what makes an effective teacher

29
7/30/2019 What Makes an Effective Teacher http://slidepdf.com/reader/full/what-makes-an-effective-teacher 1/29  What Makes an Effective Teacher? Quasi-Experimental Evidence  May 2010 Victor Lavy Preliminary Draft Work in progress  Special thanks go to Daniel Schuchalter for excellent research assistance.

Upload: lhaii-velasco

Post on 14-Apr-2018

221 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 1/29

 

What Makes an Effective Teacher? Quasi-Experimental Evidence 

May 2010

Victor Lavy

Preliminary Draft

Work in progress

 Special thanks go to Daniel Schuchalter for excellent research assistance.

Page 2: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 2/29

 

What Makes an Effective Teacher? Quasi-Experimental Evidence

Abstract

In this paper, I measure empirically the relationship between classroom teaching practices and student

achievement. Based on primary and middle school data from Israel, I find very strong evidence thatseveral elements of teaching practices cause student achievement growth. In particular, teaching thatemphasizes in the classroom instilment of knowledge and comprehension have very strong and positiveeffect on test scores, especially of boys and of pupils from low socio-economic background. Secondly,

 practicing in the classroom techniques that endow pupils with analytical and critical skills has also

 positive payoff though the test scores gains are significant mainly among pupils from educated families. Ialso find that transparency, fairness and proper feedback in the conduct of teachers with their students

lead to improved academic performance, especially of boys. Beyond identifying what works in theclassroom, these findings yield two insights for the debate about the merit of traditional versus modern

approaches to teaching since the teaching practices I study here capture the essence of these seeminglyrival classroom pedagogical approaches. The first is that one approach does not necessarily ‗crowd out‘the other and that the two can ‗coexist‘ in the classroom production function. The second is that it is best

to target the two teaching technologies differentially to students of different gender and abilities. Theeffect size of effective teaching practices estimated here is very large, especially in comparison to the

effect size of other potential interventions such as reducing class size or increasing school hours of instruction. 

Page 3: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 3/29

 1

1. Introduction

While teacher quality may be important, it is driven by characteristics that are difficult or 

impossible to measure. This is often the conclusion of many past and more recent studies that failed to

 produce consistent evidence linking pupils‘ achievement to observable teacher characteristics (Hanushek,

1986, Goldhaber and Brewer, 1997)). As an alternative researchers have tried recently to separate

student achievement into a series of "fixed effects," and assigning importance to teachers, schools, pupils

and so on. For example, Rockoff (2004), Rivkin et al. (2005), Kane et al. (2006), and Aaronson et al.

(2007) demonstrate substantial and persistent variation in achievement growth among students assigned

to different teachers. An even more recent strand of research tries anew to identify specific characteristic

of teachers that makes a difference for pupils' achievements. In contrast to older studies that examined

mostly the effect of demographic and educational characteristics of teachers, these new studies (for 

example, Kane and Staiger, 2008, Kane et al 2010) focus on characteristics such as cognitive ability,

content knowledge, personality traits, and personal beliefs regarding self-efficacy. In this paper, I

continue this approach but instead of focusing on particular personal teachers' characteristics, I shift the

attention to measuring what teachers do in the classroom. In particular, I measure teaching practices in

the classroom based on data not typically collected by schools or education authorities.  Based on pupils‘ 

survey and conceptual categorization of  teacher‘s pedagogical practices developed in the educational

 psychology literature (see Bloom, 1956), I summarize information based on 35 dimensions of teaching

 pedagogy in five aggregated measures of teaching practices in the classroom. These measures are (1)

Instilment of knowledge and enhancement of comprehension; (2) instilment of applicative, analytical and

critical skills; (3) instilment of capacity for individual study; (4) transparency, fairness and feedback; and

(5) individual treatment of students. I then examine which of these characteristics causes improvement in

student test score outcomes. Even though individually only a few of the over 30 teaching characteristics

have statistically significant effect on student outcomes, three of the five primary factors summarizing

cognitive and non-cognitive teaching practices have a large and statistically significant effect on

student‘s test scores.

Beyond the importance of identifying, which classroom practices are most productive for pupil‘s

learning, the evidence that I present here provide insights relevant for the ongoing debate regarding the

relative merit of traditional versus modern teaching‘s methods. In some countries policy action has taken

 place recently where modern teaching replaced traditional methods or vice versa1. Instilment of 

1  For example, the shadow education secretary in England Michael Gove's prescription for improving state schools

is to go back to learning based on memorization and comprehension, for example children reciting times tables,

learning to conjugate verbs and memorizing important dates and figures in the country history. He also claimed that

"it is often the poorest children who suffer most from trendy teaching. When synthetic phonics was abandoned as

Page 4: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 4/29

 2

knowledge and enhancement of comprehension includes a variety of classroom practices that are viewed

as the 'heart' of traditional teaching practices, requiring that students acquire knowledge through drill,

 practice and memorization (Salomon, Perkins and Theroux 2002). The Iinstilment of applicative,

analytical and critical skills aggregate measure captures the main elements of modern teaching style,

focusing on instilment of learning skills and creative thinking (Resnick, L. 1987). The evidence

 presented in this paper suggests important heterogeneity by socio-economic background of pupils in the

effect of these two teaching practices: the effect of the ‗traditional teaching‘ is generally positive but it is

three fold larger among pupils from less educated families. However, the effect of the ‗modern teaching‘

measure is positive as well but it is twice as large among pupils from educated families. The coexistence

of positive and heterogeneous effects of these seemingly contradicting teaching practices has important

 policy implication about potential improvement in pupils' knowledge through targeting of teaching

methods.

An additional interesting result in the paper is the statistically significant positive effect on

student achievement of the practice of fairness and feedback in teaching. However, I also find that the

other two remaining teaching practices have no systematic relationship with test scores improvement of 

 pupils. It is also important to note that the effects of teaching practices are mainly important for boys

while no significant effects are evident for girls and that the effect size of the three teaching practices that

are statistically significant is very large relative to other educational intervention such as reduced class

size or improved teachers training. For example, if pupils are moved from the minimum to the maximum

exposure observed in the sample of each of these three teaching practice measures, average test score in

each of the subjects would increase by 0.66 standard deviation of the test score distribution. The

contribution to this change of the ‗traditional teaching‘ practice is 0.25, that of the modern teaching

 practice is 0.14 and that of transparency, fairness and feedback is 0.27, partly because the change in

exposure is largest.

The size of the effect size of the teaching practices that were found to be effective and the

consistency of the findings reported in the paper are important as many countries are searching for ways

to promote ―teacher quality‖ or ―teacher effectiveness‖. This great interest, which partly is a response to

the way of teaching children to read because it was too authoritarian, children from book-rich backgrounds survived

 but those who were already book poor fell behind," See the Daily Telegraph 20 Oct 2007, Rachel Sylvester and

Alice Thomson interview England's shadow education secretary and more recently Michael Gove article in the

Times, March 17, 2010. In Israel, the Ministry of Education initiated a change in the opposite direction. A 2008

director gener al circular (entitled ‗pedagogical horizon‘) outlined that as of September 1 2008 teaching in post

 primary education should change from being based on memorization, repetition and practice to teaching that

emphasis development of deep understanding and of  learning and thinking skills. To facilitate this ―pedagogical the

reform included other changes, including in the curriculum and standards, teachers‘ training and in evaluation of 

students.

Page 5: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 5/29

 3

the anxiety in some countries as a result of the poor performance of their students in international tests

and comparisons such as TIMS and PISA, is not met with much reliable evidence and the findings

reported in this paper is a step in meeting this demand.

I use in the empirical work panel data of primary schools pupils in Israel observed in 2002 while

in 5th grade and later observed again in middle school while in 8 th grade. These pupils were tested in four 

subjects (English, Hebrew, math and science) both in fifth and i9n 8th grade as part of a national testing

 program. Identification is based on within pupil analysis (using pupil fixed effects) together with primary

and middle school fixed effects that net out any confounding factors at schools or individuals. Within

school/grade and across classes‘ variation in teaching practices are used to estimate the effect of interest.

In practice the within pupil estimation eliminates almost all the selection or sorting of pupils into primary

and middle schools and the estimates do not change much when primary and school fixed effects are

added to the estimated equations.

The paper proceeds as follows. I present in section 2 a brief review of the most relevant

literature, in section 3 I describe the data and in section 4 the empirical methodology. Results, robustness

checks and heterogeneity in treatment effects of teaching practices are presented and discussed in section

5 while the last section concludes and suggests policy implications.

2. Related Literature

A  long list of studies represent the effort of researchers in using non-experimental data to estimate

teacher effects on pupils learning outcomes (for example, Armour (1971), Hanushek (1976), McCaffrey

et. al. (2004), Murnane and Phillips (1981), Rockoff (2004), Hanushek, Rivkin and Kain (2005), Jacob

and Lefgren (2005), Aaronson, Barrow and Sander (2007), Kane, Rockoff and Staiger (2006), Gordon,

Kane and Staiger (2006)). There are also few studies that used random assignment to estimate the

variation in teacher effects. In that analysis, Nye, Konstantopoulous and Hedges (2004) re-analyzed the

results of the STAR experiment in Tennessee. After accounting for the effect of different classroom size

groupings, their estimate of the variance in teacher effects was well within the range typically reported in

the non-experimental literature.

Some researchers have found that teachers with stronger academic backgrounds produce larger 

 performance gains for their children (see, for example, Clotfelter et al. (2006, 2007). However, there are

also a number of studies, which do not find this relationship (e.g., Harris and Sass (2006) on graduate

course work and Kane et al. (2006) on college selectivity). A small number of studies have found a link 

 between teachers‘ scores on certification examinations and teacher effectiveness (e.g., Clotfelter et al.

(2006, 2007) and Goldhaber (2007), although Harris and Sass (2006) do not find this link).

Page 6: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 6/29

 4

Researchers recognize the possibility that non-random assignment of students to teachers could

distort measures of teacher effectiveness. Some teachers are assigned better students who would have

achieved highly in many different classrooms. Some researchers have questioned whether a teacher‘s

specific contribution can be accurately estimated given the possibility that students are assigned to

teachers based on unmeasured characteristics not captured by test scores and demographics (Rothstein

(2009)). Other researchers, recognizing the potential for bias, are more optimistic (Koedel and Betts

(2009)). One recent study compared experimental (i.e., classes randomly assigned to teachers) and non-

experimental estimates of teachers‘ effects on student achievement growth for a small sample of teachers

in Los Angeles. In that sample the non-experimental or observational measures predicted the

experimental measures with little bias —as long as the observational models controlled for each student‘s

 prior achievement (Kane and Staiger (2008)).

In a number of studies the effect of teachers in one grade fade out as students progress through

subsequent grades (McCaffrey, Lockwood, Koretz, Louis and Hamilton (2004), Kane and Staiger (2008),

Jacob, Lefgren and Sims (2008), Rothstein (2009)). Hypotheses for fade out range from artifacts of 

empirical strategy to the heterogeneity of teacher quality within schools tothe relevance of skills gained

this year for skills tested next year (Kane and Staiger (2008).

A few recent studies have found a relationship between a teacher‘s measured effect on student

achievement and overall subjective administrator ratings ((Jacob and Lefgren (2008), Rockoff and

Speroni (2009), Rockoff, Staiger, Kane and Taylor (2009)). However, those studies do not identify the

criteria or behaviors principals used to make their judgments. Using data from the early years of 

Cincinnati‘s evaluation program, Holtzapple (2003) and Milanowski (2004a and 2004b) demonstrated a

 positive relationship between teachers‘ final overall scores and student achievement.

Cooper et al. (2006) surveyed research conducted in the United States since 1987 on the effects

of homework. The authors found that all studies, regardless of type, had design flaws. However, both

within and across design types, there was generally consistent evidence for a positive influence of 

homework on achievement.

3. Data

I use in the empirical analysis two samples, the first includes 5th

grade elementary school

students and the second includes 8 th grade middle school students. Both samples include only Jewish

schools. Both samples are drawn from the Growth and Effectiveness Measures for Schools (GEMS) -

Meizav in Hebrew - datasets for the years 2002 and 2005. The GEMS includes a series of tests and

questionnaires administered by the Division of Evaluation and Measurement of the Ministry of 

Page 7: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 7/29

 5

Education.2

The GEMS is administered at the midterm of each school year to a representative 1-in-2

sample of all elementary and middle schools in Israel, so that each school participates in GEMS once

every two years.

The GEMS student data include test scores of fifth and eighth graders in math, science, Hebrew,

and English, as well as the responses of fifth through 9th grade students to questionnaires. In principle, all

students except those in special education classes are tested and required to complete the questionnaire.

The proportion of students who are tested is above 90 percent, and the rate of questionnaire completion is

roughly 91 percent. The raw test scores used a 1-to-100 scale that we transform into z-scores to facilitate

interpretation of the results.

The GEMS student questionnaire addresses various aspects of the school and learning

environment. I use the section that focuses on teaching style and practices and includes 29 items, which

are listed in the appendix. In this section, students are asked to rate in a 6-point scale ranging from one

(strongly agree) to six (strongly disagree) the extent to which they agree with a series of statements. I

grouped the individual items of the student questionnaire under five categories that describe teachers‘

 pedagogical practices in the classroom as follows: (1) instilment of knowledge and enhancement of 

comprehension (seven items); (2) instilment of applicative, analytical and critical skills (nine items) ; (3)

instilment of capacity for individual study (three items); (4) transparency, fairness and feedback (three

items); and (5) individual treatment of students (seven items). These categories of teacher‘s pedagogical

 practices are common and accepted terminology in the literature of educational psychology (see Bloom,

1956).

The student questionnaire data and test scores for the years 2002 and 2005 were linked to student

administrative records collected by the Israel Ministry of Education (identical in structure to the data

used for high school students). The administrative records include student demographics and are used to

construct all peer measures of students‘ background characteristics. Using the linked datasets, I built a

 panel for elementary schools and a panel for middle schools. The elementary school 2002 file includes

data from 606 elementary schools with test score student-questionnaire data. The middle school 2005 file

includes data for 506 schools with 8th grade student questionnaires and test scores. The panel data sample

includes students from 187 primary and middle schools. The mean characteristics of pupils included in

the panel data set are very similar to the mean characteristics of the students included in the full sample

(See Table 1).

2  The GEMS are not administered for school accountability purposes and only aggregated results at the district

level are published. For more information on the GEMS see the Division of Evaluation and Measurement website

(in Hebrew): http://cms.education.gov.il/educationcms/units/rama/odotrama/odot.htm. 

Page 8: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 8/29

 6

4. Empirical Strategy  

The structure of the GEMS allows following a sample of students from elementary schools (at 5th in

2002) to middle schools (at grade 8th in 2005).3  I take advantage of this feature and construct a

longitudinal dataset at the student level to examine how changes in teaching practices (styles and

methods) are causing changes in test scores of pupils. It is important to emphasize here that the mobility

of student that I track here is due to their transition from elementary to middle school. Also important is

the fact that there is no free school choice at the primary and assigned middle school level in Israel and

 pupils are assigned to their neighborhood primary school and middle school, the later often having a

catchment area that includes several primary schools.

Since a student fixed effect and a school fixed effect are included in the estimated regression

identification is therefore based on contrasting the change in the exposure of students to the various

teaching practices during 5th and 8th grade school, within students who followed the same transition path

from primary to middle school. More formally, I assume that the cognitive achievements of pupils in 5th 

and 8th

grade 8 are determined by the following equation:

' '

1 2ics ics s t ics cs sc ics cs y x S T    (1)

where i denote individuals, c denotes class (within a grade) and  s denotes schools. Since the school

indicator is perfectly correlated with the grade indicator (5th or 8th grade), there is no need to add a grade

effect in equation (1).ics

 y is an achievement measure for student i in class c and school s;i

  is a pupil

effect, s

   is a school effect,ics

 x is a vector of student‘s covariates that includes mother‘s and father‘s

years of schooling, number of siblings, immigration status, and ethnic origin, and indicators for missing

values in these covariates,cs

S  is a vector of characteristics of a class c in school s and it includes a set of 

variables for the average characteristics of the students in the class, the characteristics of the class

learning environment and climate (such as levels of noise and violence in the classroom, … ) ;cs

T  is a

vector of teaching practices and methods in class c and school  s, and ics  is the error term, which is

composed of a school-specific random element that allows for any type of correlation within

observations of the same school across classes and an individual random element.4 The coefficient of 

interest is θ , which captures the effects of the different teaching practices.

3  I do not link between datasets from consecutive years because almost all localities were sampled once every two

years.4 While the fixed effect coefficient in equation (1) captures much of the unobserved correlation within observations

of the same school, it is still important to account for within school correlations that are not fixed.

Page 9: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 9/29

 7

In order to estimate equation (1) I need to observe students while in 5 th and 8th grade. However,

for the estimates in equation (1) to have a causal interpretation, the unobserved determinant of 

achievement must be uncorrelated with the treatment variable. Including school fixed effects and pupil

fixed effects controls for the most obvious potential confounding factor  –  the endogenous sorting of 

students across schools. However, one may be concerned that there are within school and across classes‘ 

unobserved factors that are also correlated with changes in the teaching practices of teachers. If some

classroom characteristics are not controlled, the estimated effects of interest will be biased. Random

assignment of students and teachers to classrooms solves this problem because random assignment

 breaks the link between teaching practices and extraneous effects on the class, like unobserved peer 

quality. However, True random assignment variation is rare in an education context and unavailable in

many countries. However, students in Israel‘s primary and in middle schools are not generally grouped

into classes based on ability or family background. Actually, such practices are forbidden by law and

therefore classes in primary schools with multiple classrooms at the same grade level are typically

formed more or less on a random basis and in middle schools classes are formed in a way to create social

integration by mixing students from different socio-economic background.5

Since all classes within a

grade of equal average ability, teachers are assigned to classes more or less randomly and the possibility

that better teachers avoid assignment to classes with more poorly performing students is irrelevant as is

the possibility for ―teacher shopping‖ by parents.

All these imply thatcs

T  will be uncorrelated with the class-level shockscs conditional on a set

of school fixed effects, pupil fixed effects and the class mean-characteristics. Thus, the basic identifying

assumption in this study is that the systematic components of teaching practices in school arise only at

the school level and not at the class level. A necessary condition for the within-school estimation to work 

is, of course, that there is sufficient variance in teaching practices within a school, which is the case in

our data.

The identification strategy I use in this paper is most closely related to that of Ammermueller 

and Pischke (2009) who use a school fixed effects framework to estimate peer effect based on within

school across classes variation in peer ability. They demonstrate that conditioned on school fixed effect

class composition within a grade is random.

Evidence on the Vali dity of the I denti fi cation Strategy 

5 A 1968 education reform established in Israel a three-tier structure of schooling: primary (grades 1 – 6), middle (7 – 

9), and high school (10 – 12). The reform established neighborhood school zoning as the basis of primary enrollment

and of the integration, sometime with busing, of students out of their neighborhoods in middle school. Tacking and

setting of students into classes in primary and middle schools were made illegal by law and are strictly enforced

since.

Page 10: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 10/29

 8

The key identifying assumption I make in this paper postulates that changes in teaching practices

within a school are uncorrelated with changes in unobserved factors that could affect students‘ outcomes.

I assess here, from different angles, the feasibility of this assumption. I first discuss the assignment of 

students both between and within schools and present evidence that shed light on the question whether 

classes are formed (more or less) randomly and whether different classrooms systematically get different

resources. However, even if the variation in teaching practices within a school resembles a random

 process, these variations could be correlated with additional class-to-class changes that might affect

student outcomes. To assess this possibility, we check whether changes in teaching practices within a

school is associated with changes in student background characteristics such as parental education,

family size, ethnicity, and the proportion of new immigrants.

Students in Israel attend a primary school from school enrollment to grade six and a middle

school from grade 7 to 9. In general, there is no school choice at the primary and middle school level

except for cities that have few charter schools that allow opting out of the neighborhood schools. Primary

school assignment depends in general on the place of residence. Each middle school catchment area

includes few primary schools and this assignment is designed with an objective of achieving social

integration by mixing in middle schools pupils from different socio-economic background. However,

 parents have some means to influence the choice of schools, for example by choosing to live near the

school of their choice. The heads of the school are responsible for the assignment of students to classes

within schools. Class size cannot exceed a maximum of 40 students but extra resources are allocated to

for schools with a high share disadvantaged students or of students who recently immigrated to the

country. An important regulation of the Ministry of education does not allow grouping of students by

ability in primary and middle school [General Director Circular, March 2000, Ministry of Education].6 

Even when parents fund additional weekly instruction hours, these resources cannot be used for forming

study groups by ability (tracking) or any other criterion.7 A similar Ministry directive applies for middle

schools where classes have to be heterogeneous. For example, a Director General Circular that outlines

the responsibilities of the middle school principal relative to the responsibilities and authority of the

overall secondary school principal, states explicitly that it is the responsibility of the former to create

heteroegenous classes and that ability tracking is allowed only after 9 th grade. 8 These institutional

descriptions are supported from evidence from PIRLS 2003 based on the question in the school

6 See http://cms.education.gov.il/EducationCMS/applications/mankal/arc//s7bk3_1_8.htm. 

7 See http://cms.education.gov.il/EducationCMS/applications/mankal/arc//sc3ak3_11_9.htm. 

8 See http://cms.education.gov.il/EducationCMS/Units/Sherut/Takanon/Perek7/Chativa/.

Page 11: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 11/29

 9

questionnaire regarding whether the school forms classes based on ability. The fraction of students in

schools that report some ability grouping at the class level is close to zero. Similar evidence is obtained

from the data of TIMSS 1999 and PIRLS 2003, which included a similar question about the extent to

which classes are formed based on students‘ ability. Jakubowski (2009) who used the PIRL 2003 data to

study the effect of tracking included Israel among the countries that their primary and middle school

systems do not practice tracking of students by ability.

Following the above institutional evidence that suggest that classes are formed randomly within

schools, I used the sample of all primary and middle schools to test whether the data support this claim.

In order to test whether classrooms are formed randomly with respect to a particular student

characteristic, I perform a series of Pearson X2 tests for four characteristics, gender, father‘s years of 

schooling, mother‘s years of schooling and number of siblings. If classes are formed randomly, any

 particular student characteristic and the class the student is assigned to should be statistically

independent. Consider father‘s schooling, for example, the Pearson X2 test asks whether there are more

 pupils with high father‘s schooling in a particular class than is consistent with independence, given the

number of students in the school. Ammermuler and Piscke (2009) describe and apply this test in their 

study of peer effects. Formally, I perform the Pearson test for each school and under the assumption that

schools in Israel are independent, I also add up the test statistics for all schools to get an aggregate test

statistic described in DeGroot (1984). Obviously, the test can be carried out only on the subsample of 

schools (482 out of 605 schools) with two or more classrooms. Of the 1928 p values for primary schools,

83 were lower than 5 percent, which is only 4% of the sample. The aggregate p values for each

characteristic are way above 20 percent. The middle school data yield similar results. Overall, I conclude

that there is no evidence of systematic formation of classrooms with respect to the four measures of 

student‘s family background measures.

The second question I investigate in this section is whether classrooms that differ in teaching

 practices are different in pupils‘ characteristics as well. In order to test whether classrooms teaching

 practices are statistically independent with respect to each of the student characteristic, I run balancing

tests as done in randomized trials. Table 2 presents estimates from regressions when the dependent

variable is a student characteristic and the explanatory variable is one of the five teaching practice

measures. The sample used in these regressions includes both 5th

and 8th

grade students. In appendix

Table A1, I report balancing tests results based separately on primary schools and in Table A2 separately

for middle schools. These results are very similar to those obtained from the pooled data of 5 th and 8th 

grade pupils and therefore I discuss below only the latter.

I present in Table 2 results from two specifications. The first is an OLS regression without any

additional control variables, the second includes primary and middle school fixed effects. I cannot add

Page 12: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 12/29

 10

individual fixed effects because given the nature of the background pupil characteristics there are very

few cases where a characteristic has changed between 5th

and 8th

grade. The background characteristics

are father‘s years of schooling, mother‘s years of schooling, number of siblings, an indicator of recent

immigration and an indicator of weather a pupil‘s parents are native Israeli.

Table 2 suggests little evidence that students of different family backgrounds are more likely to

 be in classes with certain teaching practices, conditional on the school they attend. Only three of the 25

 balancing coefficients‘ estimates presented in Table 3 are significantly different zero when school fixed

effects are included in the regressions. This pattern is in sharp contrast to the balancing estimates

obtained from the OLS equations (without school fixed effects) that yield 25 estimates, which are all

significantly different from zero. The pattern of selection between schools with respect to parental

schooling is negative, namely schools with large enrolment of pupils with higher father or mother year of 

schooling have lower intensities of all five teaching practices. The sign of this selection remains negative

even after adding the school fixed effects to the regressions though remarkably only one of the ten

 balancing estimates of parental schooling remains significantly different from zero. The between school

selection pattern with respect to number of siblings is positive: the higher the number of siblings the

lower are the intensities of all five teaching practice measures. This is an opposite selection pattern in

comparison to the negative one seen for parental schooling. As higher number of siblings and low

 parental schooling are both proxies for lower socio economic status, the opposite sign of selection that

they reveal is contradictory. Similarly inconsistent is the positive selection pattern in the distribution of 

teaching practices between schools based on the balancing tests estimates of the proportion of immigrant

 pupils. I view these inconsistencies in the selection patterns across the various socio-economic proxies as

suggestive evidence that even across schools the variation in teaching practices is not an indication of 

clear meaningful selection that can confound in certain direction the effect of teaching practices on

 pupils‘ academic outcomes. However, I should emphasize again that once school fixed effects are added

in balancing regressions all signs of potential selection disappear.

The evidence presented above largely confirms that classes in the sample schools are formed

roughly randomly within schools. There is little evidence that students of different family backgrounds

are more likely grouped in certain classes conditional on the school they attend. However, even if 

classrooms are formed randomly, they may still receive other school resources differentially. For 

example, a class may end up with more children from less advantaged family backgrounds purely by

chance, and the school might assign this class a smaller class size. In order to shed light on this question,

I ran a set of regressions of teaching practice variables described in the previous section on class size and

its square. The last two columns in Table 2 show that there are no meaningful correlations between class

size and any of the five teaching practice variables.

Page 13: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 13/29

 11

5. Results

I now turn to the analysis of the effects of the various measures of teachers‘ teaching practices on

students‘ test scores. Table 3 presents descriptive statistics for the five teaching practices measures. In

the appendix, I present the items that I average into each of these five indices. Even though each of the

items ranges from one to six, when aggregated into the five teaching practice measures the range is

narrowed, from about 1.5-2 to about 5-5.5 though there are some cases when the maximum value of the

measure is six. There are not significant differences in the descriptive statistics of the panel and full

samples.

Table 4 reports estimates based on pooling all four subjects together and each of the

specifications in the table include subject fixed effects indicators. I report the effect of each category of 

teaching practice. In Table A3 in the appendix, I also report the estimates for all individual items that I

use to construct the aggregate teaching practice measures. Most of these estimates are not precisely

measured and some have negative values, partly because of the high correlation among the various items.

It is therefore appropriate and necessary to aggregate the items in several principal components as I do in

this paper. As there is no prior information to justify a particular weighting, I assign equal weight to all

items grouped into a given teaching style characteristic as this provides a more transparent interpretation.

I computed the class level mean of these teaching practice characteristics for each student while

excluding his own answer. Results were not different when I used instead measures based on means that

included also students‘ own answer .

The first and second columns in Table 4 report OLS estimates from an equation that included as

control variables the pupil‘s background characteristics (gender, father and mother year of schooling,

number of siblings, an indicator of immigration status, five indicators of ethnic origin (Europe/America,

Asia/NorthAfrica, Soviet Union, Ethiopia and Israel). I also included as the controls the class level

means of all these indicators and also of measures of the class climate and learning environment such as

level of noise, violence, and so on [add names of these variables]. In column 1, the estimates are from

separate regression where each of the teaching practices enters as a single treatment variable. In column

2, the estimates are from one regression that includes all the teaching practice measures as multiple

treatments. In columns 3 and 4, I omit the set of individual characteristics and include instead individual

fixed effects. In the specification presented in columns 4 and 5, I include also primary and middle school

fixed effects in addition to the controls included in the specification of columns 3-4.

Focusing first on the OLS estimates of the effects on teachers‘ pedagogical methods, we see that

the sign and precision of each the various measures is similar in columns 1-2, respectively. The estimated

coefficients of T1 and T2 are positive thought only those of T1 are precisely measured. The estimates of 

T3-T5 are negative though only those of T5 are statistically significant. However, adding to the equation

Page 14: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 14/29

 12

a pupil‘s fixed effect leads to a large change in the size and sign of the various estimates of the teaching

 practice measures. Three estimates are positive and statistically significant. Focusing on the specification

when all treatment measures are included jointly in the regression we observe the following: the estimate

of T1 (Instilment of Knowledge and Enhancement of Comprehension) drops from 0.365 to 0.140 but its

standard error declines as well, so actually its t-statistics is unchanged. The estimate of T2 (Instilment of 

Analytical and Critical Skills) is practically unchanged at 0.123 (sd= 0.039) relative to its respective OLS

estimate. The estimate of T4 (Transparency, Fairness and Feedback) is reversed in sign and it is now

 positive (0.089) and significant (sd=0.025). The estimates of the other two teaching practice measures

(individual treatment of students and instilment of capacity for individual study) are actually negative

and significant. Adding the primary and school fixed effects to the regression leaves the results almost

unchanged: the point estimates of T1is only marginally lower, the estimate of the T4 is unchanged at all

and that of T2 declines by half. The point estimates of the remaining two measures are still negative but

they are now much smaller and are not statistically different from zero.

There are two remarkable features of the estimates reported in columns 5-6 that I should note.

First is the similarity between the estimates when only one of the teaching practice measures is used a

treatment measure and when all five are used jointly. This pattern suggests that there is very little omitted

variable bias when four of the five teaching practice measures are left out of the equation, probably

 because they are not highly correlated. The implication is that if it is a selection that derives the results

about the positive effect of T1, T2 and T4, it must be very different for each of these five measures

which is very unlikely. A second important feature of the estimates in column 6 is their quantitative

similarity to those presented in 4. The fact that adding the school fixed effects does not change much the

 point estimates of T1 and T4 is an indication that the selection in the distribution of teachers' teaching

style across schools and classes is practically accounted for by adding to the regression the pupil fixed

effects. Adding in addition the primary and middle school fixed effects helps only marginally to uncover 

the causal effect of teaching style because condition on pupil fixed effects the variation within school

across classes is practically unrelated to potential outcome. I should add and emphasize that these

robustness to adding the school fixed effects is not because they are not adding significantly explanatory

 power to the regression because they do (as seen by F tests). Actually some of the estimates of the other 

teaching practice measures do change when adding the school fixed effects, for example that of T2, but

not those of T1 and T4. It is also important to note that the estimated effect do not change at all when I

use a sample that includes only schools with at least 10 students observed in the panel data. This

restriction allow for more precision in estimating the school fixed effects and it is important that the

estimated effects of T1-T5 are not sensitive to this sample restriction.

Page 15: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 15/29

 13

Overall, the evidence in Table 4 is strongly suggestive of the fact that three of the five teaching

style and methods have positive and meaningful effects on pupil‘s learning. The most important in t erm

of effect size is the indicator of the extent that teachers make sure that their students know the material

and understand it by using examples, memorization techniques, homework and class work and so on.

Moving from the minimum (value 2.5) to the maximum (value 6.0) of this teaching practice measure

observed in the data changes test score by an average of 0.43 standard deviation of the test score

distribution. Second in order of importance is the teacher extent of being transparent, faire and providing

feedback to students. Moving from an environment with relative lack of these attributes of teachers

(value of 2.3) to an environment when it is fully present (value of 6) improves test scores by 0.33 of a

standard deviation. The effect of enhancing the analytical and critical skills of pupils from the minimum

to the maximum observed I the data is of smaller magnitude, ‗only‘ 0.23 of a standard deviation. Adding

to the maximum level observed in the data all three aspect of teaching practice to a teacher not having

any of them (measured by the minimum observed in the data) will increase pupils‘ test score by almost

one (0.99) standard deviation. A more realistic simulation is to compute the effect size of a change of 

each of the three teaching practice‘s intensity from the mean in the sample (about 3) to the best possible

(about 6). This simultaneous change will improve average test score in each subject by 0.83 of a standard

deviation.

The results reported so far assume that the effect of each of the teaching practices is the same for 

all subject but these effects may vary by subjects. In Table 5, columns 1-2, I present evidence based on

 pooling math and science test scores and in columns 3-4, I present evidence based on pooling Hebrew

and English test scores. I view these groupings as less restrictive than pooling all four subjects together 

 because it is very plausible that pedagogy and teaching style are more similar for the two subjects that

are intensive in math and rigorous analysis and for the two languish subjects. The comparison of the

estimates in columns 2 and 4 reveals a surprisingly close similarity for two of the three measures that

have positive effect on test score. The knowledge/comprehension measure is 0.159 for math and science

and 0.135 for Hebrew and English. The transparency/fairness/feedback measure is 0.081 for both set of 

subjects. This suggests that pooling all four subjects in estimation is not too restrictive.

As noted above the measure of individual treatment of students has a negative sign in all

specification. It is important to note that dropping this variable from the regression did not change the

 point estimates of the other measures.

Heterogeneous Tr eatment Ef fects 

To gain further insights on the extent of effects of teaching styles and methods on test score of 

students, I explore heterogeneous effects across different dimensions. In Table 6, I report heterogeneous

Page 16: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 16/29

 14

treatment effects of the teaching style measures for two sub-samples stratified by gender and for two

samples stratified by the average years of schooling of both parents (average above or below 12 years of 

schooling).9 The results clearly show that the average positive impact of the teaching style measures

reflected mainly the effect of these variables on boys. This is particularly the case for the

knowledge/comprehension measure for which the sample of boys yields an estimated effect of 0.154 and

for the transpanency/fairness/ feedback measure with an estimated effect for boys of 0.187. These two

estimates reflect very large effect size on test score of boys. The respective estimates for girls are

 practically zero. The result that suggests that the quality of teaching affect boys much more than girls is

somewhat surprising because several studies have shown that other schooling intervention either affected

only girls (for example, pupil's monetary incentives) or affected both gender equally (for example,

teacher's financial incentives or class size reduction).

 Next, we split the sample into two groups defined by parental schooling. Since the variable

father‘s years of schooling has less missing values I used it to define the two groups though the evidence

 based on mother years of schooling is very similar. Pupils with fathers who completed secondary or 

higher schooling are grouped in the high education group and others were assigned to the low education

group. The estimated effect of the teaching style variables are presented in column 3 of Table 6 for the

high education group and in column 4 for the low education group. The estimates suggest that both

groups benefit from some of the characteristics of the teaching styles of their teachers but not in the same

way. For the high education group all the first three measures have positive and significant effects.

However, what stands out in comparison of these estimates to those based on the full sample is that the

estimated effect of ‗instilment of analytical and critical skills‘ is twice larger and much more precisely

measured. On the other hand, for the lower education group the estimated effect of this same variable is

 practically zero while the estimated effect of ‗instilment of knowledge and comprehension‘ is

dramatically higher from that obtained from the full sample, 0.218 versus 0.124. If we accept the

assumption that parental schooling is a good proxy for pupil ability than this pattern of results suggest

that high ability pupils benefit relatively more from a learning style that instill analytical and critical

skills while kids of lower ability benefit more from a teaching style that emphasize memorization and

other techniques that instill knowledge and comprehension.

6. Conclusions

9 Students with missing values in parental education (4 percent of the total sample) are excluded from this analysis.

The results are not sensitive to the inclusion of these students in the low or high education group. We also estimate

heterogeneous treatment effects by stratifying the sample by father's or mother's schooling and we obtain very

similar results to those based on the stratification by the mean of parental schooling reported here.

Page 17: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 17/29

 15

In this paper, I measure empirically the relationship between classroom teaching practices and student

achievement. I find very strong evidence that several elements of teaching practices cause student

achievement growth. In particular, teaching that emphasizes in the classroom instilment of knowledge

and comprehension have very strong and positive effect on test scores, especially of boys and of pupils

from low socio-economic background. Secondly, practicing in the classroom techniques that endow

 pupils with analytical and critical skills has high payoff as well, especially among pupils from educated

families. Additionally, transparency in evaluation of pupils, proper and timely feed back to students and

fairness in assessing pupils also lead to cognitive achievement gains, especially of boys.

The evidence I presented in the paper provide important insights about what does and what does

not work in the classroom and it is among the first set of results that clearly identify actions of teachers

that have pay off versus others activities that do not. However, this study has policy relevance also

 because its evidence shed light on the merit of traditional versus modern approaches to teaching. The

contrast between these two approaches featured in recent policy debate and educational reforms in few

countries and this study is perhaps the first to demonstrate that one approach does not have to ‗crowd

out‘ the other  and that the two can ‗coexist‘. The estimated heterogeneity in treatment effects of the two

styles and essence of teaching that I estimated in this paper implies that it is best to target certain

teaching practices to relevant ‗customers‘ and also mix the two in the classroom.

The effect size estimated for some of teaching practices are truly impressive, especially in

comparison to the effect size of other potential interventions such as reducing class size, increasing

school hours of instruction and providing more teacher training. These three alternative interventions and

 possible other educational programs are much more expensive or difficult to implement than ‗installing‘

the appropriate teaching practices in the classroom.

Page 18: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 18/29

 16

Appendix: Questionnaire details items

I nstilment of knowledge and enhancement of comprehension  

1.  The teachers give exercises and assignments that help memorize the material.2.  The teachers ask many questions in class that check whether we know the material well.

3.  The teachers commend students who know the material well.4.  The teachers provide many examples that help understand the material.5.  The teachers hold discussions in class that help understand the material.

6.  During lessons, the teachers ask many questions that check whether we understand the materialwell.

7.  I understand the teachers' scholastic requirements well.

I nstilment of analytical and cri tical skil ls 

1.  The teachers give exercises and assignments whose answers have not been studied in class andare not in the textbooks. 

2.  The teachers require that we use what we have studied to explain various phenomena.3.  The teachers ask that we find new examples by ourselves for the material we have studied.

4.  The teachers ask that we try to find several ways to solve a certain problem.5.  The teachers teach us to find a single common explanation for different phenomena.6.  The teachers give assignments where it is required to analyze material and to relate it to other 

things we have studied.7.  When there are several ways to solve a problem, the teachers require that we check them all and

find the best one.

8.  The teachers expect us to ask ourselves whether what we have learned is correct.9.  The teachers teach us how to know whether information we have found is important, relevant

and can be used.

I nstilment of capacity for individual study 

1.  The teachers teach us how to learn new topics by ourselves.

2.  The teachers require students to utilize many and varied sources of information (newspapers, books, databases etc.).

3.  The teachers teach us to observe our environment and to follow phenomena that occur in it.

Transparency, fair ness and feedback 

1.  The teachers explain to me exactly what I have to do to improve my studies.

2.  The teachers explain according to what they determine the grades / assessments.3.  The teachers often tell me what my situation is regarding schoolwork.

I ndividual treatment of students 

1.  The teachers know what the educational difficulties of each student are.

2.  When a student has difficulty with a certain topic, the teachers give him more time to study it.

3.  The teachers give homework to every student according to his place in the material.4.  The teachers help every student to learn topics interest him.5.  The teachers give me a feeling that if I make an effort I will succeed more at studies.6.  When a student fails, the teachers encourage him to try again.

7.  The teachers always assist me when I need help with studies.

Page 19: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 19/29

 17

7. References

Ammermueller, Andreas and Jorn-Steffen Pischke, (2009) "Peer Effects in European Primary Schools:

Evidence from PIRLS," Journal of Labor Economics, vol. 27, no. 3, pp. 315-348.

Aaronson, Daniel, Lisa Barrow and William Sander (2007) ―Teachers and Student Achievement in

Chicago Public High Schools‖ Journal of Labor Economics Vol. 24, No. 1, pp. 95-135.

Bloom, Benjamin S. (ed.), Taxonomy of Educational Objectives, Handbook 1: Cognitive Domain, New

York, N.Y: Longmans, Green and co.

.Boyd, D., Lankford, H., Loeb, S., Rockoff, J., and Wyckoff, J. (2008a) ―The Narrowing Gap in  New

York City Teacher Qualifications and its Implications for Student Achievement in High-Poverty

Schools‖, NBER Working Paper #14021.

Brunello Giorgio and Daniele Checchi, ―Does School Tracking Affect Equality of Opportunity? New

International Evidence‖ Discussion Papers Series, Forschungsinstitut zur Zukunft der Arbeit

Institute for the Study of Labor September 2006

Clotfelter, C.T., Ladd, H.F., Vigdor, J.L. ―Teacher -Student Matching and the Assessment of Teacher 

Effectiveness‖ Journal of Human Resources 41(4):778-820 (2006).

Clotfelter, C.T., Ladd, H.F., Vigdor, J.L. (2007) ―How and Why Do Teacher Credentials Matter for 

Student Achievement?‖ NBER Working Paper 12828.

Cooper Harris, Jorgianne Civey Robinson and Erika A Patall, "Does Homework Improve Academic

Achievement? A Synthesis of Research, 1987 – 2003", Review of Educational Research, Vol. 76,

 No. 1, 1-62 (2006).

Danielson, Charlotte., & Thomas L. McGreal. (2000). Teacher Evaluation to Enhance Professional

Practice. Alexandria, Va.: Association for Supervision and Curriculum Development.

DeGroot, Morris. 1984. Probability and statistics. 2nd ed. Reading, MA: Addison-Wesley.

Donaldson, Morgaen L. (2009). So long, Lake Wobegon? Using Teacher Evaluation to Raise Teacher 

Quality. Washington, DC: Center for American Progress.

Goe, Laura. and Andrew Croft. (2009). Methods of Evaluating Teacher Effectiveness. Washington, DC:

 National Comprehensive Center for Teacher Quality.

Gordon, Robert, Thomas J. Kane and Douglas O. Staiger. (2006). ―Identifying Effective Teachers Using

Performance on the Job‖ Hamilton Project Discussion Paper, Published by the Brookings

Institution.

Goldhaber, D. and Brewer, D. (1997) ―Why Don't Schools and Teachers Seem to Matter? Assessing the

Impact of Unobservables on Educational Productivity‖ Journal of Human Resources 32(3): 505-

523.

Page 20: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 20/29

 18

Cantrell, S., J. Fullerton, T.J. Kane, and D.O. Staiger, ―National Board Certification and Teacher 

Effectiveness: Evidence from a Random Assignment Experiment,‖ working paper, March 2007. 

Freiberg, H. J. (Ed.). (1999). School climate: Measuring, improving and sustaining healthy learning

environments. Philadelphia, PA: Falmer Press.

Fraser, B. J. (1998). ―Classroom Environment Instruments: Development, Validity, and Applications‖,

 Learning Environment Research, Vol. 1 No. 1.

Gordon, Robert, Thomas J. Kane and Douglas O. Staiger, (2006) ―Identifying Effective Teachers Using

Performance on the Job‖ Hamilton Project Discussion Paper, Published by the Brookings

Institution.

Hanushek, Eric A. (1971). ―Teacher Characteristics and Gains in Student Achievement; Estimation

Using Micro Data‖. American Economic Review, 61, 280-288.

Hanushek, Eric A. "The Economics of Schooling: Pro- duction and Efficiency in Public Schools."

 Journal of Economic Literature, September 1986, 24(3), pp. 1141-77.

Hanushek, Eric A., John F. Kain, Steven G. Rivkin, ―Why Public Schools Lose Teachers‖,  Journal of 

 Human Resources 39(2), Spring 2004b.

Jacob, Brian and Lars Lefgren (2005) ―Principals as Agents: Student Performance Measurement in

Education‖ NBER Working Paper No. 11463. 

Jacob, B.A., L. Lefgren, and D. Sims, ―The Persistence of Teacher -Induced Learning Gains,‖ NBER 

working paper #14065, June 2008.

Jakubowski Maciej "Early tracking and achievement growth" Directorate for Education, OECD,

December 2009.

Kane, Thomas J., Jonah Rockoff and Douglas Staiger, (Forthcoming) ―What Does Certification Tell Us

about Teacher Effectiveness?: Evidence from New York City‖  Economics of Education Review 

(Also NBER Working Paper No. 12155, April 2006).

K onstantopoulos, S. ―How Long Do Teacher Effects Persist?‖ IZA Discussion Paper No. 2893, June

2007.

Murnane, R. J. & Phillips, B. R. (1981). ―What Do Effective Teachers of Inner-City Children Have in

Common?‖ Social Science Research, 10, 83-100.

Resnick, L. (1987), Education and Learning to Think, Washington D.C.: NationalAcademy Press.

Rockoff, J. E. (2004) ―The Impact of Individual Teachers on Student Achievement: Evidence from Panel

Data,‖ American Economic Review, 94(2): 247-252.

Rockoff, J.E. (2008) ―Does Mentoring Reduce Turnover and Improve Skills of New Employees?

Evidence from Teachers in New York City‖ NBER Working Paper 13868. 

Page 21: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 21/29

 19

Rockoff, Jonah E. and Cecilia Speroni. (2009). ―Subjective and Objective Evaluations of Teacher 

Effectiveness.‖ Papers and Proceedings of the American Economic Association, Forthcoming.

Rothstein, J., ―Teacher Quality in Educational Production: Tracking, Decay, and Student Achievement,‖

Princeton University Working Paper, May 2008.

Rivkin, S., Hanushek, E. A. and Kain, J. (2005) ―Teachers, Schools, and Academic Achievement,‖ 

 Econometrica, 73(2): 417 – 458.

Salomon, g. Perkins, D. Theroux, P, 2001 Comparing Traditional Teaching and Student Centered,

Collaborative Learning [Online, accessed 14/6/02] URL:

http://shaw.ca/priscillatheroux/collaborative.html.

Page 22: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 22/29

Means,

sample of all

5th

grade

students in

2002

Means, panel

sample of 5th

grade

students in

2002

Means, panel

sample of 5th

grade

students in

2002: schools

with at least

5 students

T-test,

differences

 between

means in

column 1 and

column 2

T-test,

differences

 between

means in

column 1 and

column 3

12.541 12.749 12.831 0.244 0.331(0.136) (0.159)

12.791 12.979 13.008 0.221 0.248

(0.122) (0.143)

2.399 2.281 2.289 -0.139 -0.126

(0.064) (0.076)

0.499 0.522 0.529 0.027 0.034

(0.012) (0.014)

0.127 0.171 0.159 0.051 0.036

(0.014) (0.017)

0.515 0.451 0.452 -0.076 -0.072

(0.017) (0.020)

 Number of schools 606 506 187

 Number of pupils 35059 5268 4375

Table 1: Descriptive Statistics and Mean Differences Between Panel and Full Sample

Note: Each figure presented in columns 4 and 5 are estiamted in a separate regression. Standard errors

clustered by school.

Parents born in Israel

 Number of siblings

Father's years of schooling

Mother's years of 

schooling

Gender (female=1)

Immigration status

(immigrant=1)

Page 23: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 23/29

OLS School FE OLS School FE OLS School FE

-0.692 -0.022 -0.577 0.249 -0.290 -0.026

(0.244) (0.249) (0.226) (0.235) (0.115) (0.107)

-1.025 -0.365 -0.936 -0.207 -0.280 -0.011

(0.248) (0.254) (0.220) (0.235) (0.115) (0.122)

-0.412 -0.094 -0.394 -0.030 -0.179 0.025

(0.169) (0.153) (0.153) (0.129) (0.080) (0.074)

-1.408 -0.242 -1.143 -0.160 -0.325 -0.005

(0.217) (0.175) (0.188) (0.150) (0.114) (0.065)

-0.531 -0.356 -0.502 -0.229 -0.045 0.055

(0.156) (0.206) (0.140) (0.175) (0.069) (0.071)

0.089 0.066 -0.072 -0.013 -3.866 0.591

(0.021) (0.030) (0.031) (0.031) (0.749) (0.775)

0.097 0.082 -0.073 -0.055 -4.581 0.754

(0.028) (0.032) (0.035) (0.031) (0.827) (0.868)

0.051 0.019 -0.036 -0.022 -3.517 0.364

(0.019) (0.021) (0.025) (0.023) (0.545) (0.590)

0.107 0.018 -0.123 -0.018 -0.768 0.638

(0.020) (0.018) (0.026) (0.019) (0.782) (0.498)

0.045 0.031 -0.040 -0.010 -3.922 0.753

(0.017) (0.018) (0.024) (0.024) (0.485) (0.612)T5: Individual treatment of students

Note: Each coefficient in this table is estimated in a separate regressoin. Standard errors are clustering by

school. The sample is schools with 5+ students.

Father's years of 

schooling

Mother's years of 

schooling

Number of siblings

Table 2: Balancing Tests, Pooled Fifth and Eighth Grade Pupils

T2: Instilment of analytical and

critical skills

T3: Instilment of capacity for 

individual study

T4: Transparency, fairness and

feedback 

Class sizeRecent immigration Parents born in

Israel

T1: Instilment of knowledge and

enhancement of comprehension

T5: Individual treatment of students

T1: Instilment of knowledge and

enhancement of comprehension

T2: Instilment of analytical and

critical skills

T3: Instilment of capacity for 

individual study

T4: Transparency, fairness and

feedback 

Page 24: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 24/29

Mean Min Max Mean Min Max

 A: Panel sample

4.8 3.7 5.7 4.2 2.5 5.2

(0.3) (0.4)

.

3.9 2.9 5.3 3.3 2.4 4.2

(0.3) (0.3)

3.9 2.2 5.6 3.1 1.9 4.4

(0.4) (0.4)

4.2 2.5 5.5 4.0 2.4 5.2

(0.4) (0.4)

4.0 2.8 5.5 3.0 1.7 4.8

(0.4) (0.4)

B: Full sample

4.8 3.3 6.0 4.2 2.5 6.0

(0.3) (0.3)

3.9 2.0 5.8 3.3 2.4 5.3

(0.3) (0.3)

3.9 1.8 6.0 3.1 1.4 5.7

(0.4) (0.4)

4.1 2.3 6.0 4.1 2.3 6.0

(0.4) (0.4)

4.0 2.1 5.9 3.0 1.6 5.9

(0.4) (0.4)

Note: The figures in the table are based on the full sample of 5th (in 2002) and 8th (in 2005) grade pupils.

Standard deviations are presented in parenthesis.

grade 5 grade 8

T1:Instilment of knowledge and

enhancement of comprehension

T2:Instilment of analytical and

critical skills

T3:Instilment of capacity for 

individual study

T4:Transparency, fairness and

feedback 

T4:Transparency, fairness and

feedback 

T5:Individual treatment of students

Table 3: Descriptive Statistics of the Teaching Practices Measures

T1:Instilment of knowledge and

enhancement of comprehension

T2:Instilment of analytical and

critical skills

T3:Instilment of capacity for 

individual study

T5:Individual treatment of students

Page 25: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 25/29

Each measure

separately

All measures

combined

Each measure

separately

All measures

combined

Each measure

separately

All measures

combined

0.268 0.365 0.173 0.140 0.166 0.124

(0.116) (0.124) (0.035) (0.040) (0.049) (0.057)

. .0.070 0.115 0.101 0.123 0.099 0.063

(0.079) (0.098) (0.028) (0.039) (0.044) (0.056). .

-0.028 -0.003 -0.004 -0.059 0.014 -0.033

(0.060) (0.075) (0.021) (0.028) (0.032) (0.039). .

-0.021 0.014 0.088 0.089 0.106 0.089

(0.052) (0.062) (0.022) (0.025) (0.032) (0.036). .

-0.264 -0.410 -0.020 -0.124 0.036 -0.055

(0.061) (0.078) (0.027) (0.034) (0.040) (0.050)

N 26867 26971 26971

Sample 5+ reffers to the minimum number of pupils per school within the panel sample (smaller schools are not included in the sample).

Note: All regressiosns include as control variables indicators for the different subjects, class mean of parental education, gender, number of 

siblings, immigrant status and ethnic background. as well as means of eight variables that measure the class climate such as level of noise

during lessons, incidence of violence and bulling and so on. The regressions include also as controls the student individual characteristics.The regressions reported in columns 1-4 include also subject indicators as control variables. The sample is 5+.

OLS Pupil fixed effect Pupil and school fixed effect

Table 4: Estimates of the Effect of Teachers’ Teaching Practices on Pupils’ Test Scores

T5: Individual treatment of students

T3: Instilment of capacity for 

individual study

T1: Instilment of knowledge and

enhancement of comprehension

T2: Instilment of analytical and

critical skills

T4: Transparency, fairness and

feedback 

Page 26: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 26/29

Each measure

separately

All measures

combined

Each measure

separately

All measures

combined

0.162 0.159 0.207 0.135

(0.067) (0.078) (0.069) (0.080). .

0.061 0.030 0.142 0.091

(0.060) (0.076) (0.061) (0.078)

. .-0.001 -0.019 0.032 -0.050

(0.043) (0.054) (0.045) (0.056). .

0.083 0.081 0.124 0.081

(0.043) (0.050) (0.044) (0.051). .

-0.019 -0.111 0.098 0.006

(0.055) (0.068) (0.057) (0.070). .

N 13545 13538 13441 13433

Table 5: Estimates of the Effect of Teachers’ Teaching Practices on Pupils’ Test Scores

T4: Transparency, fairness and

feedback 

T5: Individual treatment of students

Pupil and school fixed effect:

math and science only

Pupil and school fixed effect:

Hebrew and English only

Note: All regressiosns include as control variables indicators for the different subjects, class mean of 

parental education, gender, number of siblings, immigrant status and ethnic background. as well as

means of eight variables that measure the class climate such as level of noise during lessons, incidence

of violence and bulling and so on. The regressions include also as controls the student individual

characteristics. The regressions reported in columns 1-4 include also subject indicators as control

variables. The sample is 5+ and it refers to the minimum number of pupils per school within the panel

sample (smaller schools are not included in the sample)

T1: Instilment of knowledge and

enhancement of comprehension

T2: Instilment of analytical and

critical skills

T3: Instilment of capacity for 

individual study

Page 27: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 27/29

Male Female Father's years

of schooling ≥

12

Father's years

of schooling

< 12

Sample of non

immigrants

0.154 0.034 0.117 0.218 0.135

(0.088) (0.086) (0.067) (0.192) (0.062)

.0.060 0.072 0.118 -0.043 0.046

(0.090) (0.083) (0.066) (0.185) (0.061).

-0.061 -0.037 -0.079 0.088 -0.054

(0.063) (0.056) (0.047) (0.123) (0.043).

0.187 0.008 0.079 0.105 0.099

(0.058) (0.051) (0.043) (0.114) (0.040).

0.071 -0.150 -0.061 0.041 -0.079

(0.077) (0.072) (0.057) (0.173) (0.053).

N 12626 14345 20935 6036 22932

Note: The Estimates in this table are based on regressions that include pupil and primary and middle schools fixed effects.

Table 6: Heterogeneity in Effect Estimates of Teachers’ Teaching Practices on Pupils’ Test Scores

T4: Individual treatment of students

T1: Instilment of knowledge and

enhancement of comprehension

T2: Instilment of analytical and

critical skills

T5: Instilment of capacity for 

individual study

T3: Transparency, fairness and

feedback 

Page 28: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 28/29

OLS School FE OLS School FE OLS School FE

-2.951 -0.870 -2.342 -0.105 -0.942 0.029

(0.470) (0.455) (0.376) (0.363) (0.316) (0.185)

. . . . . .

-2.773 -0.554 -2.203 0.030 -0.595 -0.128

(0.386) (0.414) (0.345) (0.301) (0.225) (0.144)

. . . . . .

-1.794 -0.317 -1.466 0.019 -0.529 -0.021

(0.314) (0.304) (0.255) (0.233) (0.157) (0.112)

. . . . . .

-1.807 -0.024 -1.475 0.043 -0.448 -0.011

(0.287) (0.249) (0.246) (0.168) (0.167) (0.080). . . . . .

-2.142 -0.351 -1.672 0.037 -0.318 -0.160

(0.302) (0.284) (0.274) (0.214) (0.186) (0.101)

0.173 -0.003 -0.129 0.029 0.867 2.222

(0.039) (0.045) (0.058) (0.061) (1.738) (1.326)

. . . . . .

0.204 0.045 -0.162 0.004 0.047 0.322

(0.042) (0.033) (0.054) (0.056) (1.400) (0.723)

. . . . . .0.128 0.027 -0.115 -0.028 -0.148 1.571

(0.030) (0.027) (0.039) (0.042) (1.025) (0.702)

. . . . . .

0.109 -0.012 -0.129 -0.030 -0.545 0.364

(0.028) (0.022) (0.036) (0.030) (0.989) (0.668)

. . . . . .

0.127 0.011 -0.125 -0.038 -2.316 1.310

(0.030) (0.027) (0.041) (0.041) (1.017) (0.739)

T3: Instilment of capacity for 

individual study

T4: Transparency, fairness and

feedback 

T5: Individual treatment of students

T1: Instilment of knowledge and

enhancement of comprehension

T2: Instilment of analytical and

critical skills

T3: Instilment of capacity for 

individual study

T4: Transparency, fairness and

feedback 

T5: Individual treatment of students

Table A1: Balancing Tests, Fifth Grade

Father's years of 

schooling

Mother's years of 

schooling

Number of siblings

Note: Each coefficient in this table is estimated in a separate regression. Standard errors are clustered by school.

The sample includes schools with 5+ students.

Recent immigration Parents born in

Israel

Class size

T1: Instilment of knowledge and

enhancement of comprehension

T2: Instilment of analytical and

critical skills

Page 29: What Makes an Effective Teacher

7/30/2019 What Makes an Effective Teacher

http://slidepdf.com/reader/full/what-makes-an-effective-teacher 29/29

OLS School FE OLS School FE OLS School FE

-0.576 0.217 -0.217 0.350 -0.223 -0.041

(0.372) (0.273) (0.353) (0.278) (0.178) (0.128)

. . . . . .

-1.393 -0.259 -1.120 -0.342 -0.331 0.055

(0.553) (0.320) (0.460) (0.317) (0.243) (0.167)

. . . . . .

-0.053 0.014 0.062 -0.053 -0.078 0.047

(0.295) (0.170) (0.254) (0.154) (0.153) (0.093). . . . . .

-1.082 -0.381 -0.809 -0.290 -0.188 -0.001

(0.321) (0.236) (0.284) (0.219) (0.155) (0.093)

. . . . . .

-0.761 -0.359 -0.626 -0.369 0.232 0.169

(0.365) (0.277) (0.312) (0.243) (0.133) (0.085)

0.151 0.085 -0.134 -0.025 -0.308 0.129

(0.044) (0.038) (0.053) (0.036) (1.038) (0.929)

. . . . . .0.138 0.104 -0.104 -0.088 -0.600 0.998

(0.052) (0.047) (0.060) (0.037) (1.318) (1.288)

. . . . . .

0.059 0.016 -0.019 -0.020 -1.746 -0.218

(0.036) (0.028) (0.045) (0.027) (0.842) (0.813)

. . . . . .

0.107 0.037 -0.122 -0.010 1.030 0.814

(0.031) (0.026) (0.037) (0.025) (0.915) (0.696)

. . . . . .

0.074 0.041 -0.063 0.004 -1.929 0.460

(0.030) (0.024) (0.045) (0.030) (0.804) (0.859)

T3: Instilment of capacity for 

individual study

T4: Transparency, fairness and

feedback 

T5: Individual treatment of students

T1: Instilment of knowledge and

enhancement of comprehension

T2: Instilment of analytical and

critical skills

T3: Instilment of capacity for 

individual study

T4: Transparency, fairness and

feedback 

T5: Individual treatment of students

Table A2: Balancing Tests, Eighth Grade

Father's years of 

schooling

Mother's years of 

schooling

Number of siblings

Note: Each coefficient in this table is estimated in a separate regression. Standard errors are clustered by

school. The sample includes schools with 5+ students.

Recent immigration Parents born in

Israel

Class size

T1: Instilment of knowledge and

enhancement of comprehension

T2: Instilment of analytical and

critical skills