b) the temperature of the pulp

A single categoric independent variable

A single continuous independent variable

Two categoric independent variables

Two continuous independent variables

Progression and differentiation by outcome

The structuring and sequencing of activities is thebasis of well-planned schemes of work. Choosinginvestigations which gradually increase in theirdemand is just one element of that structure.Differentiation within a particular lesson relies onthe same principle. So the same examples can beused in the one lesson, to enable pupils working atdifferent levels of understanding to be challengedbut not overwhelmed. We shall consider first how

9

to structure the procedural demand, and second,the conceptual demand.

The procedural demands of investigationswithin the same context can be matched to pupilsof different ability. Table 6.3 gives examples,within the context of juice extraction, of' this differ-entiation by task. Usinsz the same context meansthat th,.: conceptual element can be kept thc samewhile the procedural complexity is changed.

Foulds et al. (1992) point out that an investiga-tion set at a high level of procedural demand can be


Table 6.4 Investigations with an increasing conceptual demand within one context

97

Find out whether thedistance travelled dependson the.amount by whichthe elastic band is stretched

Find out whether thedistance travelled by themodel depends on theamount of energy storedin the elastic band

Find out whether theaverage speed of the modeldepends on the forceexerted by the elastic band

Find out whether theaverage speed of the modeldepends on the amount ofenergy stored in the elasticband

performed in practice at each of the lower levelswith differentiation by outcome. For example. theinvestigation with two continuous independent vari-ables could be restructured by pupils as having onerather than two categoric independent variables.

In the same way that procedural demand can bevaried to meet the needs of differentiation andprogression, so can conceptual demand. Theexamples in Table 6.4 show how investigations canbe targeted so that the conceptual demand is lowor high. while the procedural demand is kept rela-tively constant.

By making both the conceptual and proceduraldemands of an investigation high. as in the fourthcolumn of Table 6.1. investigations can exploreboth conceptual and procedural understanding.These investigations make high demands onpupils. However, the logic of our argument withregard to the overall aims of the science curricu-lum and using the research findings in Chapters 4and 5. is that having developed an effective base ofprocedural and conceptual understanding. thepupil is then in a position to recognise when andw here the various procedures are applicable.

Summary

In this chapter, we have shown how the under-standing of concepts of evidence can be taughtthrough carefully chosen investigations whichhinge on particular concepts. Follow-up exercises.written or practical, can and should serve to re-inforce the concept in another context, to beginthe task of allowing pupils to apply the concept.As pupils' repertoire increases, so we can begin towiden the menu of investigations, allowing fortasks to be selected, within the same overall con-cept. for groups of pupils at different stages ofprocedural and conceptual understanding.

References

Adey, P.. Shayer. M. and Yates. C. (1989). ThinkingScience: The Materials ()laic CASE No./et!. Walton-on-Thames. Nelson.

Foulds. K.. Mashiter..I. and Gott. R. (199(t). investigationsin Science. Glasgow. Blackie.

Foulds, K., Gnu. R. and Fease. . (1992). lnvevigatireWork in Science. Durham. UM crsity of Durham.

Jo

F CHAPTER 7

Incorporating investigations into a scheme of work

One school in the North-East has been developingschemes of work incorporating investigations forsome time. What follows is part of a discussionwith the school on how to teach for proceduralunderstanding and the difficulties of devisingschemes of work.

Investigations do need to be specifically targetedbut not in the 'follow this instruction', teacher'sway. The teacher has to be aware of the conceptsof evidence they are trying to develop and givethem practice at them and be aware of how wellthey are coping just like teaching (substantive)concepts. except they are the same concepts (ofevidence) every time, but in different situations.And just like (substantive) concepts. some kidsdon't get them all. But more kids can get more ofthese substantive concepts than they can of knowl-edge and what's more it applies to other places intheir lives where they think. So you're going to geta developing 'thinking method' going as a frame-work in which to explore the world, vnd ifnecessary you can 'test out' the knowledge you arelearning. Seeing whether you really understand.

You've got to get them going. and then stopthem in their tracks and make them think aboutwhat they're doing. Shock them into believing intheir own reasoning powers. That there's no rightand wrong except what they decide based on theevidence in front of them. And you've got to usestra'egies that show some of them are thinkinglike this use the differences in investigations thatarise to learn from each other. Point out that it issignificant when someone does use their reasoningpowers successfully.

All a scheme of work can do is:

put in investigations regularly content chosennormally (based in traditional topic areas).think about investigations in terms of proce-dural complexity don't make them too hardto start with. then they can get harder. Contextvarying: sometimes scientific. In other words.build in progression.allow opportunities in between investigationsfor practising procedural unde:-Ltanding(follow-up activities). Just have as many waysas possible for getting them in those situationswhere they have to think. So giving them agraph to draw and interpret, getting informa-tion from tables, brainstormine. We're tryingto find more and better ways. Ways to getthem to think about these things. You have tokeep varying the approach or they get boredand won't think.now the flexibility must come in. as with anyscheme of work. What do the kids do in aninvestigation? Once you know this. you canthen focus on the weaknesses and strengths of(their understanding of) concepts of evidencein discussion. Do they (strengths and weak-nesses) match the ones planned for in thescheme of work? And then whether the kids arestrong in that al ea. or weak, there is either areinforcing aspect or a teaching aspect.

There can be no linevr formula. it is the build-ing of a round picture where all the aspectsinterlink gradually and individually and differ-ently in different situations.


I Open investigationThe topic of friction is introduced with an openinvestigation:

What affects the slippiness of training shoes?

It is set in context orally perhaps in the gym, or out-side if it happens to be winter and icy underfoot.There is no specific teaching focus to the task: theaim is to open up the topic of friction and allow chil-dren's ideas to emerge. It may transpire that mostchildren understand that, for instance, weight andarea have an effect. But it could reveal misconcep-tions about force or problems associated with itsmeasurement. In any event, it sets the scene for whatis to follow. The understanding revealed in this firstinvestigation may mean that some of the followingparts can be omitted. Alternatively, if unexpectedgaps in understanding are revealed, some additionsmay need to be made.

2 Choosing and using forcemetersThe appropriate skill of measuring force is intro-duced using a skills exercise where pupils are askedto measure different things such as opening adrawer, a door, etc. Pupils are allowed access to arange of forcemeters so that they have to decidewhich is the most appropriate for the task theyhave selected.

3 The concept offrictionTo extend pupils' ideas which the teacher had begun toelicit in the training shoe c .ercise, the question 'Whatis friction?' can be used as the topic for a teacher-leddiscussion of friction and its causes and effects.

4 Illustrating the concept and its measurementHere a demonstration using blocks with differentsurfaces can be used to illustrate the idea that fric-tion depends on weight and surface. Alternatively,a tightly controlled worksheet-driven session couldbe used.

5 A targeted investigation which uses these ideasTo put the investigation into a context, a video onthe theme of the making of the pyramids can he usedas an introduction. Pupils are then asked:

What do you think might affect how difficult itwould be to pull a brick up a slope?

(One of the ideas was probably the angle of the slope)

Do an investigation to:

Find out how the angle of the slope affects howeasy it is to pull a brick up.

When you have finished, your teacher will ask youto work out the best slope for somebody who canonly pull with a certain force. So you will need tocollect data that will allow you to do this.

This investigation can be presented in a variety ofother ways, some of which can be used as supple-mentary or alternative investigations for the moreable pupils. Or, lest boredom set in. one of themcould be the subject of a teacher-led demonstrationwith the pupils directing operations.

Another way of moving the stones is to use a blockand tackle. Make one using two pulley blocks. Findout what difference the block and tackle makes.

The smoothness of the surface makes a lot ofdifference. If you were a pyramid builder would itbe better to spend time making a very smooth slopeor use the time to make a long shallow slope?

6 Follow-up exercises ( used at various points in thesequence as appropriate)The investigation will have revealed strengths andweaknesses in pupils' procedural understanding.Exercises such as those in Chapter 6 can be usedhere to target particular concepts of evidence. Ifsome groups find interpreting data difficult, forexample. they could be given other groups' data topractise interpretation. If pupils are unsure of theuse of tables, they could be given a worksheet onhow tables can be used to structure design anddata collection.

Some of these exercises may be needed by mostpupils, so that it may be useful to organise a classexercise and/or discussion. Alternatively, if only a fewpupils are weak in a particular area, then the exercisecan be given for homework for reinforcement.

0

INCORPORATING INVESTIGATIONS INTO A SCHEME OF WORK

After three years of work in this area, themessage is that schemes of work must be com-prised of a series of signposts key learningoutcomes. Navigation between these signpostsmust be at the discretion of the teacher, and inresponse to all the vagaries of real-life teaching.But it needs a resource of investigations of vari-ous types, and follow-up work, and strategiesfor teaching and learning to hand. Having saidthat, schemes of work have to start somewhere.The following is an example of a teachingsequence that could be developed into a schemeof work. The substantive concept we havechosen is friction, on no better grounds thanthat all of the activities have been tested in theclassroom at one time or another.

The scheme of work can then continue in asimilar vein, interspersing investigations in orderto use substantive concepts which have alreadybeen introduced and to develop proceduralunderstanding. Table 7.1 summarises the pur-poses of each lesson.

It is from such a table that we can begin thetask of creating schemes of work. All too often ascheme of work is little more than sets of lessonnotes. What we should be aiming for is a route

Table 7.1 Beginning a scheme of work on friction

T-1-1:1.

map through the learning outcomes the thingsthat we hope pupils will understand afterwardsthat they didn't before. Once the learning out-comes are identified, the activities that are bestsuited to those outcomes can be designed.

Table 7.2 is an edited version of part of onesuch scheme of work. The start of the first year(pupils aged 11 in a comprehensive school) con-sisted of a three-week module called 'Be scientific',followed by a longer module introducing materialsin a topic called 'Our earth'.

Progression

Looking back over a pupil's work, it should bepossible to see that there has been progression ona variety of fronts. The example on p. 102 is byMichelle, an average year 7 (12-year-old) pupil ina comprehensive school in the North-East. Wehave chosen it from many examples because it isnot untypical of the way pupils learn, pi-ogressingand then regressing. What follows is a summary ofMichelle's progression in procedural understand-ing beginning with the work in the schemeoutlined on p. 102.

The lessonlactivity Principal learning outcome Conceptualunderstanding

Proceduralunderstanding

Open investigation

Choosing and usingforcemeters

Diagnostic assessment of conceptualunderstanding

Skill and concept of evidence(choice of' instrument)

The concept of friction To reinforce the concept of friction by relatingit to familiar situations

Illustrating the conceptand its measurement

To introduce and/or develop the idea ofmeasuring ffiction as a force

A targeted investigation Appling the concept of force and assessinghich uses these ideas procedural understanding

Follow-up exercises Reinforcing particular concepts of evidence

fl

jTable 7.2 Elements of a scheme of work


Be scientific Learning outcome Suggested activities

General introduction to lab safety Key danger points in a laboratory

Heating things safely and measuring Understanding tables andtemperature interpreting data

Heat transfer via conduction Insulators and conductors of heat

Insulators

Follow-up work

Applying understanding abouttables and checking on the notionof a fair test

The idea of variables, identifyingvariables (as independent and so on),recording the data and theirinterpretation, and the link betweena fair test and the validity of the data

Discussion centred on pictures ofdangerous lab situations, postersession. Using a Bunsen burner

Basic skill activity on thermometers

Investigation into insulatingmaterials to keep hot potato hot

Discussion plus demonstration orclass experiment on conduction

Investigation into the best type ofmaterial for a disposable hot drinkscup

Discussion of class results on the drinkscup investigation. Follow-up workon interpreting other people's data

In the first investigation, which involved differ-ent materials wrapped around hot potatoes. hcrrecord consisted of:

Change: materialKeep the same: size/type of potato. timeResults: 46°C before, 46°C after.

The class then pooled their results and one of theoutcomes was a table on the board which wastranscribed into their books. In a second, related.investigation concerning the type of cup and itsability to keep drinks warm, there was somedegree of progress. but in the idea of a fair testrather than in tables!

Timed it for 5 minutes and it dropped 8°C.Another one (the polystyrene cup) dropped 7°C.This means that the polystyrene cup is the bestinsulator. But we did not measure how muchwater we put in each. I think we could of (sic] hada more accurate answer.

Later, in the second module on 'materials'.there followed an exercise on interpretation ofgraphs. Michelle began to make progress in under-standing bar charts here and then moved into an

investigation to find out which paper is best touse for holdino. chips? Her results table was a con-siderable improvement:

Recyckd Grease New

paper proof improvedpaper paper

First trial 500 g 300 g 400 g

Second trial 400 g 200 g 500 g

Third trial 400 g 200 g 500 g

Her notebook also showed evidence of under-standing fair tests and repeats. Two furtherinvestigations interspersed within the module werecarried out reasonably well including tables, barcharts and fair tests. The final investigation in themodule concerned the bounciness of squash ballshow does the bounciness depend on temperature?This is a type 2 investigation with a continuousindependent variable. All the others had been typeI. And at this point Michelle goes to pieces:


We did our test and dropped the ball from 2 inWe did this three timesThen we wrote our results tableMy results.The first trial was 96 cm.The second trial was 90 cm.The third trial was 96 cm.The average was 94 cm.

She drew a bar chart of these four values. Herconclusion was: 'The pattern in my graph goes upand down.'

Like the ball. Michelle was completely thrownby the continuous variable issue to such an extentthat she seems to have fallen back on the last thingshe got a positive comment on, the averaging ofthree results. This regression happens often. Theonly remedy is repetition, so that the notion ofunderstanding, evidenced in the early investiga-tions, is transformed into the ability to apply innovel situations.

Some practical considerations

A note on introducing the languageof investigations

It is not uncommon when discussing investiga-tions with pupils, for them to express a sense ofuncertainty as to what it is they have learned.With other practical work such as an experimentcalled 'Hooke's Law', they know that that is whatthey are supposed to have learned. Furthermore.they can refer to the activity in their books andduring revision as the 'Hooke's Law' experiment.even if they don't fully understand it.

For example, the extent to which the language ofthe concepts of evidence we have proposed shouldhe part of the language of teaching is not at all

Table 7.3 Alternative ways of describing variables

103

clear It is certainly true that before pupils can begininvestigations, they need to understand some of thebasic language which is associated with them Thequestion is: Which ideas need to have their own lan-guage and what is that language to be? Terms suchas 'variable' and `control' can present difficulty forsome pupils. Various alternative phraseologies havebeen suggested to try to overcome the problem,some of which are shown in Table 7.3.

There is an argument for the `thing to change'approach to do with the accessibility of the lan-guage. But there is a counter-argument whichsays that the search for simple words can lead toconfusion and that it is better to deal with the lan-guage head-on using the 'correct' terms from theword go. The first three activities in the CASEproject adopt this approach by introducing theterms input and outcome variables. The groundrules in terms of language are carefully estab-lished and.then used consistently throughout theremaining activities.

In the example overleaf. taken from the NCCresearch, we see that Jonathan, when prompted towrite what he would change, says that he would'change nothing', whereas in the sense that theprompt wa:; intended, he did indeed change theindependent variable. What he was referring to,we may surmise, is that he controlled the otherappropriate variables and did a 'fair test'.

The word 'control' in relation to variables canalso create difficulties because of its confusion withthe biological meaning of 'control'. In the latter,the control refers to the condition of no treatment(e.g. the effect of no fertiliser on growth), which ininvestigations equates with a zero value of theindependent variable. This is quite different fromthe meaning of control variables. A further diffi-culty arises with categoric variables. The term

The independent variable The dependent variable The variables to be e im trolled

The 'thing' to change (systematically) The 'thing' to measure (lOr each The 'things' to keep the same tovalue of the independent variable) make the test fair

The input variable The outcome variable

106


Jonathan (year 7, aged 11)

Find out. him the diatance moved by the'model &lien& on the amount the elastic band

The bigger the elastic band stretched the farther itwill go.

We are going to change nothing.

The things we will measure are the farther the tubgone and how fare we pull the elastic band back.

'variables' is sometimes associated only withnumerical quantities, so that the idea that the typeof insulating material can also be a variable (a cat-egoric variable) with 'values' such as polystyreneor cloth, needs careful introduction.

The 'fair test', too, has presented difficulty, par-ticularly in primary science. There is someanecdotal evidence to show that pupils see a fairtest as being concerned with keeping everything insight the same because 'he very name 'fair' suggeststhat not to do that is somehow unfair. To otherchildren, the notion of a fair test is akin to a handi-cap in horse racing. If the horse is fast. it shouldcarry extra weight to slow it down. 'Fairness' in sci-ence means something quite ditThrent from'fairness' in a horse race or a playground brawl.

The above examples show the importance ofestablishing the ground rules in language early onto avoid misunderstandings and confusion. It isalso necessary to reinforce these terms by their con-sistent use in all experimental work, whichever setof terms is selected. Some concepts are closely asso-ciated, so that they arc likely to be taught together.For example, the concept of the pattern in linegraphs cannot be understood without some under-standing of range and interval. Whether otherterms such as relative scale, validity and so On needto be used, and if so when they should be intro-duced, needs further classroom-based research.

loj

Organising investigations in the science laboratory

The organisation of the working groups withinthe class depends to some extent on the nature ofthe problem to be investigated and its degree ofopenness. Where the problem is closely defined,the whole class will usually carry out the sameinvestigation, each group deciding on its ownmethodology. In terms of preparation and con-trol, many teachers will find this is the moststraightforward arrangement. More open prob-lems may lead to different investigations, in thesame context, being carried out within the sameclass, which will clearly require a wider range ofequipment. An example of this situation is whenthe teacher begins the lesson by raising questionsabout a concept such as friction (as in the samplescheme of work on p. 100) and asks the pupils toidentify factors which could affect the slippinessof a training shoe. Variables such as the type orarea of' the sole, the weight of the person wearingthe shoe. the pattern of the tread and so on, canall be developed into investigations by differentgroups. Jones f t al. (1992) provide some usefulcase studies of ways of engaging pupils throughthe use of brainstorming techniques or stimulusactivities or events.

A further possibility is to have apparentlyunrelated investigations being done simultan-eously where a particular concept of evidence isthe common factor. Hence the lesson might focuson data interpretation or the significance ofrepeated readings. Having children present theirdata to the rest of the class who have not done thesame task, can be a very useful exercise which canpromote 'bridging' of procedural understanding(cf. CASE, Chapter 2). Clearly, there are prob-lems of organisation in this approach in thatmuch apparatus is necessary, but a further advan-tage in this approach is that it can emphasise topupils that acquiring procedural understanding isa significant part of science.

Prompt sheets

Various ways can be used to offer guidance to thepupils as to what is expected of them. The prompt


Your reportMake sure you think about what you were trying tofind out.Did you include:what you were trying to find out?what you altered?what you measured?how you made your investigation a fair test?how many measurements you made?what instruments you used to make your measurements?how accurate your measurements were (e.g. to thenearest millimetre)?a result table?a graph or a description of the pattern in words?what you found out?why your results could be believed by other people?a conclusion which matched your results?an explanation of your results using scientific fdeas? .

Fig. 7.1 A prompt sheet (Gott et al.. 1992)

sheet shown in Fig. 7.1 is one possibility. Theprompt sheet can also serve as a reminder to theteacher to emphasise to pupils the significance oftheir record as a scientific report for others whomay not have done the activity. It is thereforeimportant that the validity and reliability of theirevidence is clear in their write-ups. Prompt sheetscan be adapted to the needs of the children (e.g.according to age) and the focus of the lesson.

The danger of such prompt sheets is that theybecome much like the rigid format of the pastwhere each experiment had set headings such astitle, apparatus. method, results and conclusions.Nevertheless, the pros probably outweigh the consin that they can be used to establish and reinforcelearning objectives. They also familiarise the pupilswith the language used to describe learning objec-tives which is also the language of assessment. It isalso possible to use a variety of styles for differenttasks and year groups.

'I he role of the teacher in the investigation

Skilful and appropriate questioning is probablythe most important role for the teacher in investi-

105

gations. There has been a tendency to see investi-gations as a 'do-it-yourself' exercise for .pupilswith minimal or no teacher intervention. Indeed,the swing away from didactic to child-centredteaching may have further reinforced the idea thatintervention is a 'bad thing'. Investigations areindeed about empowering children to applyknowledge, but it is essential that the teacher actsas a skilful mediator in the process. Indeed.Foulds et al. (1992) suggest that investigationsrequire greater interaction. The teacher has aboveall to be flexible in meeting the needs of individualpupils. Revell (1993) suggests planning 'enablingquestions' before the lesson. These are questionsthat 'the teacher plans to ask in order to advancechildren's learning or reveal achievement of alearning objective'. Some examples of enablingquestions are shown in Table 7.4.

Foulds et al. (1992) point to the critical role ofthe timing of the intervention here:

[The eacher] must recognise when to intervene,and also when it is more appropriate not to inter-vene. Intervention too early means that pupilswill be unable to recognise the relevance of the

Table: 7.4 Examples from an enabling question planner(Revell, 1993)

Learning objectives Enabling questions

Interpret

Communicate

E aluate

What do your results mean?

Was your prediction right?

Is there a mathematical pattern inyour results?

Which is the best way to tellothers what you did and what youfound out'?

Do your results go best in a table/chart /graph?

What do your results mean?

What else could your resultsmean?

How can you tell that is whatyour results really mean'?


intervening discussion and suggestit s theintervention may be seen as direction, rather thansupport, removing ownership from the pupil

As the pupils collect data, they should be encour-aged by the teacher to reflect on their meaning sothat ongoing modifications can be made to theinvestigation:

'do you think this (the dependent variable) willdouble if that (the independent variable) is dou-bled?' can encourage reflection on the number,range and value of the data being collected, andhelp pupils extend the scope of their findings.

(Foulds et aL, 1992).

Foulds et a/. (1992) suggest that the teacher maytake on a variety of other less obvious roles through-out the investigation such as: manager in terms ofcontrolling behaviour, motivator and provider ofencouragement thus empowering the less confidentpupils to proceed with their own strategy.

Reporting-back session

After the practical part of the investigation is over,the discussion which follows and centres ongroups reporting-back is vital. The importance ofthis stage of the investigation is frequently under-estimated and on occasion bypassed altogether(admittedly sometimes because of the pressure oftime). The reporting-back session has to be carefully handled so that its learning outcomes areachieved. It is at this point that most concepts ofevidence take on their full impact because pupilsnow have to use their evidence to support theirconclusions. The notion of audience discussed inChapter 2 is particularly relevant here in encour-aging pupils to provide valid and reliable evidenceto support a conclusion.

Jones et at (1992) suggest that the reportingstage be followed by a 'consolidation' stage wherethe pupils are encouraged to use the informationgained to further their knowledge and understand-ing. This kind of reflective discussion, where thelearning outcomes of' the groups in the class areshared, can be very useful.

103

A note on resources and safety

Resources are a particular problem in investigativework A wide range of Instruments must be madeavailable, e.g. an adequate number of forcemeterswith different scales. If only the basic equipmentfor the investigation is displayed, then the elementof choosing the most appropriate instrument islost. While this might appear to be unrealistic intimes of economic restraint, the equipment forinvestigations is in the main relatively basic, themore complex and specialised equipment beingonly necessary at the higher levels.

The related issue of accessibility to equipmentis also important, enabling pupils to select easilyfrom a central store or trolley. Where it is feas-ible, a set of storage trays which are labelled withthe name of the investigation end which containthose resources which are particular to it (bags ofdifferent types of sugar, for instance) can be setup. A separate rack of labelled trays can be keptfor measuring equipment. For instance, a traylabelled measuring forces would contain a rangeof appropriate instruments.

Revell (1993) points out that an understandingof safety is part of science education, so that thechildren should be involved in assessing the risk ofthe investigation. Among other helpful practicalstrategies for overcoming the problem. Revell etal. suggest using a colour code to indicate thedegree of rie..k associated with an activity:

red activities: greater risk and a lot of super-vision requiredamber activities: medium risk, some supervisionreq ui redgreen activities: low risk with little direct super-vision necessary.

The authors suggest that the class can then beorganised so that, for instance, only one red activ-ity is going on, amber activities are positionednear the teacher and the rest of the class areengaged in green or non-practical activities.

Summary

Schemes of work are an essentiA element of ateacher's armoury. In the context of' procedural


understanding, a scheme of work has to cope withthe idea that the concepts of evidence, which arerelatively few in number compared to substant:veconcepts, reappear in many contexts. The task ofincorporating progression into the scheme, along-side progression in substantive concepts, is noteasy. The best that can be done is to highlight thesianposts the concepts of evidence that consti-tute the learning outcomes that defineprogression and then work in targeted investiga-tions and follow-up work where it is most sensibleto do so. And be ready for pupils who go back-wards in their use of tables and fair tests and so onwhen they hit a new idea like continuous vari-ables. Then those ideas have to be recycled toshow pupils that they still apply.

Another part of the armoury is a set of resourcessuch as prompt sheets and pupil self-assessmentrecords (see the next chapter). These in turn rely onan agreed language to describe such things as cate-goric or independent variables. We have made afew tentative suggestion.3 in this regard, but theissue is far from resolved due mainly to the fact thatinvestigative work has only just started to becomethe focus for development in Key Stage 4, where itbecomes far more important as pupils need acommon language to describe their understandings.

107

And yet another weapon must be the quality ofteacher intervention in the classroom. There is atendency to regard investigative work as being thesame as the discovery learning of the 1960s. Fromthis stance, the received wisdom would have beento allow pupils to discover things for themselveswith little if any intervention by the teacher. In faa,teacher intervention, and indeed straightforwardteaching, is an absolutely essential part of investiga-tive work. Pupils do need to try things out, but thenthey need to be brought to see that there are betterways of doing things; encouraged and led and, ifnecessary. driven to itn:-.:.ove their understanding.

References

Foulds, K., Gott. R. and Feasey, R. (1992). InvestigativeWork in Science. Durham, University of Durham.

Gott, R., Costello. H. and Foulds. K. (1992). Materialsto support the assessment of Scl . Draft document,Durham University.

Jones, A.T.. Simon, S.A.. Black. P.J., Fairbrother. R.W.and Watson, J.R. (1992). Open Work in Science:Development of Investigations in Schools. Hatfield,Association for Science Education.

Revell. M. (ed.) (1993). The Scl Book of Investigations.5-16. Northamptonshire Inspection and AdvisoryService. Northampton. Northamptonshire CountyCouncil.

CHAPTER 8

Assessment

Assessment can refer to the assessment of thepupil's potential ability or it can refer to attain-ment. It is the second of these, attainment, whichwill primarily concern us here. While attainmentdepends on underlying ability, it also depends onother factors such as motivation and teaching.

The purpose of the assessment of a pupil'sattainment might be:

formative: aimed at deciding what the next stepin a curriculum plan or the teacher's scheme ofwork should be for that individual or group; or&agnostic: aimed at determining lack of under-standing or any weak points, for any individualor group, so that remedial action can be taken; orsummative: aimed at providing a measure of thelevel of attainment which pupils have achieved.This measure, grade or mark may be requiredfor pupils to progress to the next stage of theireducation or to other institutions in the educa-tion system.

Formative and diagnostic assessments aim toassess in a detailed and thorough way the breadthand depth of the pupils' knowledge of a particularsubject area. Summative assessment, on thc otherhand, samples knowledge and understanding inthat only some parts of the subject area will beselected for testing. All of these types of as.iess-ment are assessments of. and for the benefit of,the individual.

In recent years, the A PU, NAEP (NationalAssessment of Educational Progress) in the USA

and other such organisations have organisedassessments which are in reality evaluations of thecurriculum and the school system rather than ofthe individual. These assessments are, of course,carried out through the testing of individual pupilsbut in other regards this form of assessment can bevery different. There is, for instance, no need toassess all pupils; a random sample is adequateprovided the sampling fran takes into accountthe smallest unit of interest to policy makers. Thatunit of interest could be, for example, a school, alocal education authority or a region, or it may bea particular group of' pupils. such as those of lowability, or boys. There is no need to insist that allpupils do the same 'test' provided the groups arematched. So one group could do a practical inves-tigation, while another matched group might do adifferent one, or a related skills circus.

The point we are making is that it is essential tobe clear about the purpose and nature of theassessment because this will determine not so muchthe test items themselves, but the structure andorganisation of thc testing and the way the result-ing scores are recorded, aggregated and used.

This chapter will deal with a variety of tech-niques of assessment linked to the type ofquantitative investigations used in the NCC'research. The ideas discussed here are notintended to be a complete catalogue for assess-ment. Rather, they are a menu from whichappropriate activities can be chosen, augmentedand refined for whichever of the above purposes isunder consideration. We shall consider first what

4051


it is we are trying to assess in investigations andthen move on to explore ways of carrying out theassessment.

What is to be assessed?

In the last chapter. we noted that while pupilsapply their conceptual understanding at particularpoints in the investi2ation, facts aril concepts are.arguably, more efficiently taught. and likewiseassessed. by means of more traditional methods.Investigations can and do allow opportunities forchildren to demonstrate their understanding ofscience concepts: we have argued throughout thisbook that investigative work involves both con-ceptual and procedural understanding. But usinginvestigations to assess conceptual understandingwould be to select the wrong tool for the job.What can be noted, in passing. is that much diag-nostic information about pupils' understandingsdoes emerge from the way they tackle an investi-gation. We quoted examples in Chapter 7 whichshow how implicit understandings can surfacefrom investigations: insights such as the idea thatheat can be packed into a hard-boiled egg. which

may be unobtainable from more traditional test-ing methods.

The thrust of our argument has been thatinvestigations are a unique opportunity in practi-cal science to teach procedural understanding andthat this understanding has its own 'content', ofskills and concepts of evidence which have to betaught. It follows that assessment must be basedaround this selfsame content. So we cannot beginwith the assessment of investigations as if that werean end in itself. Rather, we must begin with thethings that pupils should 'know. understand andbe able to do' in the area of procedural under-standing. This distinction is an important one inthat it now allows us to see investigations as oneform of assessment among many. Before we moveon to look at this, it may be useful to set tradi-tional forms of examining within this view ofprocedural understanding.

Traditional assessment

Evamining practical work

When we look at most examination syllabuses forscience, the influence of Bloom's taxonomy can beseen clearly. The Northern Examining Association's

Table 8.1 The specific assessment objectives for 'Experimental work' (NEA, 1992. 1993) and their main emphases

The examination will test:

1 Experimental work

Candidates should be able to:

1 follow instructions for practical work:2 select appropriate apparatus:3 handle and manipulate chemical apparatus and

material...safely:4 make accurate observations and measurements.

being aware of the possible sources of error:5 record accurately and clearly the results of

experiments:6 draw conclusions id make generalisations from

experiments:7 plan and organise experimental investigations to test

hypotheses.

Comment

If the assessment is made of each point separately, thcemphasis is likely to be very firmly on skills and, to alesser extent, isolated concepts of evidence. Thecontext may be a skills circus or one based inillustrative practicals where pupils arc asked to drawconclusions, having carried out a recipe type activity.Only point 7 will result in pupils having to apply andsynthesise skills and concepts of evidence.

If. on the other hand, the assessment is made in thecontext of investigative work, which is quite possiblealthough unusual, most of the points will be interlinkedand shift the emphasis somewhat more towardsconcepts of evidence.

What is clear, however, is that only some concepts ofevidence are addressed explicitly.

ASSESSMENT

(NEA) syllabus for the 1992 and 1993 examinationsin chemistry, for example, has four assessmentobjectives, three of which are principally concernedwith substantive concepts (2, 3 and 4):

1 Experimental work2 Knowledge and recall3 Understanding4 Application, analysis, synthesis and evaluation

It is interesting to note here how experimentalwork has been detached from the other threeobjectives as if it were a separate entity. We canlook more closely at this breakdown in the light ofour definition of procedural understanding. Table8.1 shows how the NEA defines 'Experimentalwork' and how it might address aspects of pro-cedural understanding.

Non-practical examining

A modular science scheme by the NEA included anassessment of data interpretation (for example.Scheme A Single Award /First Award module,1989). These written papers were innovative in thesense that the pupils were not required to bring anyparticular science knowledge to the exam. Instead.

111

they were given accessible concepts and contexts inwhich to show their understanding of data inter-pretation. An example is given in Table 8.2.

The examples cited demonstrate how differentexamination boards and schemes met the NationalCriteria for Science which were developed in the1980s prior to the introduction of the NationalCurriculum. What was emerging, tentatively e ndrather patchily, was a more differentiated form ofassessment which was creating, and thereforeenhancing, the importance of what we have calledskills and concepts of evidence in more open-endedtypes of practical work. The framework was notparticularly well-articulated, nor were terms partic-ularly well-defined, but the seeds had been sown.We shall see in the next chapter how these seedswere transplanted into the National Curriculum.

Assessing procedural understanding

We have already suggested that the proceduraltaxonomy, linked to a definition of skills and con-cepts of evidence, allows for a more focusedapproach in the teaching of investigative work.Clearly, those same ith.as. constituting as they do

Table 8.2 Example of an examination question on data interpretation: NEA Science (Modular) Paper I. Scheme A -Single Award/First Award 1989

A small house was built about 50 years ago. It costs £500 a year to heat the house.

The table below shows how much money could be saved by improving the housc in various ways.

Improrentent Cost Saving on .fitel each year Time taken to pay for improvement

Insulating spaces between walls £300 £60 5 years

Insulating loft . £300 £100

Draught proofing doors and windows £25 £50

Double glazing £1500 £50

(a) Which improvement saves most money on fuel each year?

(b) Complete the last column in till; table ahove.

(c) From the information given in the table, suggest which improvement should he madefirst.

Give a reason for your answer.

Cri


a 'content' for procedural understanding. can actas a set of assessment objectives For complete-ness, we reproduce the taxonomy here

Procedural taxonomy

Knowledge and recall of skillsUnderstanding of concepts of evidenceApplication of concepts of evidenceSynthesis of concepts of evidence

Assessments can be produced for eaeh of theelements of the taxonomy, which we shah considerin turn.

Assessing the recall and use of skills

The definition of skills we have adopted is to someextent experience-dependent. So the first task inassessing skills is to decide what activities areappropriate for the age range, ability and back-ground of the pupils. Having done that. a choicemust be made as to whether the assessment is tobe embedded in an illustrative experiment or aninvestigation. We sl ,I1 deal first with the assess-ment of the skill in isolation. The assessment taskthen must clearly focus on the appropriate skill(s)

Table 8.3 A measuring skills exercise (from Gott et al.1988).

I 'sing instrunwntk

In front of you are a number of measuring instrumentswhich you will need to use to answer these questions.Do not forget to say what UNITS you are using.

What is the temperature of this laboratory?

The temperature is

flow long does it take for the ball bearing to roll downthe track?

The time is

How hem y is the ball hearing you have just used?

The mass is

I 0 5

with as little distraction as possible An example ofa circus practical where several measurement skillsare tested is given in Table 8 3

The pupils are given one of each type of measur-ing instrument a suitable thermometer, stopwatchand scales so :hat the choice is one of selecting thethermometer to measure temperature rather than ofchoosing one of several thermometers. The assess-ment is, therefore, primarily of skills and only to alimited extent of concepts of evidence (pupilsmight. for instance, show understanding of appro-priate accuracy and repeatability). Other examplessuitable for assessment purposes can be found in anumber of published texts (e.g. Coles et al.. 1988,1989: Gott et al., 1991).

We would argue that such assessment shouldbe used for diagnostic purposes since skilldeficits may have unlooked-for. and possiblyunrecognised, consequences for more complexassessments. For instance, an assessment inwhich pupils are required to weigh an item, maygive quite the wrong impression of the pupil'sattainment if he or she avoids weighing simplybecause of ignorance as to how to use the scales.Similarly. Strang (1990). reporting the work ofthe APU. showed how pupils often read individ-ual divisions on a particular scale incorrectly sothat all subsequent readings are inaccurate.

Skills assessment. therefore. can be diagnos-tic in that common errors can be spotted. It mayalso be used for summative assessment. such asin the circus practical (Table 8.3) where a 'markscheme' can be implemented by simply countingthe number of skills a pupil can accomplishsuccessfully.

Such marking can be applied, in theory, toassessing the same skill in the context of an illus-trative or investigative task. The danger of sodoing is that this element of thc assessment.because it is relatively easy to carry out, takesover from the more important assessment ofunderstanding and application which is moretime-consuming and requires judgement ratherthan relying mechanistically on 'bits' of evidence.Diagnostic assessment of skills at an early stage isnevertheless essential if basic errors in. for ex-ample. using equipment, are to be avoided.

ASSESSMENT

Assessing the understanding and application ofindividual concepts of evidence in shortassessment tasks

We argued in the last chapter that the teaching ofsome concepts of evidence in isolation is just notpossible because the ideas are so enmeshed in thewhole investigation that any attempt to isolatethem ends up with an unwieldy and, probably,invalid activity. Nevertheless, some concepts ofevidence can be isolated. Examples of those asso-ciated with the concepts of fair testing, choice ofinstrument, data handling and interpretation weregiven in the last chapter. These exercises can, ofcourse, be used for summative, formative or diag-nostic assessment.

Assessing the synthesis of concepts of evidence

Putting together ideas about how evidence can begenerated, interpreted and evaluated can, intheory, be assessed using a variety of activities.Investigations are, we would argue, the principaltype of practical science activity in which pupilsare given the opportunity to synthesise skills andconcepts of evidence into an overall strategy. Weintend to discuss only the sorts of investigationdescribed in earlier chapters because those are theonly ones for which data on pupil performance areavailable in a form appropriate to the discussion.

Using investigations for assessment purposes

Choosing the investigation

In Chapter 6. we discussed the issue of' progressionin investigative work and its link to the complexityand sophistication of the data in the light of theresearch evidence presented in Chapters 4 and 5.The arguments raised there are relevant to thechoice of task, since it is important to achievesome sort of match between task type and struc-ture and the attainment of any individual pupil orgroup. The choice of an investigation also requiresa not inconsiderable degree of experience to allowfor anticipation of the wide range of ideas thatpupils come up with.

113

The research evidence, however, points tosome key factors which help to govern the choiceof investigations:

Substantive concepts

In Chapter 4, we saw thc strong but sometimes unex-pected effect that concepts have on performance. Interms of assessing procedural understanding, it fol-lows that it is essential to use several different conceptareas that are at a level well within the grasp of thepupils. If the concepts are too hard or unfamiliar,their effect can be to block the pupils at an earlystage of the investigation. Several tasks allow for thefact that pupils find some tasks easier. or more moti-vating. or both, and perform better.

The :ontext

The research evidence shows that, if the context is'everyday', then children can regress in their per-formance. For assessment purposes, it mighttherefore be tempting to use only scientific con-texts. Our suggestion is that if everyday contextsare used regularly in the normal teaching situationinterspersed with scientific contexts, then childrenwill begin to relate science to everyday situationsand perform as well in either context.

Procedural complexity

Differentiation in assessment can be by outcomeor by task. Differentiation by outcome relies onthe use of a task which allows all pupils to makeprogress, but encourages the more able to carryout a more sophisticated investigation. By defini-tion almost, the task will need to be set in a moreopen form, which can cause problems as pupilsopting for the easy life carry out an investigationwell below their capacity. For instance, in theinvestigation 'Which is the best fucl?', able pupilsmay opt to design an investigation where theyinterpret thc dependent variable as 'the smokiest'and then proceed to measure qualitatively. Theymay well be capable of carrying out highly compe-tent work of the sort exemplified at the end of thelast chapter.


Alternatively, as we saw in the last chapter inthe case of the apple juice task, investigations canbe relatively easily manipulated in terms of proce-dural complexity on the basis of variable structurewithin one context. Thus, in a mixed-ability class,assessment can be carried out by using a series ofslightly different investigations of varying com-plexity appropriate to the ability of the pupils.Even within a more targeted task, there is an ele-ment of differentiation by outcome which willallow pupils who fail to rise to the challenge tocarry out a lower-level investigatio,..

Openness

The effect of the openness of the task definition wasto depress performance because children tended toopt for 'easier' investigations and regress to qualita-tive comparisons, although encouragement fromthe teacher or prompt sheets may help here.

It is clear that from the point of view of assess-ment of procedut.al understanding, the favourite,in terms of manageability, is an in), estigationwhich is defined relatively tightly but which is thenquite open in terms of method and solution. Thiswill allow for a degree of control over predictionswhile retaining the key elements of allowing pupilsto select their own method and put together theirown solution.

The importance of question wording

The wording or language of the task ir alsoimportant. Obviously, the language must be at alevel easily understandable by the pupils, particu-larly in an assessment situation. There is noreason, of course, why a written form of the taskshould not be supplemented by an oral introduc-tion from the teacher.

The wording is also important in determiningthe choice of design. For example in the investiga-tion 'Does the temperature of the water affect howquickly thc sugar dissolves?', it could be arguedthat testing hot and cold water is sufficient toanswer the question. Changing the wording to'How does the temperature of the water affect thetime the sugar takes to dissolve ?' might convey to

the expert the idea that a relationship is lookedfor, but it is a very subtle point for pupils whichmay easily be missed. Various techniques can beused to encourage pupils to explore relationshipsmore fully. An example for younger secondarypupils is given in Fig. 8.1. The competition at theend serves to focus the pupils' attention on theneed for some method of prediction, based on therelationship between the variables, which willshow how far the bottle will go. A line graph isjust such a predictor.

The last point concerning wording is to dowith the pupil's written account. We will see in thenext section that these accounts are valuablepieces of evidence, so we should endeavour tomake them as complete as possible. Pupils can beencouraged to give fuller accounts by giving thema target audience to write for, such as a relative, ora friend in a different class, or the environmentalhealth inspector. Occasionally, this technique canbackfire unexpectedly with pupils writing in unsci-entific terms; they need constantly to be remindedthat they are to present scientific evidence whichwill convince someone else.

Class 3 were running a Cola bar in aidof a local charity. They set up a

catapault to shoot Cola bottles downthe bar to the customer.

This is the question you have to find an answer to:

How does the distance thebottle travels depend on howfar the elastic is pulled back?

When you have finished, your teacher will tall youwhere the first customer will sit.

You have to be able to shoot the Cola bottleetraight to them. No trial runs!

Prizes for the nearest.

Fig. 8.1 The Cola bottle investigation (Gott, 1993)

ASSESSMENT

In summary, then, we suggest that the followingbe considered as criteria for selecting investigationsfor assessment purposes:

three or four tasks in different concept areaswhich have been taught and are, in general, rea-sonably well understood by the pupils;set in a mix of everyday and scientific contexts:differentiated by task, with some pupils beingasked to carry out type I investigations, otherstype 2 and so on;with a well-defined question but choice ofmethod and solution;an oral introduction to set the scene and to sup-plement the written task;having a clearly stated purpose;a target audience for the pupils' reports whichencourages as complete a description as possible;access to a wide range of approptiate resourceswhere possible so that ideally pupils will be ableto select an appropriat instrumertt;accompanied by constant, but neutral, encour-agement from the teach:s.

I-laving selected the investigations, the next issueconcerns the collection of evidence on which tobase the assessment. It is all too easy for the ses-sion to become dominated by checklists to theexclusion of judgement. It is to the management ofthe assessment session that we must now turn.

Collecting evidence for assessment

The assessment of investigations can be donethrough a combination of observation of the pupilsIn action' by the teacher, children's writtenaccounts and, if possible, through questioning ofpupils. In addition, children's self-assessmentrecords can be used to inform the overall assess-ment. There is also the pragmatic issue of assessingchildren while they are working in groups.

Group work

While tra iition dictates that pupils should beassessed inuividually in exam conditions (for sum-mative reporting at any rate), the realities ofpractical work in the classroom are likely to make

115

such an approach unreasonably time-consumingand disruptive. Pupils working in a group inevitablywork differently from any one of them alone, that isthe nature of communal effort. But when pupils areasked to write up their work individually, theirunderstanding or lack of it can show up starkly.Since we are assessing the understanding of conceptsof evidence, then these written reports are vital.Th .t is not to say that other foims of assessment arenot necessary. We shall see in the following sectionshow different assessment methods can all contributeto an overview of pupil performance. Experienceshows that professional judgement over a series oftasks will allow for assessment of an individual inthe context of group work. What cannot be said asyet is the degree to which working in a group influ-ences an individual's understanding of the task.

Observation

The observation of practical work by the teachercan be carried out in two ways. The teacher cansimply watch groups of pupils and come to ajudgement. Or, the teacher can use a standardisedchecklist in an attempt to gain objectivity.Checklists, or the 'tick y box' approach as oneteacher in a local school calls it, can be cumber-some and are an anathema to many professionalteachers. They also run the risk of teachers notseeing the wood for the trees: because they arefocusing on :he detail, they may lose sight of thequality of the investigation as a whole.

We will declare an interest here on behalf ofthe judges rather than the accountants. Webelieve that the 'stand back' judgement is a betterreflection of the synthesis which we believe to bethe key to investigative work. Our arguments inthis regard are not based on any reliable empiricalevidence but on experience of teachers engaged inthe task.

But there is a proviso. We believe that a mentalchecklist is needed to inform that judgement.Experience suggests that, in practice, it might wellbe necessary to use the checklist approach, at leastinitially, until it becomes an automatic one whichwill inform, but not be the only component of, the

iii


overall judgement. The checklist is also a means ofassessing the understanding and application ofindividual, or a particular subset of, concepts ofevidence such as those associated with data hand-ling or design.

The complex checklists used by the APU andby the NCC project were designed for researchpurposes and so are not necessarily useful insituations where teacher judgement is, quite

rightly, an important element. To provide evi-dence about procedural understanding and toinform the judgement which has to be madeagainst the criteria, we suggest that a muchshorter checklist based on skills and concepts ofevidence should be used and, further, that thechecklist should be general rather than specific.An example in use in a local school at the time ofwriting is shown in Fig. 8.2.

Practical assessment Date:

Year:

Investigation title:

Class:

Group Pupils:

A

B

C

D

E

Dependentvariable

What are theymeasuring?

Measuringinstruments

What have theychosen?

Are they suitable?

Range andaccuracy

Range ofreadings? How

accurate?

Independentvariable

What are theychanging?

Controlvariables

What are theykeeping thesame? Is it a

fair test?

Group A

Group B

Group Cetc.

Fig. 8.2 Assessment sheet

ii

ASSESSMENT

Experience suggests that checklists such as thisare most appropriate to the teacher who is learn-ing how to make judgements by highlighting thecriteria against which those judgements will ulti-mately be made. Eventually, as the teacher gets toknow the class and becomes more confident inboth the teaching and the assessment of 'nvestiga-tive work, the list becomes less important and maybe used only to ecord unexpected happenings.

Pupils' written work

Much information can be gleaned from the pupils'written work. The use of prompt sheets (see Fig.7.1) can also aid the assessment because pupils aremore likely to record the salient information.However, not all of the pupils' written informa-tion can be relied upon. This is not necessarily

117

because pupils wish to cheat. They often fail towrite down what they see as trivially obviousthings which may be revealed when they are askedto describe what they have done. More seriously.evidence shows that they can claim, in good faith,to have controlled a variable while in realityhaving failed to do so. The misunderstanding ofthe notion of the fair test (as in Jonathan's report;see p. 104), for instance, can also easily misleadthe assessor unless supported by observation. It ishere, perhaps, that the checklist in Fig. 8.2 servesits most useful function.

An example of a pupil's written report(Melanie's group) serves to illustrate the extent towhich evidence can be accumulated. The pupils hadalready been introduced to the particulate nature ofmatter. The task was set in the context of a rapidturnover drinks stall, but the teacher. in an oral

Melanie's group

Experiment: To find out which sugar dissolves fastest.

If 'hat we need: Caster, granulated and demerara sugar. warm water, stopclock and stirring spoon.

H Izat we did: We measured 4 ounces of each sugar. put them in a beaker with 100 ml of water and stirred themuntil they dissolved and put the results in a table.

What we kept the same:

1 The amount of sugar2 The amount of water3 The amount of stirring

Results: 500

4000-o

300

cr) 200

100

0

Sugar Time

Caster

Granulated

Demerara

1.35

4.34

7.02

If ha/ we f(1und out:

We found out that caster sugar was the fastest to dissolv

Caster

LI Granulated

0 Demerara

41. 1t)


Sarah's group (aged 11)

PlanI am going to find out if the material of the cupaffects how fast it cools down.

I think the cup which is thick will hold the mostheat because there will be more material for it so itwouldn't be able to escape through the side.

I am going to change the different material of thecup and take the temperature after they have cooleddown to see which is hotter.

Type of cup Amount of Temp. Temp.chocolate when hot when cold

Polystyrene 45 90 60

Plastic 45 90 58

ReportI got all my equipment together and put 4 spatulasof hot chocolate in 2 kinds of cups which were poly-styrene and plastic. I put the boiling water in thecups and took the temperature which was 90°C inboth of them. Later when they were cooling downwe took the temperature again.

The heat was a bit hotter in the polystyrene cup thanin the plastic one because the polystyrene is thicker.It took quite a long time to cool to the temperaturewe got when it was cooling.

90

80

70

60

50

40

30

20

10

0cnc b-ED. 0

Kinds of cups

Tentative ideas of energy transfer ( heat loss ) andinsulation are being used to come to a predictionthough they are not clearly formed.

She changes the type of eup though only pikstwo materials out of a range of four available. Shemeasures the temperature with a thermometer butdoes not measure the time at all.

At .first sight it appears that Sarah has carried outa fair test. She certainly knows that she should bykeeping the temperature the same at the start, and theanzount of chocolate powder used. However, she didnot measure the amount of water in the cups and itwas in fact different. She has not actually carried outa fair test although she recognises one.

Results are presented clearly in a table and used tocome to a general statement linked back to the originalquestion. She displays her results in a bar chart andrealises that the final temperatures are not much different.

Discussing this with her, she realised that she hadnot put the same amount of water in each cup and thatthis might help to explain why she had not really got areliable difference in her results. Fair testing is not justabout keeping 'something' the same but knowing theimportance of controlling certain variables.

ASSESSMENT

introduction, stressed the importance of providingscientific evidence on which to base a judgement.

Melanie's written report shows that she has aclear understanding of design; she has selected thekey independent and dependent variables andclearly stated how she controlled the relevant con-trol variables. The measurements appear to beaccurate, although this is best confirmed by obser-vation. The scale of her design is appropriate, butshe does not mention repeating any measurements.Her presentation of results is sensible, but she doesnot relate dissolving time directly and explicitly tothe underlying concept of grain size, which was theteaching context in which the investigation arose.

Another example, with a commentary and in adifferent context, shows how a pupil's writtenrecord can be misleading.

Planning vs per/brining

Can procedural understanding. as deployed ininvestigations, be assessed using tasks in whichpupils are asked to plan an investigation ratherthan actually do it? Clearly. the answer has to be'yes'. Any plan for an investigation must haveconsidered the relevant elements of skills and pro-cedural understanding. But it is important torealise that it is quite a different kind of assess-ment. Some evidence from APU work illustratesthis difference clearly.

The APU gave pupils investigations in two writ-ten forms one in prose and the other usingpictures. They also asked other pupils to actually

1101

carry out the same investigation. The results showedthat the pupils were far more successful on the prac-tical than on the written questions (Table 8.4).

They were also more successful with pictorialclues than with the question in prose alone. TheAPU also asked the same pupils to do similartasks in a written and practical form. The previousresults were confirmed. It follows that if pupils aresuccessful in the written question, then they arelikely to be successful in the practical investiga-tion. The main point here is, however, thatchildren who do not write adequate plans can stillperform well in the practical situation.

Strang et al. (1991), following Gott andMurphy (1987), explain this striking difference inperformance as due to the interaction with theapparatus in the practicaj situation. In the writ-ten form, pupils have to imagine the apparatusand design the investigation in the abstract.There is also little scope for trial and error. In thepractical situation, pupils can revise and changetheir plans as they proceed. This does not implythat written plans are of no value. Indeed. Stranget al. (1992) suggest that written plans can be auseful starting point for discussion of strategiesfor investigations.

Questioning

Questioning will give vital diagnostic and summa-tive information at all points. In the assessmentsituation, the questioning of all groups may seemimpractical but can be limited to a few carefully

Table 8.4 Examples of practical vs written responses to an investigative task (Strang et al, 1991)

W'hkhfabrk would keep you warmer?

pupils

Prose Picrorial Practical

Control of variables:

Use of measuring instrument:

Volume of water

Conditions of cooling

Stopclock

No measurements made

6

21

54

9

44

18

31

51

60

So

17

.1

4 -


focused questions to test the understanding of par-ticular concepts of evidence, which is not alwaysapparent either in the observation or in writtenaccounts. Probing pupils' understanding of particu-lar ideas can be carried out using shorter follow-up

assessments. An example focusing on data interpre-tation including the concepts of evidence ofpatterns and a fair test is shown in Fig. 8.3. Thisinformation is likely to be diagnostic and forma-tive rather than summative.

Dissolving data interpretation

Karl and Lee

Karl and Lee did the same investigation as you. This is what Lee wrote:First of all we set everything up and we had a beaker with water.We put 3 heaped spatulas of sugar in a beaker and timed it.

Sugar

white

brown

icing sugar

Time to dissolve

65 sec

84 sec

65 sec

We found out that icing sugar went fastest because the bits of brown sugar were bigger.

I Is there anything wrong with Lee's conclusion? What is wrong?

2 Do you think they carried out a fair test? Why do you think that'.'

3 What do you think they really found out?

Susan and Vicky

This is how Susan and Vick presented their r.!sults

50

40(1)

E 30I=

20

10

0

11111 Brown

111 White

o Icing

This is how Susan v, rote her report:

We set up our materials and put 200 nil of water in a beaker.We tipped the sugar in. We had 2 ml of sugar. We started tostir. We were timing while stirring. When all the sugar haddissolved, we stopped the clock. The brown sugar tooklonger. Brown sugar is harder than icing sugar. That's why ittook longer.

4 Do you think Susan and Vicky carried out a fair test? Why do you think that?

5 Who do you think wrote the best description of v, hat they did. Lee or Susan? Why do you think that?

Fig. 8.3 A follow-up exercise on data interpretation ( (Jott et al. 1993)

ASSESSMENT

Pupil self-assessment

The idea of pupils assessing themselves raises thespectre of pupils awarding themselves at least asmany marks as their closest friend, and probablyone more. The reality is very different. In schoolswhere this approach has been tried, it has beenfound that most pupils are their own greatest crit-ics. They are unreliable as sources of evidence.

But the notion that mark schemes should besecret does not equate with the idea that pupilshave a right to know and indeed should knowwhat is expected of them rather than having toguess what is in the assessor's mind. Making theassessment targets clear, or as clear as possible,through a self-assessment sheet, has the advantageof providing an agenda for instant feedback, andhence converts the assessment into a learning situ-

Assessment sheet

Investigation title:My level is (good, very good) because I haveshown that I can:

decide what to change

decide what to measure

decide what to keep the same

choose the best measuring instruments

take measurements over a wide range

repeat measurements for accuracy

carry out a fair test

record my results clearly in a table

draw and label a graph

get information from my graph

describe patterns in my results

explain my conclusion using scientificknowledge

EVALUATE how my results are useful, onlyif I have done a fair test

-suggest DIFFERENT interpretations of myresults

describe step by step what I did..aiiiiraira

121

ation as well. Self-assessment sheets can also serveas 'prompt' sheets for pupils when they are writingreports of their work.

An example of a self-assessment sheet whichfocuses on concepts of data handling and evaluationis given here. It is specifically designed to encouragepupils to use their results rather than spending toolong saying what they did: an altogether lessdemanding exercise and, hence, more popular!

It is clear that these assessment sheets can bedesigned and used to focus on the relevant con-cepts of evidence which are being assessed by theteacher. For example, the assessment sheet couldbe more tightly focused on design by restricting itsheadings, if that is to be the focus of a particulartask. The teacher must also be aware of the diffi-culty of language that we have referred toelsewhere in this book, so that some time shouldbe spent on 'teaching' pupils what the statementsmean as well as how to use self-assessment, to givethem confidence in completing the sheets. Theexercise is intended to promote pupils' under-standing of their own progress as well as being auseful assessment tool for the teacher.

Making judgements

The overview

Whatever the form of assessment, the methods wehave described can contribute to an overview (Fig.8.4) which can be used to generate the 'raw data'.Each method of a'ssessment can contribute in differ-ent ways to the overview. The pupils' interpretationof the data and their ideas on the reliability of theirdata and on the validity of their investigations isoften not made explicit in their written accounts butmay be elicited by careful questioning. The self-assessment sheet can also serve as a guide whilepupils are writing their reports. Experience suggeststhat controls, scale, the choice of instrument andaccuracy are best assessed by observation. Assessingtheir ability to produce reliable and valid data is amore complex exercise. It can only be done by refer-ence to all the evidence. Valid data require thatpupils have identified the correct variables and hatthere are no serious systematic errors in their experi-

122

Teacherobservation Questic ling

Teacherchecklist

Pupil'sself-assessment

sheet


written record

Fig. 8.4 The informed overview of the assessment ofinvestigations

mental design. Reliability depends on accuracy andrepeatability, which in turn depends on the choiceof instrument and the accuracy with which it isused. the number of readings taken and so on.

Table 8.5 suggests the principal source of evi-dence for each of the concepts.

Table 8.5 Principal sources of evidence

Norm-referenced and criterion-referenced assessment

Before we can look at how the information canbe used to inform judgements, we must take abrief sidestep into the issue of norm- and crite-rion-referencing in the context of summativeassessment. In planning an assessment we mustconsider what sort of information is requiredfrom that assessment. We shall consider twobasic ways of dealing with the accumulated dataabout pupil performance.

The first is based on the addition, somehow, of'marks' for any particular action that was 'correct'.This method resembles the task and interpretationscores described on pp. 55.-6. By adding thesemarks in various ways. particular strengths andweaknesses - in broad terms at least can berevealed. Thus. for instance, if we wanted to see ifa class had grasped the concepts associated withthe design of investigations, we might assess theperformance of children by adding the number ofpoints that they achieved in the design section

Observation Written work Questioningsupported byMeckliss andself-assessmentsheet

Design Identified key variables as independent anddependent variables correctly

Appropriate variables controlled for fair testVariables identified as categoric or continuous

(as appropriate)

Measurement Scale quantities chosen stmsiblyRange quantities spread over an appropriate

rangeInstruments chosen and used appropriatelyTo give suitable accuracyMeasurements repeated if necessaryAppropriate sample size (if relevant)

Data handling Table used to organise data collectionGraph chosen (or omitted) as appropriateData interpreted correctlyLink to reality included in interpretation r1

ASSESSMENT

only. This kind of assessment will give data suit-able for putting pupils in a 'pecking order', eitherfor the whole or parts of the investigation, the tra-ditional norm-referenced approach.

The second way, criterion-referenced assess-ment, involves using the checklists and pupils' workto make judgements against particular criteria. Theissue of criterion-referencing will be deferred untilthe next chapter where we look at the NationalCurriculum in the UK, which purports to be an'example of this method of assessment.

The understanding of individual concepts ofevidence within investigations

In a traditional norm-referenced assessment, the'domain' to be assessed is defined and a series oftasks set to sample that domain. The resultingscore is then taken to be representative of abilitywithin that domain. In our case, the domain isthat of procedural understanding. The moretasks used, across a greater spread of' conceptareas and contexts (both everyday vs scientificand open vs closed), the greater the validity andreliability of the score.

123

Using simple task scores

Table 8.6 gives a summary of the sort of data thatcould be collected for assessment. It would bedrawn, as we have argued earlier, from a varietyof sources: observation, pupil records and so on.It can be seen from the table that the additive taskscore is easily obtained. The disadvantage of thistype of score is that any failure at an early stagedoes not preclude getting a 'mark' at a later stage,so that two children obtaining the same score mayhave performed very differently and be able toapply quite different concepts of evidence. Thebreakdown of the task score into the broad cat-egories of design, measurement, data handling orevaluation can be more meaningful in pinpointingbroad areas of relative strength or weakness.Additive scores, however, take no account of howthe concepts of evidence relevant to the investiga-tion are put together.

Nobody would think it sensible to carry out adriving test by assessing, with the car stationaryand the engine off, whether the driver knowshow to drive by looking at the mechanics ofchanging gear, using the brake pedal and so on.

Table 8.6 Using the task score and its components in assessment

Overall view Assessment report

Design

Measurement

Data handling

Identified key variables as independentand dependent variabl correctly

Appropriate variables controlled forfair test

Appropriate sample size (if releant)Variables identified as categoric or

cont.nuous (as appropriate)

Scale quantities chosen sensiblyRange and interval appropriateInstruments chosen and used appropriatelyMeasurements repeatedTo give suitable accuracy

Table used to organise data collectionGraph chosen (or omitted) as appropriatePattern in data recognised.

interpreted and linked to reality

Task score

Design

Measurement

Data handling

9

3

3

124 INVESTIGATIVE WORK IN THE S IEN URRICULUM

What matters is, can he or she drive? If the driv-ing as a whole is judged to be imperfect, then itmay be useful to ascertain why by assessing indi-vidual components of the process. The key hereis the realisation that in a complex activity suchas a scientific investigation, the whole is fargreater than the sum of the parts, which leads usto the assessment of that whole the synthesis ofskills and concepts of evidence.

The synthesis of concepts of evidence withininvestigations

We have argued throughout this book from thestandpoint that the ability to understand con-cepts of evidence and hence be able to synthesisethem into a task solution is the ultimate goal ofinvestigative work. But how can this synthesis beassessed?

We have argued that the notion of evaluationand its related ideas of validity and reliability arethe key to all other concepts of evidence, sincethese concepts encompass all of them and whichdetermine the progress of the whole task. If theseideas of validity and reliability are understood,pupils will make sensible choices of variables, vari-able types, data requirements and interpretation.

To assess the synthesis of pupils' investigativework, teachers need to use the overview describedin Table 8.5. In the light of all this information.the teacher must then decide whether or not:

the question has been answered (validity)the data are 'good', e.g. have they been collectedin sufficient quantity and with appropriateaccuracy? (reliability)

This judgement, based on the extent to which anobjective observer would believe any conclu-sions, can be used as the basis for assessment.Quite how that could be done in a reliable andstraightforward fashion is a matter which canonly be tested in the classroom. Attempts to doso can be successful, hut there is a learning curveas indicated by this extract from a discussionwith a head of department:

I think that trying to pin down th4ssessment of'synthesis' is very, very difficult. Ll1ke in Englishyou can give separate marks for s ling, sentenceconstruction, paragraph usage and s on. Buthow do you give marks for imaginatio , unity ofstyle? Or in art, you can give separate tnarks fordraughtsmanship, use of colour or composition.But the overall piece of writing or the finished pic-ture is much more than this. It is a subjectivejudgement, but you can try to be more objectiveabout the bits. So what art teachers have alwaysdone is to decide for themselves, come to a consen-sus and then have their judgement moderated.

Unless and until all science moderators are will-ing to make judgements like this, science teacherswill be floundering because, unlike arty types, theylike things to be cut and dried. And what theyneed to see is that there is something else to sciencethat they have missed along the way as theybecome ever more logical and nit picking in theirtraining. There is that 'certain something' whichmakes an investigation better than others, and thatcomes from practice with checklists and things.There are certain indicators of quality a resultstable is one such, variables and fair test are others.and then, above all, the idea of believability.

No doubt there will be some who will feel suchan approach, itself, is not very scientific. It is allright, they may argue, for arts teachers to adoptthis approach relying, as it does, on teachersmaking judgements and then coming to a consen-sus on the judgements through discussion. But inscience, we must have evidence, preferably in thepupils' own writing. That, we would an ue. is toconfuse science with its assessment. Science relieson objective evidence: indeed, the thesis of thisbook is that such evidence is undervalued inschools. But assessment is not a science, evenwhen it is applied to science. It is an art.

But we fail to accord that art the status it merits.How often do we hear that teachers have under (orover) estimated pupils' ability compared to theexam'? This question assumes that the exam is cor-rect and that the teacher is wrong. Discrepanciesbetween teacher judgement and 'objective' externalassessment are far more likely to be examinationerror for individual pupils, or a bad day, or ques-tions that they didn't grasp properly. The comment

ASSESSMENT

should really say 'the exams have' under (or over)estimated pupils' ability compared to the judgementof the expert teacher'.

If we go down the line of a more subjective, butinformed, judgement, we must have some criteriaagainst which to make that judgement. The quotesabove were made in the context of the NationalCurriculum in the UK which is, in theory, cri-terion-referenced and will be treated as a casestudy in the next chapter.

Summary

In this chapter, we have suggested how investiga-tions can be used to assess procedural under-standing and how information from a variety ofsources can be collected prior to making the finalassessment. A key decision that has to be made byany assessor is whether that judgement is to concen-trate on individual concepts of evidence or theirsynthesis. The former allows for a more traditionalpattern of aggregating 'scores' in that marks foreach concept of evidence can simply be added. Thelatter relies on a more subjective approach which.while relying on the same evidence, seeks to standback and come to an overall judgement which

125 I

allows for the fact that an investigation is more thanthe sum of its parts. This, we believe, is more inkeeping with the spirit of investigative work.

References

Coles, M., Gott, R. and Thornley. T. (1988). ActiveScience /. London. Collins Educational.

Coles, M., Gott, R. and Thornley, T. (1989). ActiveScience 2. London, Collins Educational.

Gott, R. (1993). investigative Tasks for Assessment.Durham. University of Durham.

Gott, R. and Murphy, P. (1987). Assessing Investigationsat Ages 13 and 15. APU Science Report for TeachersNo. 9. London, HMSO.

Gott. R., Welford, G. and Foulds, K. (1988). Assessmentof Practical Work in S,ience. Oxford, Blackwell.

Gott, R., Price. G. and Thornley, T. (1991). ActiveScience 3. London, Collins Educational.

Gott, R., Costello. H. and Foulds, K. (1993). Materials tosupport the assessment of Sc 1 . Draft document.Durham University.

Northern Examining Association (1992. 1993). GCSEchemistry: Syllabus A. Manchester, NEA.

Strang. J. (1990). Measurement in School Science.Assessment Matters No. 2. London. SEAC/EMU.

Strang, J.. Daniels, S. and Bell, J. (1991). Planning andCarrying Out Investigation.s. Assessment MattersNo. 6. London. SEAC/EMU.

4 0

CHAPTER 9

Investigations and the UK National Curriculum

In the previous chapter, we looked at variousways of collecting evidence for assessment pur-poses and making judgements which reflectdifferent emphases. The UK National Curriculumpresents us with a different problem. It is premisedon a criterion-referenced system of assessmentwhich revolves around a set of criteria (theStatements of Attainment) within a curriculumdefined by the Programmes of Study (DES. 1991).The curriculum is essentially assessment-driven,which is why we are dealing with it in this chapter.

Criterion-referenced assessment, if it deals withtrivia, is relatively easy to operate (Gipps, 1992).For instance, whether or not a pupil takes a ther-mometer out of its case before using it could be acriterion. But it is hardly vital, except of course inthat particular and localised instance. If the cri-teria are to refer to something more generallysignificant. then they inevitably become morejudgemental in nature. The real question. then, inconstructing the National Curriculum, has beenthat of determining at what level of generalitythese criteria are to be and how judgements are tobe made against them.

The National Curriculum has included investi-gations in the curriculum since its formalinception in 1989. This official recognition of thesignificance of in\ estigative work in science edu-cation is an innovative move which has beeninfluenced, in part, by the work of the APU andlater by the recommendal ions of' the Task Groupon Assessment and Testing (TGAT) report(DES. 1988a).

The overall structure of the National Curriculumfor science has been modified considerably sinceit was introduced. By looking briefly at thesechanges. we shall see how the present structure ofthe National Curriculum has evolved and thethinking behind the definition and assessment ofthe investigative component.

The development of the National Curriculum

The National Curriculum which was proposed in1988 consisted of twenty-two 'Attainment Targets'(ATs) for science. These attainment targets remain.although they have been reduced in number. Theyare simply descriptions of things which pupils shouldknow, understand or be able to do. Each attainmenttarget is then divided into ten levels defined by'Statements of Attainment' (SoA), through whichpupils can progress from the ages of 5 to 16. Thesestatements of attainment are intended to act as thecriteria in a criterion-referenced system. The twenty-two attainment targets in the original version weregrouped together for the purposes of reporting, intofour profile components:

1 Knowledge and Understanding (ATs 1-16)2 Exploration and Investigation (ATs 17 and 18)3 Communication (ATs 19 and 20)4 Science in Action (ATs 21 and 22)

The National Curriculum was originally intendedfor formative as well as summative assessment lur-

122


poses. The very nature of the twenty-two attainmenttargets, with their very detailed criteria, was mostappropriate for formative assessment to guidefuture teaching and learning. Its early structure washeavily influenced by the work of the Assessment ofPerformance Unit (APU), which was charged withassessing the system through the pupils, rather thanassessing the pupils as individuals. The confusionbetween the assessment of a system, which requiresmerely that pupils are sampled, and assessing pupilssurnmatively resulted in fault lines. A systemdesigned to measure populations was used, indeedabused, to assess individuals. Readers are referredto Gipps (1992) for further discussion.

The assessment, then, came to be used in a sum-mative way because the assessment results were tobe summarised for publication by schools for theso-called 'league tables'. In this way, schools couldbe publicly compared in terms of thc attainmentof their pupils. So we have the paradoxical situ-ation of a criterion-referenced system, designedlargely for diagnostic and formative purposes andrelying heavily on teacher judgements, being askedto produce relatively simplistic norm-referencedinformation to put schools in a 'pecking order'.

We will look below at the development of thesecriteria in relation to investigative work. It will benoted that the original criteria have become drasti-cally reduced in number and, hence, increased intheir level of generality. They have, then, becomemore like descriptions of a domain than a set ofcriteria to be met.

If we look more closely, we can see the role ofinvestigations (Profile Component 1 as it thenwas) in this original version (Table 9.1).

It can be seen that Profile Component I was amixture of skills, concepts of evidence, substantiveconcepts, and observation. Profile Component 2,'Communication', also included some concepts ofevidence. Table 9.2 gives some examples.

Both these profile components include many ofthe concepts of evidence which we have suggestedare at the heart of procedural understanding, butthey are mixed up with other aspects of science.The message as to what, precisely, investigationswere to be used to assess was far from deal

The proposed structure consisting of twenty-

125

Table 9.1 Profile Component 1: Exploration andInvestigation (DES, 1988b)

Explore events and phenomena seeking regularitiesand noting the unexpectedFormulate hypotheses which can be tested experi-mentallyPlan and carry out investigations using apparatus,materials and methods appropriate to the problembeing investigatedMake and systematically record observations whichare relevant to the problem being investigatedRepresent experimental findings using graphs. tables,charts, symbols and conventions as appropriateSelect measuring instruments which are suitable to atask and use them to an appropriate level of accuracyRecognise variability and unreliability in measure-mentsMake inferences and justify them in the light of thedataEvaluate the design of experimentsUse a range of measuring instrumentsEstimate quantitiesFollow instructions in verbal and written formWork with an awareness of safety aspectsTreat living things with respect

Table 9.2 Profile Component 2: Communication(DES, 1988b)

Represent experimental findings using graphs, tables,charts, symbols and conventions as appropriateUse secondary sources, including the media, otherpeople, reference books, databases and select infor-mation relevant to a particular topic of studyTranslate information between graphical, tabular,pictorial and prose formsCommunicate information on a scientific topic toothers in written and oral formConsider alternative theories, hypotheses and models(including personal theories) and assess their claimsin relation to observations and other evidence

two attainment targets, each of which was to beassessed, was soon found to be too complex andunwieldy as the move from a sampling idea andteacher assessment was replaced by the requirement

INVESTIGATIONS AND THE UK NATIONAL CURRICULUM

to assess all pupils on all aspeOts of the Programmeof Study. In December 1988, a consultation reportrecommended reducing the attainment targets toseventeen, within two profile components:

Pr(!file component I Attainment Target 1Exploration of science, communication and theapplication of knowledge and understanding.

Prqile component 2 (Attainment Targets 2171Knowledge and understanding of science, com-munication. and the applications and implicationsof' science.

(NCC. 1988)

This recommendation was implemented in March1989. A further reduction was proposed in 1991(NCC, 1991) and implemented soon .after. Theattainment targets were now reduced to four andthe term 'profile components' subsequently aban-doned. This structure applies at the time ofwriting, although there are already signs that itmay change again.

The present National curriculum

The shrinking of the curriculum, together with thefact that teacher assessment has been reduced, hasmeant that it can no longer claim to be formativeassessment, at least in the terms originally defined.It is now akin to a traditional 'sampling knowl-edge' type examination and hence primarilysummative in nature. Its purpose has becomelargely one of accountability.

So how do investigations and procedural under-standing fit into 'today's National Curriculum? Thefour attainment targets currently are Investigations(Sel ), Life and living processes (Sc2). Materials andtheir properties (Sc3) and Physical processes (Se2).Attainment Target 1. scientific investigation, isdefined within the Programme of Study, as follows:

Pupils should develop intellectual and practicalskills which allow them to explore and investigatethe world of science and develop understanding ofscientific phenomena. the nature of theories andprocedures of scientific exploration and investiga-tion. This should take place through activities thatrequire a progressively moi e systematic and quan-tified approach which develops and draws on an

129

increasing knowledge and understanding of sci-ence. The activities should develop the ability toplan and carry out investigations in which pupils:

(if ask questions, predict and hypothesise(ii) observe, measure and manipulate variables(iii) interpret their results and evaluate scientific

evidence( DES. 1991)

These three components or 'strands' form thebasis of the assessment.

Strand (i) brings in the substantive concepts ofscience in the 'raising of questions' or hypotheses(Table 9.3). This poses a number of problems.Principally, of course, it represents an attempt toassess the substantive structure of science which isalready covered by Se2 4. Over and above that, itis very difficult for pupils if the hypothesis is seenas being the source of any investigation.Experience shows that it is hard enough for any ofus to generate questions suitable for investigationswhich will target skills and concepts of evidence inany sort of coherent fashion.

In Strand (ii), Table 9.3 shows that the lowerlevels are concerned with observation hut fromlevel 3 to level 9 concepts of evidence concernedwith design and measurement appear. Level 1(1includes evaluation. Strand (iii) is a conflation ofdata interpretation and evaluation of evidence andthe drawing of inferences based on that evidence(Table 9.3). The first locates within concepts ofevidence: the second, once again, moves into thesubstantive concepts of ATs 2 4.

How does this assessment relate to our view ofthe content of procedural understanding? It' wedraw out the concepts of evidence as they appearin the levels of the curriculum (Table 9.4), we cansee that while most are included, some such asrepeatability or sample size arc not mentionedspecifically in any of the strands

Like the original version. the overall picture ofassessment in Sel is rather confused. On the onehand, it seems to be assessing some concepts ofevidence, while on the other hand, it is aboutapplying substantive concepts. There appears tobe no clear underlying philosophy. It is basedaround the 'doing' of practical activities rather

INVES1IGAlIVE WORK IN I HE SCIENCE CURRICULUM

Table 9.3 Strands in Attainment Target I: Scientific investigation in the National Curriculum (DES. 1991)

Pupils should carry out investigations in which they:

Level Strand (i)Ask questions, predict andhypothesise

Strand (ii)Observe, measure and manipulatevariables

Strand (iii)Interpret their results and evaluatescientific esidence

Pupils should tarry out invettigatum., in which they:

(al ask questions such as 'hms...T and 'vvhat skill happen

if ...T. suggest ideas and makepredictions

(a) suggest questions, ideas andpredictions. based on everydayexperience, which can he tested

4 (al ask questions. suggest ideas andmake predictions. based on sonierelevant prior knovsledge, in aform vs hich can he investigated

tai formulate hypotheses where thecausal link is based on scientificknowledge. understanding ortheory

6 (al use scientific knots ledge. under-standing or theory to predictrelationships betvveen continuousvariables

7 (a) Me scientific knott ledge. under-standing or theory to predict therelative efko of a number ofvariables

8 Ia ) use scientific knovv ledge, under-standing or theory to generatequantitative predictions and astrategy for the investigation.

9 tat use a scientific theory to makequantitative predictions andorganise the collection of validand reliable data

(al use scientilk knovv ledge and anunderstanding of lay\ s. theoriesand models to develop Its potheses%%Inch seek to explain thebehaviour ol objects andevents they have studied

(al observe familiar materials and events.

(b) make a series of related observations

obsers e closely and quantify bymeasuring using appropriateinstruments

carry out a fair test in vvhich theyselect and use appropriateinstruments to measure quantitiessuch as olurne and temperature

choose the range of each o: thevariabks involved to producemeaningful results

consider the range of factors involved.identify the key ariahles and thoseto be controlled and /or takenaccount of and make qualitative orquantitative observations ins cls ingline discrimination

(b1 manipulate or take account ofthe relative effect of MO or moreindependent variables

select and use measuring instrunlents%cinch provide the degree of accuracycommensurate \soh the outcome theyhase predicted

(hi systematically use a range ofinvestigatory techniques to judge therelative Oleo of the factors in\ olved

rho collect data %%Inch are sufficientlyvalid and reliable to enable them tomake a critical evaluation id the lanc .dleor> or Model

tel use their observations to supportconclusions and compare Nhat theyhave observed %kith %hat they expected

(ci recognise that their conclusions maynot be valid unless a fair test has heencarried out

(d ) distinguish hemeen a description o!'\chat they observed and a simpleexplanation of hoc% and vshv ithappened

let drays conclusions cv Melt link patternsin observations or results to theoriginal question, prediction or idea

(ci evaluate the validity of theirconclusions by considering differentinterpretations of their experimentalevidence

tcl use their results to draw Lonclusions.explain the relationship betsseenvariables and refer to a model toexplain their results

I c use observations Or results to dray,conclusions \cinch state the relativeeffects of the independent variablesand explain the limitations of theevidence obtained

Ict justify each aspect of the investigationin terms of the contribution to theoverall conclusion

anidy se and Interpret the dataobtained. in terms ol complex functionsyv here appropriate. in a \kir>demonstrates an appreciation of theuncertainty of evidence and thetentative nature of conclusions

(ci use and analyse the data obtainedto evaluate the lass, theory or model inteims ol the extent to %%Inch it canexplain the observed behaviour

4 04 kJ

INVESTIGATIONS AND 1 HE UK NATIONAL CURRICULUM

Table 9.4 Concepts of evidence in the NationalCurriculum

Concepts ()I. evidence( not in any particular order,

Level at which theyappear ( in strand 2unless noted )

Identifying variables ifsindependent and dependent

Fair test and control variables

Categoric, discrete andcontinuous variables

Scale

Range

Accuracy

Choice of instrument

Repeatability

Variation, sample size and probability

Tables

Choice of graph type

Recognising and interpreting patterns

Interpreting multivariate data

Validity

Reliability

3 (strands 2 and 3)

3

3

5

6

4

4,5

7

3.9

9.10

than the understanding that underpins that activ-ity. It is no wonder, then, that its implementationhas been fraught with problems. Some teacherswho are unsure about assessing the 'content' ofprocedural understanding, focus on the assess-ment of the application of substantive concepts.with the result that the understanding of conceptsof evidence becomes marginalised. Others see it asno more than traditional practical work as before.

The issue of whether the assessment is aboutindividual concepts of evidence or their applicationand synthesis is also unclear. On the one hand, thecurriculum emphasises that assessment is to be car-ried out 'in the context of complete investigations'(SEAC. 1993). While on the other, the suggestionis that the three strands be assessed separatelywithin the context of whole investigations:

All of the abilities that pupils need to demonstratehave therefore to be assessed in the context ofwhole investigations. This does not mean thateach investigation has to offer the opportunity toassess all three strands. Nor does it necessarilymean that where an investigation allows allstrands to be assessed, the teacher has to assessthem all. It may be more manageable to record apupil's performance on each strand in the contextof different investigat;ons.

(SEAC. 1992)

So while there is clear guidance to use whole inves-tigations for assessment, confusion reigns overwhether to focus on individual concepts of evi-dence or their application and synthesis.

The research evidence and the present curriculum

Strand (i) asks pupils to generate their own hypothe-ses to test. Many teachers see this, not unreasonably,as the starting point. But there could hardly be amore open start. As a consequence, some pupils willask questions that are incapable of being answered atall, or certainly not with the apparatus available. Butmore importantly, they will tend to ask questions, thesolutions for which are already within their grasp. Aswe saw in the research, the move to open questionsled pupils to rcgress to a more qualitative solution.Such an approach is not going to bring out the bestin pupils, which surely we should be aiming for.

The emphasis on conceptual understandingacross strand (i) and strand (iii) to a lesserextent at the higher levels makes heavy demandson pupils. As we have seen in our research. theeffect of the concept area on performance is oneof the most influential factors. The assessment ofprocedural understanding will be, to all intentsand purposes, impossible if the biggest hurdle forpupils is conceptual. The existing NationalCurriculum attempts to structure progression onvariable complexity. There is a clear thread defin-ing the required independent variable(s), whichsuggests a progression from the type 1 investiga-tions (as defined in the research) to type 2 totypes 3 and 4. The research has demonstratedthat this order is consonant with assessing appli-cation and synthesis.

2I 0

INVES I IGATIVE WORK IN 1 HE SCIENCE CURRICULUM

At the same time, pupils are being asked toapply concepts at a level commensurate with thelevel of Sc2-4:

There should be approximate parity between thelevel at which pupils are ww king in Scl and thelevel of knowledge and understvnding in Sc2. 3 or4 needed to make appropriate predictions orhypotheses (strand i) and to make sense of theoutcomes (strand

(SEAC. 1993)

The requirement that the concepts should bethe same or at a similar level to that of the investi-gation is to ask pupils to apply newly acquiredunderstanding of substantive concepts and con-cepts of evidence. This fundamental problemremains. The research reported here, and theincreasing criticisms levelled by teachers, suggeststhat such an approach is untenable, elevating as itdoes the higher levels of Scl into what one teacherhas described as a PhD thesis. Sc I must refocusitself on the procedural understanding which setsit apart from Sc2- 4. As our much quoted head ofdepartment says:

Getting the pupils to raise a question themselves.writing perfect plans, relating their conclusions toscientific ! owledge and so on are all very worthythings to do. and I try to do them. But they get inthe way of pupils showing their procedural under-standing and should be assessed separately. Theyare the points at which teachers get bogged downand pupils get discouraged because they can't doall the hits perfectly and stay at one level. Theinsistence on [the substantive concepts being partof assessment as well] is hindering investigationstaking off properly. And if that looks as though Iam ignoring how concepts fit in with imestiga-tions. I am not. That is where the scheme or workis so important.

Add to that our often reiterated comment that theresearch, and the underpinning of the NationalCurriculum, is biased towards quantitative workand we begin to see that, as it stands, the structurehas its problems but is well worth building on.Time will be needed for it to develop. So we should

be looking for ways of improvement: two possibleways forward arc suggested below.

The ways forward

A structure based on the complexity andsophistication of evidence

Of prime importance in any revision is to recognisethat a wholesale rewriting is simply not possible atthis particular stage: the system above all needstime to settle in. But what is also quite clear is thata philosophy must be provided if some coherentprogression is to be built into the criteria. Onepossible way of structuring progression based ona 'content' of procedural understanding whichattempts to reflect the existing structure of the cur-riculum is shown in Table 9.5. We have used theexisting ten-level scale as our starting point: the cri-teria could be used in a number of different ways,one of which we shall explore in a later section. Wehave also tried to keep the levels as close as possibleto the existing ones, not because we believe them tobe necessarily correct, but for purely pragmatic rea-sons in that continual tinkering is causing teachersto lose that trust necessary for the investment oftime and effort in new schemes of work. It is alsoworth noting that the structure is, we suggest, onlya part of any revision which might well include ele-ments connected with skills, written assessmentsoased on data interpretation and the nature of evi-dence in science and engine.ring.

This structure (Table 9.5) is based on the com-plexity and sophistication of the evidence plannedfor, generated and interpreted within the task.There are a number of advantages to thisapproach: crucially. the basis is one of understand-ing, application and synthesis, rather than skills orany other sort of 'doing'. The assessment is then ofa complementary nature to conceptual understand-ing, revolving around conors of evidence, whichare the content descriptors of that understanding.

The complexity of the evidence relates to the type.number and complexity of the variable structures inan investigation. A task involving continuous vari-ables, for instance, gives both more data and more

12-1

INVESTIGATIONS AND THE UK NA'110NAL CURRICULUM

Table 9.5 Levels defined by the complexity andsophistication of the evidence

Level Key indicator

Within a mmplete task:

4 Carry out a complete task which will give datathat answer the original question

5 Recognise the importance of a fair test and ofthe scale and range of the values involved inensuring that the resulting data are 'believable'

6 Use continuous data to represent a pattern andthe relationship of that pattern (in a line graph)to scientific understanding

7 Disentangle the effect of more than one inde-pendent variable

8 Use patterns to predict what will happen andcollect data of appropriate scale, range andaccur; . y to allow the prediction to be checked

9 Synthesise evidence from a variety of sources(multiple investigations or a mixture of investi-gations and use of secondary sources) into acoherent argument and conclusion

10 Understand the interplay between theory andevidence, the use of evidence to test a theory orthe use of theory to check the reliability andvalidity of evidence

complex data as the need for range and accuracylinked to patterns comes into play. The sophistica-tion of that evidence depends on the concept(s) (andcontext) which underpins it. We have noted that, forinstance, the connection between a pattern in dataand the reality of the event that they represent is acrucial element of science.

The tasks themselves are not defined so they arenot limited by variable structures because it is theoutcome, in terms of evidence, that is defined.Tasks of other kinds which generate complex andsophisticated evidence can be used, such as simplequalitative analysis in chemistry where the logicalstructuring of the task based on the accumulationof evidence is paramount (Table 3.6 gives someexamples of other types of investigative work).

The key indicators of progression in Table 9.5can he seen as pegs along this path of increasing

133

complexity and sophistication (it is limited tolevels 4 and above, those most relevant to sec-ondary schools). It will be apparent that theyrepresent the pupils' ability to put together a strat-egy for the complete task and so are based, as weargued in the previous chapter, largely on ideas ofvalidity and reliability.

Using the criteria

How can these key indicators (Table 9.5) be relatedto the present National Curriculum? We suggestusing these key indicators as the initial criteriaagainst which to locate a pupil's investigation.'Teacher judgement is essential here. The existingcriteria at that level can then be used as supportingcriteria to assist in making a judgement as towhether the pupil has or has not met, more or less,the criteria related to and defining the level. Thismay seem a very subjective approach, but the searchfor objective criteria is bound to fail. Complexunderstandings are not capable of unambiguousdefinition when limited to a few statements.

In reality, this is criterion-related judgementrather than strict criterion-referencing and as such itmust be based on case law. Experience suggests that,given a set of criteria and some examples of pupils'work. judgements can be brought into linc andincrease teacher understanding of what is required.

Another point of some importance in using cri-teria of any sort, including our suggestions here. isthe degree to which the statements are to be inter-preted in a legalistic sense. In the existingassessment structure of the National Curriculum,pupils' work is judged against level I (for instance)and if it meets all those requirements it is judged atlevel 2. again compared to every single phrasewithin the criteria. And so on. Clearly, the effect iscumulative. In the end we find ourselves lookingfor evidence at level 10 that pupils (in one piece ofwork?) have met' every single criterion that hasgone before, all twenty-seven of them. And sonobody can succeed.

There is another way. We can look at a pupil'swork and match it against key indicators with aview to finding the hest overall match. Then the

2

134 INVESTIG ATIVE WORK IN 1 HE SCIENCE CURRICULUM

additional criteria can be used to make an overalljudgement as to whether the work is sufficientlyclose to that level, as a whole. This may seem like ahair-splitting exercise, but the change in stancefrom working upwards through all the words ofthe statements to a judgemental matching is not tobe underestimated. Level 10 then becomes some-thing which can be judged over a number of piecesof work using the criteria as canons of judgementrather than legalistically interpreted hurdles. Asthe head of department quoted before says:

I find having to obey the 'letter of the law' of theNC criteria discouraging. It needs putting right.Using judgements seems a revolutionary way ofassessing, but it is 'do-able'. You stand back fromthe pupil's work and you have the criteria (man-ageable not too many fine details) in your mindclearly before you start then you exercise yourjudgement. It is a clearer and more relaxed way ofassessment. The 'nit-picking' approach is essentialfor details of conceptual understanding. It is notthe way to judge procedural ability.

It will be noted that there is no reference toscience concepts (the substantive elements) in thesesuggested criteria. The reason is tied to the interplaybetween context and concepts which the researchshows plays such a big part in determining pupilperformance. The concepts and contexts are largelyset by the Programmes of Study defined for eachkey stage. These, quite rightly, differ from stage tostage. The consequence of that is, of course, that apupil in Key Stage 3 will be assessed on an investi-gation set in contexts defined by the Key Stage 3programmes of study and using concepts that arewithin their grasp. A pupil attempting tlw same levelin Key Stage 4 would come across a different set ofconcepts and contexts. The level of demand will bequite different. As the Dearing Report (1993) notes:

This [the problem of the 10 level scale] raises thequestion of whether. for example. the level 3 inhistory achieved by a bright Key Stage 1 pupilmeans the same thing as the level 3 achieved bythe less able 14-year-old [in relation to the com-pletely different programme of studyl.

The sensible thing, then, is to define levels withina key stage based in the concepts of that key stage.

rather than attempt a too grandiose and all-encom-passing set of criteria. One suggestion is shown inTable 9.6. The range covered and degree of overlapis an arbitrary choice. All that is required is arecognition that the assessment is to be of the abil-ity to generate. understand and use evidence in thecontext of investigations set within the contexts andconcepts defined by the Programme of Study. Thegrades. be they letters corresponding to GCSEgrades or labels such as average, above average andso on, will then be qualified by the key stage atwhich the assessment is made.

The Dearing Report (1993) suggests somethingsimilar: 'One possibility for the end of key stageassessment would be a five (or six ) point grad-ing in each of the Key Stages 1 to 3 leading on tothe General Certificate of Secondary Educationscale at Key Stage 4.'

In the introduction to this chapter, we sug-gested that the level of generality of criteria andthe problem of how judgements are to be made arethe major problems. Our suggestions above canonly be seen to deal with these questions if:

the criteria, which arc still very general. aregiven meaning through many examples or tasksand examples of pupils' work: andthe various ways of collecting evidence sug-gested in the previous chapter are used to formjudgements against those criteria.

Other strands

-1 he above amounts to a structure based on onefactor the idea of evidence. We regard this as thecentral feature of Scl. but a broadening of the def-inition to include skills, data interpretation and.potentially, the nature of scientific evidence mighthe considered. Certainly at Key Stage 4 there is aplace for more advanced skills that involve thcmore sophisticated instruments encountered atthat level. This would build on thc progress madein GCSE work within the National Criteria, whichwas proving popular before its demise at thehands of the National Curriculum. It would alsoprovide a vehicle for inclusion of tasks which areessentially based on 'finding a way' to measure


Table 9.6 A norm-referenced approach to levels

135

Current level as Key Stage I levels Key Stage 2 levels Key Stage 3 levels Key Stage 4 levels

suggested in defined within defined within defined within defined within

Table 9.5 Programmes of Programmes of Programmes of Programmes ofStudy at this Study at this Study at this Study at thiskey stage key stage key stage key stage

4

5

6

7

8

9

10

Table 9.7 Possible strands for a new structure for Sc I

Strand (a)Carrying out investigations

Strand (b)Skills

Strand (c)The nature of evidence in science andengineering

Designing tasks and collecting valid and reliable evidence

Use of apparatus and measuring instruments and the interpretation ofevidence from primary and secondary sources

Empirical evidence and its relationship to laws and theories and as apredictor of the behaviour of materials and systems

something. often a key element in science. Wehave seen examples in earlier chapters of activitiesbased on individual concepts of evidence: theN EA core skills paper is one example of thatapproach in practice. Such tasks could also formthe basis of a strand based directly on the interpre-tation of 'second-hand' data.

Inclusion of a strand linked to the nature of sci-ence itself has some attractions, particularly whenwe remind ourselves of the data presented inChapter 4, which showed the disparity in teachers'perceptions of practical work in science. Its suc-cessful incorporation would rely on it notbecoming confused with the history and philoso-phy of science, which tends to focus on historical

discoveries or ideas about inductivism or falsifica-tion. We mean nothing so grand for pupils at thisage. We do believe that thc current provision failsto prepare students, particularly those going intoengineering, to encounter empirical evidence whichdoes not have any formula to attach to it. Forexample, graphs in civil engineering often areenvelopes within which the material is likely tobehave. To a student of school physics who expectsa graph to 'prove' some law or other, this can be asignificant hurdle to learning. Whether it can beincorporated into a workable assessment system inan arca which we have shown still is on a steeplearning curve is open to doubt. The strands thenmight become as shown in Table 9.7.

4 -all

INVES1 IGATIVE WORK IN THE SCIENCE CURRICULUM

A structure based in a norm-referenced framework

A quite different alternative is to use the taxonom-ical structure in the way it is used in conventionalexamining practice. A typical examination plannerwill begin by creating a grid for assessment objec-tives and content. We give a much simplifiedversion in Table 9.8.

The grid will encompass all the assessment objec-tives and all the content areas. Then, rather thanattempt to cover all the boxes in the examination.questions are distributed around the grid to reflectthe overall weightings of the objectives. There may

be more questions in the understanding row. fewerin the application and fewer still in the recall. Thisgrid, then, represents the domain which is to beassessed and the results obtained from the samplingof that domain are used to give an estimate of apupil's ability in the subject defined by the domain.

Exactly the same approach can be adopted fora domain of procedural understanding. The gridnow contains 'concepts of evidence' rather thansubstantive concepts but in other respects the issueis identical (see Table 9.9).

Questions can now be distributed across thegrid, making sure that there is a spread of concept

Table 9.8 An assessment planning grid for (substantive) concepts

Assessment objectives Electricalcurrent

i'altage Energy Etc.

Knowledge and recall of skills

Understanding of concepts ofevidence

Application of skills and conceptsof evidence (in unfamiliar situations)

Synthesis of skills and concepts ofes idence (in problem-solving)

Table 9.9 An assessment planning grid for procedural understanding

.4ssessment objectives Skills Definevariables

Fair test Samplesize etc.

Knowledge and recall of skills ,,

l'nderstanding of concepts ofevidence

..

... I...................:......J

Application of skills arid conceptsor evidence (in unfamiliar situations)

S nthesis or skills and concepts ofe idence (in problem-solving)

131


areas. There will be some skill tasks, written andpractical, some short tasks based on individualconcepts of evidence and some practical investiga-tions. It will be remembered that a key feature ofinvestigations is their ability to differentiate byoutcome, so it will be possible to use relatively fewsuch tasks to cover a wide spread of ability. And,just as for traditional examining, marks andweightings can be allocated to each question and afinal total used to represent ability in that domain.We give this alternative as a technical examplerather than as something we would advocate.

GCSE assessment of Scl

As the final touches were being put to this manu-script, the suggestions for assessment of Scl forGCSE were dropping through school letterboxes.The method being adopted by the NorthernExamining Association (NEA) is of some interestin the light of the above arguments. They suggestthat assessment be carried out using whole investi-gations and that the level be defined in relation to'key features': 'An investigation needs to be of theright type according to these descriptions [the keyfeatures] if the mark for each skill [strand] is to beawarded' (NEA, 1993).

So the first step, they suggest, is to decide atwhat level the pupil's work is compared to the keyfeatures before deciding on what to do about anyone skill (strand). Then, the assessment calls,implicitly, for teachers to use their judgement: 'Itis now possible for these assessment criteria to beused in such a way that a candidate can beawarded a mark without necessarily having satis-fied each and every one of them in full.'

So, having decided on the overall level, theteacher is being asked to judge if each skill(strand) has been attained, whether or not everysingle element of that strand has been 'ticked'.Another suggestion made above, that pupils' workshould be matched against criteria by looking fora 'best overall match', is paralleled in the NEAdocuments: `...the criteria should be used to iden-tify best fit between the work and the criteria...provided the investigation is of the right type(defined by the key features)'.

137

The correspondence between these suggestionsand our analysis is close, which suggests that thereis a convergence in meaning as to what Scl isabout; the language is becoming established.

Summary

We have seen in this chapter how the NationalCurriculum has developed from a mixture of skillsand concepts of evidence towards somethingwhich concentrates more directly on thc synthesisof concepts of evidence. We have suggested thatthe insistence on the inclusion of substantive con-cepts with which p:apils are not necessarily athome is to miss the point of the assessment, whichsurely is not to repeat an assessment of conceptswhich can be carried out more effectively in otherways. Furthermore, the evidence suggests that itsinclusion denies some pupils the opportunity todemonstrate their procedural understanding.

A move towards subjective judgement based onkey indicators, supported by other criteria, would,we believe, not only make assessment easier tomanage in the classroom but also give a better pic-ture of pupil ability. The most recent documentson assessment from the examining boards seem torecognise and support this view. Most encourag-ingly of all, perhaps, is the extent to which thelanguage of investigations and its attendant caselaw is converging. This, we must hope, signals aconvergence of philosophy.

References

Dearing, R. (1993). The National Curriculum and ItsAssessment: An Interim Report. York, London,NCC/SEAC.

Department of Education and Science and Welsh Office(1988a). National Curriculum Task Group on Assessmentand Testing: A Report. London, HMSO.

Department of Education and Science and Welsh Office(1988b). Science for Ages 5 to /6. London, HMSO.

Department of Education and Science and Welsh Office(1991). Science in the Ivational Curriculum. London,HMSO.

3


Gipps, C.V. (1992). National Curriculum assessment: Aresearch agenda. British Education Research Journal,18(3) 277-86.

National Curriculum Council (1988). ConsultationReport: Science, December. York, NCC.

National Curriculum Council (1991). ConsultationReport: Science. September. York, NCC,

NEA (1993). Science Framework Guidance on theAssessment of Scl GCSE. Northern Examinationsand Assessment Board.

SEAC (1992). School Assessment Folder Part Two:Science. National Pilot. London, SEAC.

SEAC (1993). School Assessment Folder Key Stage 3:Assessing Scl . London, SEAC.

4 11-I 0 0

I CHAPTER 10

Postscript

This book has explored the role of investigationsin the science curriculum in the light of currentresearch. Ou central theme has been that investi-gations are not just another tool in the teachers'repertoire for teaching the substantive concepts ofscience, but a tool with a very specific purposethat of teaching for the understanding. applicationand synthesis of concepts of evidence which lie atthe heart of procedural understanding. We haveput forward ways of teaching these concepts ofevidence and of incorporating investigations intothe curriculum. We have suggested ways ofmaking investigations more meaningful to pupilsby using the notion of an 'audience' for the result-ing evidence.

The current science curriculum focuses primar-ily on substantive conceptual knowledge. We havesuggested that a more balanced view of sciencewould give greater attention to procedural under-standing. The research we have presented disputesthe assumption that children gain proceduralunderstanding simply by doing practical work inthe course of traditional science teaching. Thiscommon view was epitomised at a recent confer-ence when an apparently accepted opinion wasstated as: 'we know that practical work is import-ant but because everyone can do it. there's noneed to worry about it'.

Research findings suggest that there is indeed'something to worry about'. While it is encourag-ing that children can do practical investigations inthe sense that they can design and collect data, it isa matter of concern that they do not get past this

point and that they do not value the significanceof their own evidence. Pupils particularly lack theimportant parts of procedural understandingwhich we have defined as those concepts of evi-dence associated with the validation of evidence. Ifpupils have not been taught these ideas, then it ishardly surprising that they do not understandthem. There has also been, and still is, a tendencyto regard procedural understanding as somethingwhich once acquired is only useful in as much as itis a nieans of teaching cdricepts. We have arguedthat procedural understanding is something that isworth teaching in its own right.

We have demonstrated how the content of pro-cedural understanding can be broken down intoconcepts of evidence which serve two purposes.First, it enables the teacher to examine children'sperformance in investigations to see which con-cepts children do and do not understand. Second,it means that investigations can be designed tofocus on particular aspects of investigations andto teach particular concepts of eidence. The issueof the assessment of investigations is not withoutits problems, particularly in the UK NationalCurriculum. If we accept the view that the assess-ment of the synthesis of concepts of evidence is thegoal, then assessment should ultimately rely on theinformed judgement of the professional teacher.

So far, our arguments have stemmed from theresearch we have presented. Here we shall put for-ward a more personal view.

If we consider the aims of educating pupils in sci-ence, then we start from the point of view of equality

41 3 4


of opportunity which we take as fundamental. Aneducation system without a belief that all pupils areequally worth educating and that they deserve thesame chance would be a moral failure. This does notmean that we can guarantee, or even wish to strivefor, equality of output. That is up to each individualpupil. But they should all have the chance. In termsof science education, we are not talking about pro-ducing more top scientists but working towards amore scientifically literate society.

The phrase 'the public understanding of science'is much in vogue at present. It points to the fact thatthere is a growing awareness that there is somethingwrong with that understanding that it is in need ofimprovement. Industry is particularly concernedthat a science-trained workforce needs more 'trans-ferable skills' at its disposal. Both these issues canbe, and we would argue should be, addressed at theschool level.

How then do pupils regard science? To many, itis a subject that is hard, irrelevant and boring.Added to this is the fact that science has a negativeimage in society generally, which is passed downfrom one generation to the next. We shall considereach of these claims in turn before we considerhow these problems can be addressed.

Science is hard?

Assessment of Performance Unit (APU) data fromthe 1980s showed that something like 20 per centof the nation's 16-year-olds could make sense ofthe concepts of science. In some areas of physicsand chemistry, the proportion is nearer 2 per cent.It seems as if science is indeed 'too hard' for themajority. But this is well known and equally welldocumented:

School science has, both cwertly and covertly.become more pure. conceptually demanding andcomplex, and less concerned with the everydayreality and experience of our youngsters, theirparents and their employers: it has, in so manyways. become a complex symbolic system access-ible to the few.

(ASE, 1979)

A more personal view, and one which expressesthe sheer frustration felt by many of our best sci-ence teachers, is typified by this quote from ateacher in one of our research schools:

I want people to be interested in science. None ofmy friends ever were. I think if it had been left toschool science alone, I would definitely have fol-lowed the Arts. But my father was an engineerand he was fascinated by science and this influ-enced me.

By the 6th form, if you thought about the sci-ence it held you back - if you learned theformulae and followed the instruction you werewell away if you could stand the boredom.Science had become dead, and dead hard.

[At university], my friends did English and artand French and history: they were always beingencouraged to think for themselves, analyse butgive their own opinion and reasons, using the evi-dence they had. Would tutorials in science openup this discussion? No! - more formulae andinstructions. All in all, the message was be ableto stick the numbers in the right places in the for-mulae, but never mind what it means.

When I started teaching I decided I wouldn'tteach like that, but would try to make thingsclearer to everybody. But in the end I felt I'dconned them into opting [for physics] because.with all the formulae and instructions, there washardly time (or anyone with the staying power), orteaching suggestions in text books to try to keepit interesting. I felt like force of [the teacher's]personality alone drove th.m to success at '0'level, in the midst of incomprehension.

Science is irrelevant?

The criticism that school science is irrelevant is notnew either:

Although there are some impressive exceptiens.too much time spent learning science by too manypupils consists of the accumulation of facts andprinciples which have little perceived, or indeedactual, relevance to their daily lives as youngpeople or adults.

(DES, 1985)

nd

POSTSCRIPT

Unfortunately, there is a feeling in society andamong some scientists that school science is schoolscience and no matter whether it appears to beboring or irrelevant, you cannot fiddle about withit. This is the 'science is hard, and if they can't getinto it, well so be it' approach. Whether this is theview of those who have been through sciencetraining, or the view of 'laymen' who are over-impressed by the mysticism of pure science, isarguable. The political rhetoric in some quartersseems particularly impressed with these argu-ments: it has the feel of 'back to basics' about it:

The thrust in science towards practical. investiga-tive and experimental work has left less timeavailable for its teaching. and has underminedfurther the acquisition of knowledge.

(Centre for Policy Studies, 1988)

This extraordinary statement beggars belief. Whatwe can only assume, to be charitable, is that theauthors have confused discovery learning withpractical science. If they really believe that scienceis not about 'practical, investigative and experi-mental work', then the future for science andengineering is indeed bleak.

Science is boring?

Ideas that are difficult, set in contexts perceived asirrelevant, result in failure. Failure often leads tolack of motivation, boredom and classroom dis-ruption. Many students leaving school science.never to return, will say that it was a set of dis-jointed and incoherent facts that made increasinglyless sense. Foulds et at (1992). in discussing thiskind of fragmented teaching, write: '...this situ-ation is likely to leave many pupils with the samesterile view of science as that which pertainedduring the 1960's and which resulted in few stu-dents opting for further studies in science'.

This may all sound like an attempt to placeblame on the teachers. Not so. They can onlyteach what the curriculum asks indeed, in theUK now instructs them to teach. They willnearly all say that the curriculum in secondary

J

141

schools is too full. There is no time to exploreideas and the consequences of science in theirpupils' everyday lives. And so science stops withthe school bell.

A bad image?

Young people see science as destructive. By dam-aging the environment, contributing to wars andexperimenting on animals, science has lost itsappeal and the popular portrayal of scientistsshows them as dangerous or insane.

(Purchoii. 1991)

Science has undoubtedly caused some of theproblems in today's world but it has also foundthe cures to many, although the latter is notalways recognised.

Another contributory factor to this negativeimage is that the 'objective' nature of science canbe taken to imply that scientists are uncaring and


unemotional people They are seen as somehowremote from society. The more caring image ofbiology drives pupils towards the biological sci-ences or into the arts where that humanisinginfluence is more apparent.

Newton and Newton (1992) asked young chil-dren to draw a picture of a scientist and found ason p. 141, that many young children drew a malescientist, with a balding head, wearing a whitecoat and working with test tubes in a laboratory.This stereotype of a somewhat remote figure setapart from society may be acquired by as early assix years of age.

An alternative science?

To those 'in the know', science is a fascinating wayof looking at and understanding the world. Manyyounger pupils, particularly in primary and earlysecondary, share this view. What can we do inschools to enable children to retain this interest?

If we look at the content of the science curricu-lum, then we can see a number of core ideas or keysubstantive concepts which are essential for scien-tists, engineers and technologists and indeed forthe scientifically literate person. We cannot throwthese key concepts away, even if they are hard.The best we can do is to keep them to a minimumand present them in a way which makes themaccessible to the majority of students and in a waywhich highlights their relevance. But there remainsa lot of clutter around these big ideas. The resultof the clutter is that many graduates come toteacher training knowing a lot and understandinglittle. They are full of very erudite soundingknowledge. Ask them to explain some everydayphenomenon and the k .y gaps in those big ideasshow through all too clearly.

When we look at procedural understanding, itis here that we find the core of 'transferable skills'that industry asks for. These transferable skills arethose which make up the general scientificapproach to issues which can be applied acrossmany situations. One of the most important ofthese skills is the ability to evaluate evidence:

To decide between the competing claims of vocalinterest groups concerned about controversialissues such as 'acid rain', nuclear power, in vitrofertilisation or animal experimentation, the indi-vidual needs to know some of the factualbackground and to be able to assess the quality ofthe evidence being presented.

An uninformed public is very vulnerable to mis-leading ideas on, for example, diet or alternativemedicine. An enhanced ability to sift the plausiblefrom the implausible should be one of the benefitsfrom better public understanding of science.

(The Royal Society, 1985)

Evidence is, science must be central. It is thekey that differentiates science from maths, forinstance. In science, as in mathematics, any theoryhas to be internally consistent but in science it alsohas to be tested against reality. If it fails the test, itis, or should be, discarded. That is a pure scien-tist's view of evidence and one that is difficult tofit neatly into science for younger pupils.

But applied scientists and engineers rely on evi-dence to a far greater extent. When an engineerdesigns a bridge, the theory may help in under-standing the problem, but it would be a foolishengineer indeed who designed a bridge in theabsence of empirical evidence about real bridges.Evidence here is the rock upon which any engi-neering project is based and the quality of thatevidence and its believability must be central. It isalso more accessible to pupils in schools than the'purer' use of evidmce to test theory. Scientists inindustry rely on the ability to generate, validateand interpret evidence. But most importantly, it isthis evidence which will form the basis for the:

Humanising [of] the science curriculum by devel-oping an understanding of human nature inrelation to the natural environment to enable citi-zens to deal with problems that have ethical, valueand moral components.

(Hurd, 1993)

It is the ability to evaluate evidence and to beginto appreciate such components, that we are arga-ing for in this book and which we believe providesthe core of the transferable skills that we seek.

137

POSTSCRIPT

Turning to the issue of motivation which mustbe central to any development in science educa-tion, there is general agreement that investigativework engages pupils' interest. In particular, thereis some anecdotal evidence that girls, the biggestsource ol untapped potential in science, find theopen approach more to their liking. Why mightthat be? One suggestion that has been floatedconcerns the transfer of control, and ownership,to the pupil. As we saw in the quote above fromthe (female) teacher, she saw her fellow studentsat school in art subjects challenged to think,while she plugged numbers into formulae.Investigations require that sort of challenge tothink. What we hope we have shown in this bookis that there is a 'content', primarily of conceptsof evidence, to investigative work which under-pins and provides the rationale for their inclusionin any balanced curriculum.

143

References

Association for Science Education (1979). Alternativesfor Science Education. Hatfield, ASE.

Centre for Policy Studies (1988). Simple Curricula forEnglish, Maths and Science. Policy Study No. 93.London, Centre for Policy Studies.

Department of Education and Science ( 985). Science5-16: A Statement of Policy. London, HMSO.

Foulds, K., Feasey, R. and Gott, R. (1992). InvestigativeWork in Science. Durham, University of Durham.

Hurd, P.D. (1993). Comment on Science educationresearch: A crisis of confidence. Journal of Researchin Science Teaching, 30(8): 1009-11.

Newton, D.P. and Newton, L.D. (1992). Young children'sperceptions of science and the scientist. InternationalJournal of Science Education, 14(3): 331-48.

Purchon, V. (1991). A prayer for the new priest-craft.Times Educational Supplement, 8 November.

Royal Society (1985). The Public Understanding ofScience. London, The Royal Society.

.1 .)

Index

Adey, P., 35-7, 93age

of pupils in NCC sample, 54effect on performance of

investigations, 58-9, 64APU (Assessment of Performance

Unit)assessment framework, 41-3, 109,

128

concepts of evidence, 69, 73-9,83-4

factors affecting performance,51, 58

planning of investigations, 119problem-solving model, 44pupil motivation, 61skills, 67-8, 112

APWIS (Assessment of PracticalWork in Science), 47, 48

Armstrong. H. E., 17, 18, 23, 27assessment

choosing an investigation for.113-15

evidence for, 115-21self-assessment, 121

Bloom's taxonomy, 28-9, 110-11

CASE (Cognitive Acceleration inScience Education), 25, 34-9,93, 103

cognitive processes, 19, 20, 25, 27concepts, 53, 142

definition, 26

effect on performance ininvestigations, 56-7, 64, 72-3,113

concepts of evidence, 29-33, 63-4,69-85

assessment of, 113, 123-5, 136-7and CASE, 38-9, 43in National Curriculum, 128-31,

133

teaching, 88conceptual difficulty, 51, 53, 97-8conceptual taxonomy, 29conceptual understanding

assessment of 136definition. 26in science, 25-7in Scl in the National

Curriculum, 128, 131-2teaching for, within

investigations, 87-8, 95-6, 101,110

context, 42, 51, 54, 113effect on performance of

investigations, 59, 64

data handlingassessing, 122-3teaching, 91-3, 95understar.ding of in

investigations, 30-2, 7584design

assessing, 122-3teaching, 89- 90, 93

4 33

understanding of ininvestigations, 30-2, 69-73, 84

differentiation, 96-7, 113-14

enquiry, 21, 28evaluation

in National Curriculum, 129-30understanding of, 30-1, 33, 43, 83-4

evidence, 33-4, 133, 135, 142

GASP (Graded Assesssment inScience Project), 20

heurism, see ArmstrongHodson, D., 23-4

illustration, 21, 28, 62investigative work

definition, 20, 22, 48examples of, 37, 42-3, 53-4purpose of, 22types of, 49

Jones, A. T., 43-5, 61

language of investigations, 103, 114

measurementassessing, 122-3teaching, 90-1, 93-4understanding of, 30-2, 73-.5, 84

motivation, 60-1, 64, 143

National Curriculum, 20, 32, 48, 63,127-38

146

NCC (National Curriculum Council)project

aims, 47analysis, 55-6design, 53-5findings, 56-65,69-84methodology, 55sample, 52types of investigations, 47-8,

Nuffield Schemes, 18-19,20,21,23and model of science, 27-8

observationof investigations, 115-17of practical work, 22,28,62

openness, 54-5,114effect on performance, 59-60

OPENS (Open-Ended Work inScience Project), 43-5

organisation of investigations, 104,106

Piaget J., 35-6,49planning, 119

INVESTIGATIVE WORK IN :THE SCIENCE CURRICULUM

practical workdevelopment of, 17-21types of, 21-4

problem-solving, 26,41,43,44procedural complexity, 51,53,96-7,

113-14effect on performance of

investigations, 57-8procedural understanding, 87-8

assessment of, 110,111-25,131-2,142

. definition, 25-7,29-34,teaching for, 87-95,101

processes, 19-20processes and skills movement, 19-20progression, 96-7,102,132

questioning, 119

SAPA (Science A ProcessApproach), 19

schemes of work, 99-102,107Screen P. A., 19-20

Simon, S. A., 43-5,61skills, 20,26-7,32

assessment of, 112performance of, 67-8skills practicals, 21,28,62

substantive concepts, see concepts

TAPS (Techniques for theAssessment of Practical Skills), 20

teachersperceptions of investigations,

-62-3,65perceptions of practical work, 62,

135

role in investigations, 105thinking skills, 34

variablestypes of, 32,47variable structure, 47-9,53-4,

57-8,114

Warwick Process Science, 20

PRACTICAL SCIENCETHE ROLE AND REALITY OF PRACTICAL WORK IN SCHOOL SCIENCE

Brian Woolnough (ed.)

Science teaching is essentially a practical activity, with along tradition of pupil experimental work in schools.And yet, there are still large and fundamental questionsabout its most appropriate role and the reality of whatis actually achieved. What is the purpose of doing prac-tical work? to increase theoretical understanding or todevelop practical competencies? What does it mean tobe good at doing science? Do we have a valid model forgenuine scientific activity? and if so do we develop itby teaching the component skills or by giving experiencein doing whole investigations? What is the relationshipbetween theoretical understanding and practical perfor-mance'? How significant is the tacit knowledge of thestudent, and the scientist, in achieving success in tack-ling a scientific problem? How important are suchfactors as motivation and commitment? What do wemean by transferability and progression in respect topractical work? do they exist? can they be defined?How can we assess a student's practical ability in a waywhich is valid and reliable and at the same time encour-ages, rather than destroys, good scientific practice inschools? This book addresses such questions.

By bringing together the latest insights and research find-ings from many of the world's leading science educators,new perspectives and guidelines are developed. It pro-vides a re-affirmation of the vital importance of' practicalactivity in science, centred on problem-solving investiga-tions. It advocates the need for students to engage inwhole practical tasks, in which all aspects of knowledge(tacit as well as explicit), of practical ability, and of per-sonal attributes of commitment and creativity, areinteracting. While considering the particularly pertinent

issues arising from the National Curriculum for Sciencein England, its discussion is equally germane to all con-cerned with developing good practical work in schools .

ContentsSetting the scene Practical work in school science: ananalysis qf current practice The centrality of practicalwork in the SciencelTechnologylSociety movement-Practical science in low-income countries a means to anend: the role of processes in science education Practicalwork in science: a task-based approach? - Reconstructingtheory from practical experience Episodes, and the pur-pose and conduct of practical work Factors affectingsuccess in science investigations School laboratory life

Gender differences in pupils' reactions to practicalwork Simulation and laboratory practical activityTackling technological tasks Principles of practicalassessment Assessment and evaluation in the sciencelaboratory Practical science as a holistic activityRekreiwes Index.

ContributorsTerry Allsop. Bob Fairbrother, Geoffrey J. Giddings.Richard Gott, Richard F. Gunstone, Avi Hofstein,Richard Kimbell. Vincent Lunetta, Judith Mashiter,, thin Millar. Patricia Murphy, Joan Solomon. PinchasTamir, Kok-Aun Toh, Richard White, Brian E.Woolnough. Robert E. Yager.

224pp 0 335 09389 2 (Paperback) 0 335 09390 6(Hardback)

141

TEACHING IN LABORATORIES

David Bond, Jeffrey Dunn and Elizabeth Hegarty-Hazel

This is a complete guide to the design and organisationof laboratory activities and the conduct of laboratoryteaching. It is an exhaustive, up-to-date account andappraisal of current practice, with recommendations forchange supported by case studies.

Notoriously time and labour intensive, lab work isdoubly threatened when resources run scarce. It is par-ticularly dependent upon precise objectives, purposefulstructure, effective experiments, and close assessment.This book brings together and elucidates all that hasbeen and could yet be done in the sciences to maintainand improve effectiveness and keep lab work in the

I 4

curriculum. Research and regular course evaluation inthis field are shown to be essential to development.

ContentsIntroduction Aims, objectives and course planningTeaching strategies - Sequencing and organizationAssessment of students Monitoring laboratory teaching

Research on laboratory work Aide-memoire for thedevelopment of a new laboratory course ReferencesIndex.

192pp 0 335 15609 6 (Paperback)

BIOTECHNOLOGY IN SCHOOLSA HANDBOOK FOR TEACHERS

Jenny Henderson and Stephen Knutton

In recent years there has been spectacular growth inbiotechnology and in its importance for the school cur-riculum. This handbook offers teachers:

an overview of the significance and scope of biotech-nologyan introduction to the content of biotechnology andits relevance to the everyday worlda guide to how biotechnology fits into the NationalCurriculum, within and across subject disciplinesappropriate teaching strategiessuggestions for practical workcase studies and other material which can be useddirectly with sixth form studentsa glossary of termsa guide to resourcescoverage of safety issues.

This is an essential resource for practising and traineeteachers of science and technology.

ContentsWhat is biotechnology? Biotechnology and the schoolcurriculum Biotechnology and the food industryBiotechnology and medicine - Biotechnology in agricul-ture Biotechnology and the environment - Biotech-nology, fuels and chemicals Biotechnology throughproblem solving Biotechnology through discussion-based learning - Practical considerations - ResourcesGlossary Appendix References Index.

I76pp 0 335 09368 X (Paperback)0 335 09369 8 (Hardback)

1 4 3

LEARNING AND TEACHING IN SCHOOL SCIENCEPRACTICAL ALTERNATIVES

Di Bentley and Mike Watts

This book provides a series of different approaches toteaching school science. These approaches will be of usenot only to science teachers but also to teachers outsidescience and in different parts of the education system.

The book is organised as follows. The first chapterlooks at pressures for change: the authors show thatscience teachers need to adopt new and differentpproaches to teaching and learning. In particular. theauthors focus on the notion of active learning - a themethat runs through the remainder of the book. In the fol-lowing chapters, case studies are clustered around aseries of themes. The final chapter summarises theapproaches and their implications for teaching sciencefor the National Curriculum.

In general, the book is a useful, practical guide to a vari-ety of strategies and classroom activities: a collection ofexperience and ideas about different teaching methodswhich will benefit both trainee and practising teachers. Itwill appeal to those engaged in initial training and in-ser-vice work, as well as to teachers who are keen to innovate.

ContentsPreface Acknmvledgements Learning to make it yourown - Practicals and projects Talking and writing jifflearning Problem solving Encouraging autonomouslearning Games and simulations: aids to understandingscience Using role play and drama in science Mediaand resource-based learning Summary and discussionIndex.

The ContributorsBrigid Bubel, Bev A. Cussans, Margaret Davies, RodDicker, Mary Doherty, Hamish Fyfe, John Heaney,Martin Hollins, Joseph Hornsby, Andy Howlett.Pauline Hoyle, Harry Moore, Robin Moss, PhilMunson, Philip Naylor. Jon Nixon, Mick Nott, AnitaPride, Peter Richardson, Linda Scott, Brian Taylor,David Wallwork. Norma White, Steve Whitworth.

224pp 0 335 09513 5 (Paperback)0 335 09514 3 (Hardback)

.41111111/BMW!NMI11111111.. ,11111111==

DEVELOPING SCIENCE ANDTECHNOLOGY EDUCATIONINVESTIGATIVE WORK IN THESCIENCE CURRICULUM

The book considers the place of investigative work inthe science curriculum and presents the latest researchin this field. The authors consider the theoretical frame-work which underlies this kind of practical work whatare pupils actually learning and what are we trying toteach them? They argue that the thinking behind thedoing of science is at present undervalued and that thisis something that needs to be taught. They present in-novative ways of focusing teaching on particular aspectsof investigations and consider the issue of assessment.The interplay between theory, research and practice willappeal to readers who are involved in secondary scienceeducation and who are eager to know more about thepresent state of knowledge in investigative work in science.

After working as head of physics in a large comprehen-sive school in Northern England, Richard Gott becameDeputy Director of the Assessment of Performance Unitat Leeds University (1982-84). He is now Senior Lecturerin science education at Durham University where he hasestablished the Exploration of Science team. He is authorof a number of books and policy documents in scienceeducation.Sandra Duggan is an experienced researcher, currentlyresearChing investigative work in science at DurhamUniversity.

BEST COPY AVAILABLE

ISBN 0-335-19143-6

4

document resume se 056 829 author gott, …document resume ed 390 642 se 056 829 author...

Documents