Adaptive Testing For ClassificationOf Learners And Quiz Questions Using Clicker

MTP Stage-2 Report

Submitted in partial fulfillment of the requirements for the degree of

Master of Technology

Under guidance of

Prof. Deepak B. Phatak

Submitted By

Anoop Dobhal

Roll No: 123050062

aDepartment of Computer Science and Engineering

Indian Institute of Technology Bombay

Mumbai

1

Dissertation Approval

The dissertation entitled “Adaptive Testing For Classification of Learner andQuiz Question Using Clicker”, submitted by Anoop Dobhal (Roll No:123050062) is approved for the degree of Master of Technology in Com-puter Science and Engineering from Indian Institute of Technology, Bombay.

Examiner-I

Examiner-II

Supervisor

Chairperson

Date :

Place :

Declaration

I declare that this written submission represents my ideas in my own wordsand where others ideas or words have been included; I have adequately citedand referenced the original sources. I also declare that I have adhered to allprinciples of academic honesty and integrity and have not misrepresented orfabricated or falsified any idea/data/fact/source in my submission. I under-stand that any violation of the above will be cause for disciplinary action bythe Institute and can also evoke penal action from the sources which havethus not been properly cited or from whom proper permission has not beentaken when needed.

Anoop Dobhal

(123050062)

Acknowledgement

I would like to thank my guide, Prof. Deepak B. Phatak, for his consistentdirections and guidance throughout the project. His perpetual motivation,patience and excellent expertise in discussion progress of the project workhave benefited me to an extent, which is beyond commendation. Becauseof his consistent encouragement and right directions, we are able to do thisproject work efficiently and correctly. I would also like to thank Mr. NageshKarmali for his continuous support throughout project work. Special thanksto my colleagues and friends for providing me useful comments, suggestionsand continuous encouragement.

Abstract

Now a days, clickers are called with many names like CRS(Classroom Re-sponse System), ARS(Audience Response System), SRS(Student ResponseSystem) and many more. Many questions are still in the queue, which are yetto be answered, like, Do clickers improve learning, What type of questionswill enhance students participation and improve their knowledge?This report presents an overview of clickers system, methodologies acceptedfor improving active participation of students, from the start of hardwareclickers to software clickers implementation in portable devices like tablets,mobiles, PDA, or laptops. This report also presents kinds of problems andtheir connection to active learning in various modes of quizzes, questionstypes, devices and technologies used. Lastly, in this report, we present a sys-tem for improving learning in classrooms by classifying question accordingto their difficulty level and user/client according to their intelligence level.

Contents

1 Introduction 41.1 What is Clicker . . . . . . . . . . . . . . . . . . . . . . . . . . 41.2 Uniqueness of clicker . . . . . . . . . . . . . . . . . . . . . . . 51.3 Types of data collected . . . . . . . . . . . . . . . . . . . . . . 51.4 Downside of clickers . . . . . . . . . . . . . . . . . . . . . . . . 51.5 What Classroom Response System (CRS) can add to a learn-

ing environment . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2 Literature Review 72.1 Advantage of Clickers . . . . . . . . . . . . . . . . . . . . . . . 92.2 Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

3 Clicker development at Indian Institute of Technology Bom-bay (IITB) 113.1 CLICKER 2009A, Ver 1 . . . . . . . . . . . . . . . . . . . . . 113.2 Clickers in distance education for a multiple classroom envi-

ronment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133.2.1 Approach . . . . . . . . . . . . . . . . . . . . . . . . . 133.2.2 Architecture of the Distributed Student Response System 14

3.3 Web Based Clicker System for Aakash Tablet . . . . . . . . . 163.3.1 Different modules in AakashClicker . . . . . . . . . . . 163.3.2 General Constraints or Limitations in AakashClicker . 16

4 Problem Formulation 194.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194.2 Problem Statement . . . . . . . . . . . . . . . . . . . . . . . . 204.3 Scope of the Project . . . . . . . . . . . . . . . . . . . . . . . 204.4 Constraints of the Project . . . . . . . . . . . . . . . . . . . . 20

1

4.5 Assumptions and Dependencies . . . . . . . . . . . . . . . . . 214.6 Proposed System . . . . . . . . . . . . . . . . . . . . . . . . . 21

5 Contribution 225.1 Algorithm for Recommendation System . . . . . . . . . . . . . 22

5.1.1 Algorithm for modifying the Parameters . . . . . . . . 225.1.2 Scenarios for User Uk . . . . . . . . . . . . . . . . . . . 235.1.3 Scenarios for Question Qk . . . . . . . . . . . . . . . . 24

6 Software Implementation 256.1 Software Requirements . . . . . . . . . . . . . . . . . . . . . . 26

6.1.1 Server Side . . . . . . . . . . . . . . . . . . . . . . . . 266.1.2 Client Side . . . . . . . . . . . . . . . . . . . . . . . . . 26

6.2 Server Module . . . . . . . . . . . . . . . . . . . . . . . . . . . 266.2.1 Login Module . . . . . . . . . . . . . . . . . . . . . . . 266.2.2 Create Quiz Module . . . . . . . . . . . . . . . . . . . 276.2.3 Add Question Module . . . . . . . . . . . . . . . . . . 276.2.4 Database . . . . . . . . . . . . . . . . . . . . . . . . . 276.2.5 Web Interface . . . . . . . . . . . . . . . . . . . . . . . 29

6.3 Client Module . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

7 Test Results 347.1 Test Results When One User Attempt The Quiz . . . . . . . . 367.2 Test Results When Five Users Attempt the Quiz . . . . . . . . 407.3 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

8 Conclusion 43

9 Future Work 44

2

List of Figures

3.1 Hardware Clicker developed at IITB . . . . . . . . . . . . . . 123.2 Distributed System Setup . . . . . . . . . . . . . . . . . . . . 133.3 Flow Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . 143.4 Protocol Diagram . . . . . . . . . . . . . . . . . . . . . . . . . 153.5 Software Clicker developed at IITB . . . . . . . . . . . . . . . 18

6.1 Software Architecture Diagram . . . . . . . . . . . . . . . . . 256.2 Server Interaction Diagram . . . . . . . . . . . . . . . . . . . . 276.3 Database Relation Diagram . . . . . . . . . . . . . . . . . . . 286.4 login module . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306.5 add question module . . . . . . . . . . . . . . . . . . . . . . . 316.6 create quiz module . . . . . . . . . . . . . . . . . . . . . . . . 326.7 flow diagram of quiz application . . . . . . . . . . . . . . . . . 33

3

Chapter 1

Introduction

Clickers, known with many names like ARS (Audience Response System),CRS (Classroom Response System), SRS (Student Response System), PRS(Personal Response System), etc. which act as learning aid for better under-standing of course material and individual growth.There are a lot of educational softwares in market for e-learning environ-ment. Clicker system is one of the e-learning software, which act as readingand writing tool that help students to achieve growth and knowledge. In earlyyears, before invention of smart phones, clickers are just act as an hardwaretool based on Infrared communication. The cost and other installation over-heads led to next generation of clickers. Now a days smart phones, tablets,or laptops with internet access are used as clickers. Generally student do notask their doubts in class due to fear or a class is so big that not all get chanceto communicate with teacher, because of this major concern, clickers cameinto an existence.As we gaze out at the sea of slouching bodies and expressionlessfaces, it is hard to resist wondering if students want less educationand more entertainment. Rand W. Guthrie and Anna Carlin

1.1 What is Clicker

Clickers are act as medium of communication between teacher and students.It uses infrared or radio frequency technology for transmitting and receivingdata. Clickers as previously explained, can be an any portable handhelddevice like smart phones, tablet, laptops or any hardware are given to all

4

students in the class. The teacher can ask question or post quizzes duringclass, student then give response through their device. The overall feedbackis tabulated and displayed on the projector.

1.2 Uniqueness of clicker

The main theme which makes clicker different from other teaching pedagogyis that teacher always has a right track of lecture going on. As studentresponses are collected in teacher computer, so teacher can correspond toresponses according to priority or usefulness of responses during class.

1.3 Types of data collected

In order to better understand methodologies in use, the data collected isrepresented as interview, survey, questionnaire, quizzes, test results, obser-vation, field notes, etc..The most frequently data collection form is surveying and quizzes speciallymultiple choice quizzes. The response to this collected data can be individualor group. Sometime individual response is followed by group discussion.

1.4 Downside of clickers

• If iclickers are used in a class, then cost factor comes between learning.Which includes installation cost, maintenance cost, etc..

• As clickers generation is in rapid growth, now portable devices likeSmartphones, tablet, laptops are replacing additional cost factor andinstallation cost, as mostly students carry mobile portable device. Butstill some student is unable to arrange these devices. These portabledevices have some ill-effects on students. Student can reflect to othersides of portables device use if wifi connectivity is available all the time.

• Now a days active learning participation is implemented by using SMSservices also. Student sends their responses through SMS to teacherdevice with the help of sms-gateway. This approach has cost factor ofsms-pack. Sending sms are not free, so sometime students avoid givingtheir responses.

5

1.5 What Classroom Response System (CRS)

can add to a learning environment

There is lot of pedagogical and implementation methodologies are proposed.Many of them change the form of teaching. Studies have a proof of increasedstudent participation and active involvement. But the question in time isthat, what CRS can add to a learning environment?

• On what basis of comparison it should be rectify, either by differencein the use, or lack of use, of a CRS?

• What type of pedagogical approaches required to provide greater un-derstanding, group-based, individual, individual response followed bygroup-discussion.

• What types of question formats will help student in greater extent?

• How to adjust the pace of presentation and explanation strategies ac-cording to student feedbacks?

6

Chapter 2

Literature Review

There are a lot of changes happened in pedagogical construct, such as thetiming of feedback, affordances and limitations of traditional CRS, movingfrom current generation to next generation of CRS. In this chapter we givea brief survey or research related to clickers. Most of the literature surveylights on Pedagogical Theory and Implementation.Roy P. Pargas (2006) [14] describes a Web based tool: Message Grid, whichis a two-dimensional grid like structure for posting queries. Anyone can postor reply to queries. Later they added feature of clicker for conducting quizzesand questionnaire. It also supports ink-based animated software for writinghand written response or queries. Roy P. Pargas et al. (2006) also presentedan approach to teach algorithm and data structure course by using a web-based tool called by Message Grid.Scott Teel et al.(2012) [17] also presented a web-based student response sys-tem using existing user portable devices, a wireless network and a cloud-basedbackend server and a database. So eliminating additional cost of hardwareand installation overhead. It also supports text and image for questions. In-structor use Google doc for adding questions and later invoking web-basedphp script for adding question from Google doc to MySql database.Abderrahmane Lakas et al. (2006) [9] implemented a tool, ACP an interactiveclassroom response system for active or cooperative learning environment. Itsupports some protocols which allows the students to discover the currentserver IP address and allows the server to push the assessment results.Eusebio Scornavacca et al. (2007) describe a TXT-2-LRN, sms-based class-room interaction between students and instructor. Instructor mobile is con-nected to laptop to receive sms responses through SMS gateway, or SMS

7

studio (SMS management system).David Lindquist et al. (2007) [10], Wolfgang Hrst et al. (2007) [6], CraigPrince (2007) [15] describe the use of mobile devices in active participationduring classes. It reduces the cost and installation overhead for setting CRS.Similarly, L. Jackowska-Strumillo et al. (2013) [7] describe a concept of us-ing clickers software in portable devices like tablet, laptops, PDA or mobilephones.Joana Cruz e Costa et al. (2008) describe a browser based CRS, where stu-dent application was implemented using J2ME and instructor applicationimplemented as dynamic web page.Monika Andergassen et al. (2013) [1] presented similar approach, a browser-based mobile clicker, which is a platform independent tools operated via anyinternet-enabled device. It also support spatially distributed communication.The system was developed over package called XoWiki content flow availablein XoWiki framework. The main key of this approach is that, at differentstate we can present different content to different kinds of users.Matthias Hauswirth et al.(2009) [4] implemented software clickers in java,that support richer problem types other then only multiple choice. Theirsystem was a Java based Classroom Response System for Teaching Java, inwhich problem is defined as Syntax, Types, Coding and Control Flow. Theyrepresent the Solve and Evaluate Approach in which students who solvedtheir problem can evaluate problems of other students.Beth Simon et al. (2010) [16] also implemented java teaching software inactive classroom.Similar work is done by Daniel Zingaro et al. (2013) [19] they implementedPython Classroom Response System (PCRS), that can be used by instruc-tor in classroom to run both PI (multiple choice) and code-writing questions.The PCRS is entirely web-based tool so it can used using laptops and tablets.The code submitted by students is executed on a server which returns feed-back on student machine.T. Mantoro et al. (2010) [12] describe Survnvote a web-based tool for audi-ence response system. Another web-based tool was presented by M. Jagaret al. (2012) [8] named as Auress, the audience response system. Similarlylater, M. Llamas-Nistal et al. (2012) [11] present web-based Audience Re-sponse System using the educational platform called BeA, a web form useby students as a device, which can be operated by any portable device liketablet, mobile and laptops.Brian T. Davis (2010) [3] describes a method used in CRS for numerically

8

intensive courses. This method involve decomposition of numerical probleminto steps, which can utilize CRS. Then use of CRS in identifying the ap-propriate equation for a problem. Finally generation of numerical problemresponses.M. Miura et al.(2012) [13] describe a Device-Free Personal Response Systembased on Fiducial Markres. It involves collecting student responses by meansof printed fiducial markers. Tracking technology based on computer vision,students hold their marker sheet and camera catch their responses. The in-teraction techniques: Response for Multiple Choice Questions and Responseby Pointing (using elevation and heading angle) are the key-point in theirresearch. This system have some major drawbacks occlusions and insufficientlightning. High camera resolution which is needed to increase the overheadon instructor system and also decrease the image tracking refresh rate.A similar approach was presented by Andrew Cross et al. (2012) [2] usinga laptop and an off-the-shelf webcam, their software automatically recognizestudents responses.Reinicke B et al.(2012) [5] implemented an CRS application called uRespondon ipad. The purpose of their research was to develop a free form studentinput such as graphics, calculation, drawing, or structure creation and ma-nipulation.Palmo Thinley (2013) [18] discuss about the use of tablet in teaching andlearning environment.Robert Law (2013) presents an ongoing project based on quick response (QR)codes and Google forms for generating rapid response polls and quizzes. Theprocess followed by installing Google app on user mobile phone. The key com-ponents of this system are QR codes, Google spreadsheet, bar code reader,Google form and Google Scripts. This system has no cost overhead.

2.1 Advantage of Clickers

• Clickers technology and advancement Provide instructors, a new wayof teaching and improving methodologies of learning.

• Students participation increased, as students start involving in class bygiving feedbacks, asking questions, giving responses in form of quizzes,exams.

• As students response collected and tabulated during the class itself, so

9

instructor get a good hand on students understanding about a subject.

2.2 Observations

The following observations made through detail literature survey of researchrelated to clickers, ARS (Audience Response System), CRS (Classroom Re-sponse System), SRS (Student Response System), or PRS (Personal Re-sponse System).

• Type of response, either group or individual, or both affects activelearning, depends upon the understanding of course material.

• In spite of a list of advancement raised in clickers, results are still nothighly adequate.

• Pedagogical Theory (affordances and limitations in traditional CRS)or Implementation Studies (Where and how data are collected, typesof data collected in order to better understand methodologies in use)affects the type of learning.

• Various questionnaire submitted by students shows their interest to-wards CRS, but their responses are still not very promising.

• Selection of question type play an important role in clicker system.

10

Chapter 3

Clicker development at IndianInstitute of TechnologyBombay (IITB)

Clicker development at Indian Institute of Technology Bombay (IITB) isa Ministry of Human Resource Development(MHRD) project aimed to im-prove active learning by motivating students active participation in or outsideclass. Generally, it is used for collecting attendance, creating and conductingquizzes, generating reports, polling and feedbacks in real time.

3.1 CLICKER 2009A, Ver 1

In 2009 the first clicker system which was developed at IITB based on hand-held device. It was combination of two devices, one at student side forsending response and other one at instructor side for collecting students re-sponse. The receiver and clicker both are designed on a Texas Instruments,CC2510-F32 Low-Power SoC (System- on-Chip). The receiver has the samedesign as a clicker with some additional circuitry of level converter MAX-232. The clicker is normally in sleep mode. When you have to give response,you press certain key ST (start) only then the clicker wakes up and comes into operational mode. The response keyed in is stored in the clickers memoryand the receiver collects the responses of all the clicker by polling them ina sequence. A computer connected to the receiver collects responses usinga small program and sends the results to the IIT-B central server. The ad-

11

vantage of clicker and receiver based on the same design is that they areinterchangeable, so if a receiver fails to work, any clicker can be converted into receiver by connecting to an additional small circuitry.Figure 3.1 shows the model of hardware clicker used in IITB. The receiver at

Figure 3.1: Hardware Clicker developed at IITB

instructor side is powered through any computers USB port. The receiver,through easy-to-use software, logs and stores the data of each individualstudent. The instructor can then display voting results in a graph, to thestudents. The results are also available for later analysis, grading, and ex-porting to any grade book software or course management system.The main drawback of this system is cost overhead around 600 for first ver-sion and around 1100 for next version. To overcome this, android basedclicker system developed at IITB.

12

3.2 Clickers in distance education for a mul-

tiple classroom environment

To provide use of clicker in distance education, Divya Tiwari el al. (2010),IITB, describe a synchronous, distributed approach for use of student re-sponse system.

Figure 3.2: Distributed System Setup

3.2.1 Approach

The main idea of clicker system use in distance learning lies in the use of xmlfile, which contain response of students from different participating remotecenters. The central servers initiate the response collection at all participat-ing remote centers. The data collected is then transferred back to centralserver using FTP.

13

3.2.2 Architecture of the Distributed Student ResponseSystem

The key feature of Distributed Student Response System is that, it is devel-oped using open source software and a handheld device. The hardware partis consisting of multiple radio frequency based clickers. The lectures werebroadcast through EDUSAT, a satellite dedicated to education sector by theIndian Space Research Organization.Figure 3.2 shows the architecture of different remote server connected to cen-tral server. Figure 3.3 shows the flow diagram of how communication carriedout between central and remote server.

Figure 3.3: Flow Diagram

The software side of this system is divided into two parts:

• For displaying question to be asked during active lecture. This part isinitiated by the central server. The software at each remote centre ismanaged by respective coordinator at each centre.

• Read, store and interpret the data collected from different participatingremote centers. The responses, which are written in xml formats, are

14

Figure 3.4: Protocol Diagram

transferred to main server using FTP. This part is runs at each of theremote centre.

Figure 3.4 shows the protocol of collecting responses from different remoteserver.

15

3.3 Web Based Clicker System for Aakash

Tablet

The main purpose of web based clicker system is to provide a suitable andeasy communication between student and instructor. It can be accessed fromthe main centre or from any remote centre with ease. Earlier, desktop basedversion of clicker was dependent on specific hardware and software to run onsome specific system. But, AakashClicker provide a web based interface foractive learning regardless of type operating system.

3.3.1 Different modules in AakashClicker

• Admin: It provides facilities to admin for adding, editing, fetching anddeleting the information. Also control access permissions like studentlogin and instructor login.

• Raise Hand: This is new facility added in AakashClicker, which providethe facility to ask questions or doubt regarding course. The instructorcan reply to doubt by checking the list of student, who pressed raisedhand button.

• Polling: Used to conduct questionnaire.

• Question Bank: Provide the facility of adding and searching questionsfor conducting exams or quizzes.

• Create Quiz: This module provides the facility of creating quiz byfetching questions from Question Bank.

3.3.2 General Constraints or Limitations in AakashClicker

• Works best in latest version of Mozilla Firefox and Google Chrome.

• Lack in vocabulary of scientific symbols.

• Used for real time response.

• Limited to HTTP/HTTPS.

16

• It requires an Internet browser to fetch Clicker Web Application throughurl on AakashClicker and at the remote centers.

1

1Figure 3.2, 3.3, 3.4 are cited from Divya Tiwari el al.(2010), IITB, Figue 3.1 is citedfrom user manual of CLICKER 2009A, Ver 1, IITB, Figue 3.5 is cited from SummerInternship Project 2012 at IITB

17

Figure 3.5: Software Clicker developed at IITB

18

Chapter 4

Problem Formulation

4.1 Motivation

In spite of a lot of applications developed over the years for increasing andimproving active participation of students. Clickers with many names likeSRS, ARS, PRS, CRS or voting system are in the race of improving teachingmethodology. From start of true/false or multiple choice questions responsesto graphics, calculation, drawing, or hand-written questions responses. Toreduce the installation cost, overhead and device cost, a lot of changes cir-cumnavigated. Reinicke B et al.(2012) [5], T. Mantoro et al. (2010) [12],David Lindquist et al. (2007) [10], Wolfgang Hrst et al. (2007) [6], CraigPrince (2007) [15], and others give the idea of implementing mobile, tabletor laptop based clickers, thus reducing the cost factors.M. Miura et al.(2012) [13], Andrew Cross et al.(2012) [2] presents the use offiducial markers using the technology based on computer vision. The TXT-2-LRN approach presented by Eusebio Scornavacca et al. (2007) was able toreduce unwanted overhead greatly.Being of all these innovations and theories, still there is something missing.Different results shown by different researchers has not given rise to full par-ticipation by students. It is assumed that unlike hardware clickers, softwareclickers, implemented in mobile, tablet or other portable devices give someform of distraction to students. Due to this distraction results are not up tothe mark of pure participation like in case of SMS-based learning studentscan easily distract from learning mind. Also, it is assumed that if same

19

questions are posted to all students, then there is a chance that studentswill cheat in order to show excellence. These small perceptions and generalassumptions are the prime source of motivation in this study.

4.2 Problem Statement

In a large or even small classroom, there is different kind of students. Theyhave different scale of learning depend upon their nature and past experi-ence. So dividing students into category of learning power is difficult taskfor a instructor. There were lots of up and downs were implemented andoptimized for improving learning in effective manner but still best responseyet not acquired. Still students are not able to understand the problems.One solution to the problem is to use Recommendation System, as recom-mendation system is implemented in varieties of social world sites, there isconvergent focus of applying recommendation system in learning through useof clickers in classrooms.

4.3 Scope of the Project

• To provide better learning environment in classroom.

• To provide means to change current clicker system learning environ-ment beyond only tabulating students responses and displaying results.

• To provide users to give Reponses according to their intelligence level.

• To provide system to change questions difficulty level according to usersresponse.

4.4 Constraints of the Project

• To adjust the pace of presentation and explanation strategies accordingto student feedbacks.

• Bandwidth management problem as numbers of user increases.

• Maintaining each student data can create extra load on server.

20

• What types of question formats will help student in greater extent?

4.5 Assumptions and Dependencies

• The Smartphone or tablet on which the clicker application runs, shouldsupport WiFi.

• The user should have a stable internet connection with a minimumbandwidth of 128kbps to view and give responses.

4.6 Proposed System

How cans recommendation system can be applied:

• First same kind of question will be given to all students.

• Then students responses and some kind of feedback is collected at in-structor side.

• Based on data collected at instructor side, our recommendation systemwill give easy or tough question.

• Those students, who solved the early problem, we will recommendtough question to them and those students, who were not able to solvethe early question will by recommended by easy question.

• Our system is based on Rasch model (is a psychometric model foranalyzing categorical data, such as answers to questions on a readingassessment or questionnaire responses, as a function of the trade-offbetween:-

– The respondent’s abilities, attitudes or personality traits

– The item difficulty. For example, they may be used to estimate astudent’s reading ability

• maintaining two variables Qj and Ui, Qj is question difficulty valueand Ui is user intelligence value. The values of these variables will bechange according to user responses.

21

Chapter 5

Contribution

5.1 Algorithm for Recommendation System

Qj, Qj is difficulty level of the questionwhere

1 ≤ k ≤ n (5.1)

n is total number of questions

Ui, Every user in the system assigned with a number Ui, this parameter isuser intelligence, more the number, he/she is more capable of solving prob-lems. The Value of Ui and Qj parameters varies between 0 and 1.

Initial ValuesA new user or new question is assigned with unbiased x values. Based on thequestions taken the parameters are expected to go UP or DOWN.

5.1.1 Algorithm for modifying the Parameters

Pi = Probability that user will answer the question correctly, as a functionof (Ui and Qj)Uk is student id attempting the question Qk and have Ui intelligence.if( Uk taken Qk )

Pi = eUi−Qj/1 + eUi−Qj (5.2)

22

if( correct )Ui = Ui + (1 − Ui) ∗ (1 − Pi) (5.3)

Qj = Qj −Qj ∗ Pi (5.4)

elseUi = Ui− Ui ∗ Pi (5.5)

Qj = Qj + (1 −Qj) ∗ (1 − Pi) (5.6)

1

5.1.2 Scenarios for User Uk

1. If the users is highly expected to answer correctly (pi is larger, so 1-piis smaller), then he/she answers correctly. No surprise, add little intel-ligence to his bucket

Ui+ (1-Ui) * (1-Pi)

2. If the users is not expected to answer correctly (pi is smaller, so 1-piis larger), then he answers correctly, it is a surprise, add intelligenceparameter is tuned much

Ui + (1-Ui) * (1-Pi)

3. If the users is highly expected to answer correctly (pi is larger, so 1-piis smaller), then he answers wrongly, it is surprise, penalize intelligenceparameter

Ui - (Ui) * (Pi)

4. If the users is not expected to answer correctly (pi is smaller, so 1-pi islarger), then he answers wrongly, no surprise, penalize little

Ui - (Ui) * (Pi)

1Equation 5.2 is taken from Rasch Model

23

5.1.3 Scenarios for Question Qk

1. If the users is highly expected to answer correctly (pi is larger, so 1-pi issmaller), then he/she answers correctly. No surprise, Decrease questiondifficulty little bit

Qj = Qj - Qj*Pi

2. If the users is not expected to answer correctly (pi is smaller, so 1-piis larger), then he answers correctly, it is a surprise, Decrease questiondifficulty little bit

Qj = Qj - Qj*Pi

3. If the users is highly expected to answer correctly (pi is larger, so 1-piis smaller), then he answers wrongly, it is surprise, increase questiondifficulty more

Qj = Qj + (1-Qj)*(1-Pi)

4. If the users is not expected to answer correctly (pi is smaller, so 1-piis larger), then he answers wrongly, no surprise, increase question dif-ficulty more

Qj = Qj + (1-Qj)*(1-Pi)

24

Chapter 6

Software Implementation

The main centre of this project for conducting quizzes surrounds two parts:

• Client Application

• Server Module

Figure 6.1 shows the software architecture of this system

Figure 6.1: Software Architecture Diagram

25

6.1 Software Requirements

6.1.1 Server Side

• Apache Server

• PHP

• MySql

6.1.2 Client Side

• Java

6.2 Server Module

Server module is designed for Instructor, it provide web interface for creatingquizzes in classroom. Figure 6.2 shows the Server Interaction diagram of thissystemThe main component of server modules is:

• Login Module

• Create Quiz Module

• Add Question Module

• Save Response Module

• Database

• Web Interface

6.2.1 Login Module

This module provide the functionality of login, the credential for login areLDAP user_id and password.

26

Figure 6.2: Server Interaction Diagram

6.2.2 Create Quiz Module

This module provide the functionality of creating a new quiz, the main fieldsof these module is Quiz Name which is unique for each quiz whenever in-structor want to conduct quiz and the other is total number of questions.The number of questions is randomly selected through database from tableQuestion.

6.2.3 Add Question Module

This module provides the functionality of adding question into the tablenamed by Question. It also provide a field named by Question DifficultyLevel which specify the difficulty of question by default the value of this fieldis set to 0.5.

6.2.4 Database

The database is used for storing questions added by instructor, so that, laterthat can be used for quiz. All the responses submitted by client/user arereflected back into the database. Figure 6.3 shows the database relation

27

diagram of our application.The tables used are as follow:

Figure 6.3: Database Relation Diagram

Question

This table contains the entire question added by instructor. The main fieldsin this table are question_id and question_text, answer1, answer2, an-swer3, answer4, question_level and correct_answer. The question_levelfield store the difficulty level value for each question, this value change when-ever user give the response.

Login

The main fields of this table are login_id, quiz_id, timestamp and uid.Where login_id field keep the track of different users, when they login and

28

uid field store user_id of each user who login.

Quiz

This table stores the information about the quiz, which is created by in-structor. The main fields of this table are quiz_id, quiz_name, quiz_from,quiz_to, total_questions, question_scheduling, scheduling_parameters,user_rating, user_ranking, updated.

Response

This table store the responses submitted by user/client, whenever they sub-mit answer. The main fields of this table are response_id, question_id,response, login_id. The question_id is the question number and responseis the answer submitted by user/client.

Userdet

The main fields of this table are user_level and login_id. The user_levelfield store the user level intelligence value, which we get when user answerthe question.

6.2.5 Web Interface

This is the dashboard used by instructor for login, adding questions andcreating quizzes. Figure 6.4, 6.5, 6.6 shows the interface of login, addquestion and create quiz module.

29

Figure 6.4: login module

6.3 Client Module

Java Wrapper is used to communicate to Server using apache http client.Each Operation is done using http post request. The operations are

• Login

• Next Question

• Save Response

Main Class:

• Login the user given by command line arguments.

• Fetch the questions and save response for each question.

Figure 6.7 shows the flow diagram of quiz application.

30

Figure 6.5: add question module

31

Figure 6.6: create quiz module

32

Figure 6.7: flow diagram of quiz application

33

Chapter 7

Test Results

• Initially all users have same intelligence level

User Intelligencelogin_id user_level

1 0.052 0.05. 0.05. 0.05n 0.05

• Initially when question is selected from database the question difficultylevel assigned to it is 0.95

34

Question Level Values Before Quiz startsquestion_id correct_answer question_level

1 2 0.476881482803718562 2 13 4 0.38246048670578154 3 0.344231179990214965 3 16 3 17 1 18 3 19 4 110 3 111 1 0.235865158625298412 2 0.2485968245777563413 1 0.670167042309232514 1 0.47515 4 0.601492199596200616 2 0.493793115066178917 4 118 2 119 3 120 1 0.313111653123102221 2 0.477857052882043822 3 0.47523 4 124 3 0.9525 2 0.9526 4 0.9527 4 128 1 0.9529 1 0.95

35

7.1 Test Results When One User Attempt

The Quiz

• When all answers are correct

– User Login Details

Login Detailslogin_id quiz_id timestamp uid1 1 2014-06-19

13:15:13123050062

– User IntelligenceUser Intelligence

login_id user_level

1 0.95

36

– Question Level Status

Question Level After Quiz Attemptedquestion_id correct_answer question_level

1 2 0.476881482803718562 2 13 4 0.38246048670578154 3 0.344231179990214965 3 16 3 17 1 18 3 19 4 110 3 111 1 0.235865158625298412 2 0.2485968245777563413 1 0.3218450498539532414 1 0.47515 4 0.601492199596200616 2 0.2461303307314398317 4 118 2 119 3 120 1 0.313111653123102221 2 0.477857052882043822 3 0.47523 4 124 3 0.4940709090137741525 2 0.47798160998861526 4 0.47527 4 128 1 0.9529 1 0.95

37

• when all answers are incorrect

– User Login Details

Login Detailslogin_id quiz_id timestamp uid1 1 2014-06-19

13:15:13123050062

– User IntelligenceUser Intelligence

login_id user_level

1 0.05

38

– Question Level Status

Question Level After Quiz Attemptedquestion_id correct_answer question_level

1 2 0.476881482803718562 2 13 4 0.38246048670578154 3 0.67166153005486125 3 16 3 17 1 18 3 19 4 110 3 111 1 0.629888343305759612 2 0.65923736909538813 1 0.3218450498539532414 1 0.47515 4 0.786051286215092516 2 0.568325646221258517 4 118 2 119 3 120 1 0.701479046115138621 2 0.477857052882043822 3 0.47523 4 124 3 0.622973549370627425 2 0.2461303307314398326 4 0.47527 4 128 1 0.9529 1 0.95

39

7.2 Test Results When Five Users Attempt

the Quiz

• User Login Details

Login Detailslogin_id quiz_id timestamp uid1 1 2014-06-18

23:46:24123050068

2 1 2014-06-1823:46:37

123050062

3 1 2014-06-1823:46:42

123050078

4 1 2014-06-1823:46:50

123050061

5 1 2014-06-1823:47:33

123050002

• User IntelligenceUser Intelligence

login_id user_level

1 0.24395316569341652 0.90007886900062283 0.20838624899637064 0.140331066081478275 0.7857900233260007

40

• Question Level Status

Question Level After Quiz Attemptedquestion_id correct_answer question_level

1 2 0.72820251495535192 2 13 4 0.303251399067232364 3 0.088779315433861145 3 16 3 17 1 18 3 19 4 110 3 111 1 0.597360713595713112 2 0.2485968245777563413 1 0.622973549370627414 1 0.47515 4 0.601492199596200616 2 0.2461303307314398317 4 118 2 119 3 120 1 0.313111653123102221 2 0.477857052882043822 3 0.47523 4 124 3 0.4940709090137741525 2 0.47798160998861526 4 0.47527 4 128 1 0.9529 1 0.95

41

7.3 Limitations

There is some limitation of this application.

• When an user attempt all question correct, then intelligence value reachat maximum value of 0.95. If further that user attempt and give correctanswer to the question then intelligence value for that user will notincrease.

• Same problem occur in case when user give incorrect answer to allquestions, then intelligence value for that user will not decrease beyond0.05.

• It is possible that some question answer by a particular user is difficultfor him/her and he gives incorrect answer, but other user can solve thatproblem easily, but difficulty level for that question is already increase.Solution to this kind of problem is to update question difficulty afterall users responses are collected at server side.

42

Chapter 8

Conclusion

Many models have been proposed and implemented for improving learningenvironment in classroom. But, the question has been always remain thesame, how to improve effective learning in classroom, so that students maybe participative with more encouragement and overall interaction betweeninstructor and students can be increase effectively.We have proposed a model to solve this problem in some extent. Our solutionto this problem is to provide recommendation system that can recommendquestion to students according to their intelligence or IQ level. Our systemmaintains two variables Qj and Ui, Qj is question difficulty value and Ui isuser intelligence value. The values of these variables will change accordingto user responses. The central approach of our system is based on RaschModel.At the end of the quiz, we are able to characterize questions according totheir difficulties and user according to their intelligence.

43

Chapter 9

Future Work

• Find the correlation between users intelligence, so that question canbe recommended based on correlation. Like, if two users have sameintelligence then same question can be recommended to them.

• Maintain the history of each user responses, so that question can berecommended based on that history. Like, if user give correct answerfor particular question at one time moment and at next time moment,if same question is posted and user give wrong answer, then decreaseuser intelligence according to that, as it may be possible that previouslyuser gave random answer.

• Use other kinds of recommendation system or approaches that can beintegrated with current system, like update values for question difficultylevel once all questions are answered by user/client. This scenario isuseful in case, when we want to change question difficulty level afterall responses are recorded, it is because sometime question difficultylevel changes unexpectedly for other user when some other user markright/wrong answer.

• Provide some explanation of each answer to client/user, whenever theygive their responses, so that their learning can be improved.

44

Bibliography

[1] Monika Andergassen, Victor Guerra, Karl Ledermller, and Gustaf Neu-mann. Development of a browser-based mobile audience response sys-tem for large classrooms. International Journal of Mobile and BlendedLearning (IJMBL), 5(1):58–76, April 2013.

[2] Andrew Cross, Edward Cutrell, and William Thies. Low-cost audiencepolling using computer vision. In Proceedings of the 25th annual ACMsymposium on User interface software and technology, UIST ’12, pages45–54, New York, NY, USA, 2012. ACM.

[3] Brian T. Davis. Use of classroom response systems in numerically inten-sive courses. School of Technology, Michigan Technological University,page 6, 2010.

[4] Matthias Hauswirth and Andrea Adamoli. Solve & evaluate with in-forma: a java-based classroom response system for teaching java. InProceedings of the 7th International Conference on Principles and Prac-tice of Programming in Java, PPPJ ’09, pages 1–10, New York, NY,USA, 2009. ACM.

[5] Reinicke B Vetter R. Yaylacicegi U. Grove N. Herrmann, A. urespond: Aclassroom response system on the ipad. Annals of the Master of Sciencein Computer Science and Information Systems at UNC Wilmington,6(1):7, 2012.

[6] Wolfgang Hurst, Sabine Jung, and Martina Welte. Effective learn-quizgeneration for handheld devices. In Proceedings of the 9th internationalconference on Human computer interaction with mobile devices and ser-vices, MobileHCI ’07, pages 364–366, New York, NY, USA, 2007. ACM.

45

[7] L. Jackowska-Strumillo, J. Nowakowski, P. Strumillo, and P. Tomczak.Interactive question based learning methodology and clickers: Funda-mentals of computer science course case study. In Human System Inter-action (HSI), 2013 The 6th International Conference on, pages 439–442,2013.

[8] M. Jagar, J. Petrovic, and P. Pale. Auress: The audience responsesystem. In ELMAR, 2012 Proceedings, pages 171–174, 2012.

[9] Abderrahmane Lakas, Khaled Shuaib, and Mohammed Boulmalf. Acp:an interactive classroom response system for active learning environ-ment. In Proceedings of the 2006 international conference on Wirelesscommunications and mobile computing, IWCMC ’06, pages 1301–1306,New York, NY, USA, 2006. ACM.

[10] David Lindquist, Tamara Denning, Michael Kelly, Roshni Malani,William G. Griswold, and Beth Simon. Exploring the potential of mobilephones for active learning in the classroom. SIGCSE Bull., 39(1):384–388, March 2007.

[11] M. Llamas-Nistal, M. Caeiro-Rodriguez, and J. Gonzalez-Tato. Web-based audience response system using the educational platform calledbea. In Computers in Education (SIIE), 2012 International Symposiumon, pages 1–6, 2012.

[12] T. Mantoro, M.A. Ayu, E. Habul, and A.U. Khasanah. Survnvote: Afree web based audience response system to support interactivity in theclassroom. In Open Systems (ICOS), 2010 IEEE Conference on, pages34–39, 2010.

[13] M. Miura and T. Nakada. Device-free personal response system basedon fiducial markers. In Wireless, Mobile and Ubiquitous Technology inEducation (WMUTE), 2012 IEEE Seventh International Conference on,pages 87–91, 2012.

[14] Roy P. Pargas. Reducing lecture and increasing student activity in largecomputer science courses. In Proceedings of the 11th annual SIGCSEconference on Innovation and technology in computer science education,ITICSE ’06, pages 3–7, New York, NY, USA, 2006. ACM.

46

[15] Craig Prince. Integrating diverse student devices into the digital class-room. Dept. of Computer Science and Engineering University of Wash-ington, 2007.

[16] Beth Simon, Michael Kohanfars, Jeff Lee, Karen Tamayo, and QuintinCutts. Experience report: peer instruction in introductory computing.In Proceedings of the 41st ACM technical symposium on Computer sci-ence education, SIGCSE ’10, pages 341–345, New York, NY, USA, 2010.ACM.

[17] Scott Teel, Dino Schweitzer, and Steven Fulton. Braingame: a web-based student response system. J. Comput. Sci. Coll., 28(2):40–47, De-cember 2012.

[18] Palmo Thinley. Tablets as disruptive interactive learning innovation inteaching and learning environment. In Jan Herrington, Alec Couros,and Valerie Irvine, editors, Proceedings of World Conference on Edu-cational Multimedia, Hypermedia and Telecommunications 2013, pages2315–2319, Victoria, Canada, June 2013. AACE.

[19] Daniel Zingaro, Cynthia Bailey Lee, and Leo Porter. Peer instruction incomputing: the role of reading quizzes. In Proceeding of the 44th ACMtechnical symposium on Computer science education, SIGCSE ’13, pages47–52, New York, NY, USA, 2013. ACM.

47