Зборник Института за педагошка истраживања ISSN 0579-6431

Година 41 • Број 1 • Јун 2009 • 7-24 Прегледни чланак

УДК 37.013.2;37.014.2 DOI: 10.2298/ZIPI0901007T

SCHOOL EFFECTIVENESS:
LITERATURE REVIEW
Jelena Teodorović *
Institute for Educational Research, Belgrade

Abstract. In this paper, the reader embarks on the first part of the review of school effecti-
veness research. The aim of the review is to offer a clearer picture on whether, which, and
how much teacher and school variables impact student achievement, as there is currently no
wider and accepted consensus on this matter, in spite of the wealth of various school ef-
fectiveness studies. An introduction is followed by a section on fragmented research para-
digms. Four subsequent sections describe and critique findings from these paradigms, namely
from student background, input-output, effective-schools, and instructional effectiveness stu-
dies. The paper concludes with the section on synthesis of findings, which implicate student
background variables as the most important for student achievement, followed by instruction-
and teacher-related variables (in very poor developing countries, input-output factors are also
relevant for student success). Subsequent paper will showcase more recent school effecti-
veness studies that use appropriate methodology and conceptual framework for identification
of the most important school effectiveness factors.
Key words: school effectiveness, fragmented research paradigms, student background, input-
output, effective schools, instructional effectiveness.

Introduction
During the last four decades, the field of school effectiveness has been ex-
plored from a variety of research paradigms using both qualitative and quan-
titative methods. However, at the heart of all school effectiveness research is
an attempt to explain how school inputs, the context of schooling, and scho-
ol processes affect school outputs (Figure 1, Scheerens, 2000).
In general, school inputs include variables such as teacher qualifica-
tions, school infrastructure, and per-student expenditures. The context of
schooling consists of variables such as support from higher administrative
levels and school location. School processes encompass variables such as
teacher behaviors, orderly atmosphere, and the quality of school curricula.
* E-mail: [email protected]
Jelena Teodorović 8

Student achievement is usually considered to be the major school output.

Student background (also a school input) is discussed separately because it
is not under the control of the school system, and so it needs to be removed
from examination of the impact of school factors on student achievement.

Figure 1: A basic systems model of school effectiveness (Scheerens, 2000)

Context

Process

Inputs School level Outputs

Classroom level

In spite of the wealth of school effectiveness studies, there is still no com-

plete picture on whether, which, and how much teacher and school variables
that make up school inputs, processes, and context affect student achieve-
ment. This is largely due to the following four reasons:
 Three school effectiveness research paradigms have been used mostly
in isolation of each other and without a conceptually integrated model to
guide research, thus ignoring teacher and school variables identified in other
paradigms.
 Most prior research has been undertaken in either industrialized or
developing countries, possibly obscuring the importance of certain teacher
and school variables present in other countries.
 Most prior research has used statistical procedures intended for only
one unit of analysis (e.g., schools), whereas the hierarchical nature of scho-
oling requires management of multiple units of analysis (i.e., schools, clas-
srooms, and students).
 Most past studies have been in countries where students are assigned
to different teachers each year, obscuring the possible cumulative effects of
teachers.
This paper focuses on the first two problems; incidentally, it also repre-
sents a review of the older school effectiveness studies. The second part of
9 School effectiveness: literature review

the literature review (to be published later) will address two other problems,
as well as give insight into more recent and advanced school effectiveness
studies.

Fragmented Research Paradigms

In 1963, John B. Carroll postulated that student learning depends on time
needed to learn (determined by a student’s aptitude, ability to understand in-
struction, and quality of instruction) and time spent in learning (determined
by the time allowed for learning and the student’s perseverance). Later, Blo-
om (1976) elaborated on Carroll’s model by positing that a student’s cogni-
tive and affective entry behaviors (the prerequisite learning and motivation)
and quality of instruction determine learning outcomes. In both models,
variables can be altered to both increase learning and reduce variation
between students.
Most of research that ensued has been empirical and has focused on
various parts of the two models, aiming to measure the above-mentioned,
generalized variables in different ways. Of the thousands of studies that
have been conducted, most have been undertaken within one of three dif-
ferent paradigms: (a) input-output studies, (b) effective-schools studies, and
(c) studies on instructional effectiveness (Scheerens, 2000). Input-output
studies have focused on finding relationships between various school inputs
(excluding student background factors) Effective-schools research has em-
phasized the importance of school processes, particularly school organiza-
tion and administrative practices. The instructional-effectiveness paradigm
has centered on the links between teaching processes and school outputs.
While each research tradition has made strides in discovering factors that
consistently influence student achievement, little research has drawn simul-
taneously on all three paradigms.

Student Background Research

The now-famous 1966 Coleman report found that in the United States stu-
dent background (mostly socioeconomic) factors showed much stronger as-
sociation with student achievement than any school-related factors, such as
per-student expenditures and teacher qualifications (Coleman et al., 1966).
These findings were confirmed in numerous subsequent studies in industri-
alized countries.
Grissmer et al. (1994) used multiple linear regressions to identify the
net impact of each of several student background factors on student achie-
Jelena Teodorović 10

vement in mathematics and verbal/reading skills. For example, the dif-

ference between having a college-educated mother and a mother who did not
finish high school resulted, on average, in about a half standard deviation
difference in achievement. Being black resulted in a score that was, on ave-
rage, half a standard deviation below the score for a white student. Dif-
ferences in family income ($40,000 vs. $15,000), and family size (having
four vs. having one sibling) showed an impact of around 0.10–0.20 standard
deviations.
In the re-analysis of IEA data on student achievement, Heyneman and
Loxley (1982) also used multiple linear regression to show that in indus-
trialized countries various student background variables (the father’s and
mother’s education, the father’s occupation, the number of books in the
home, the use of the dictionary in home, the sex of the student, and the age
of the student) explained around 20% of total variance in science achie-
vement (this constituted about 50% of explainable variance1).
PISA 2000 showed that several student background factors explained
significant variance in the achievement of 15-year old students: parental
occupational status, cultural possessions at home, parental involvement,
home educational resources, participation in cultural activities, and family
wealth (OECD, 2001). The same was true of some variables that capture
learning strategies and attitudes: engagement in reading, interest in reading,
control strategies, and time spent on homework (OECD, 2001). In single
linear regression analyses, mother’s education exhibited a large significant
effect on student achievement (the difference between students whose
mothers did not finish high school and those whose mothers finished high
school was 45 points, with 500 being the average). Students from single-
parent families performed 12 points below those from two-parent families.
Students who were not born in the country of testing scored 71 points lower
(a full proficiency level) than native students.
In multiple linear regression analyses where the above-mentioned va-
riables (excluding learning strategies and attitudes variables) were simul-
taneously examined, their effect changed somewhat. Increase in one stan-
dard deviation in parental occupational status increased student achievement
by 28 points on average. Students born in the country of testing performed
better than non-native students by 26 points on average. Each additional
year of parental education added 5 points to the student’s score, while one

1 Another 20% of variance was explained by school track (general, vocational, or academic),
school program (focusing on social science, natural science, etc.), and school variables (e.g., using
textbook in science class). The remaining 60% could not be explained by any of the above va-
riables.
11 School effectiveness: literature review

standard deviation increase in indices of home educational resources and

cultural possessions raised student achievement by 12 and 13 points, res-
pectively. Living in a two-parent family was associated with 11-point higher
achievement (OECD, 2001).
PISA 2000 highlighted the need to analyze the factors simultaneously
with each other rather than independently, since only then their net effect on
achievement can be accurately determined. Also, it is important to examine
a wide range of literature-implicated variables, as the effect of a certain
variable can considerably vary depending on the presence of other variables
that potentially impact achievement.
Student background factors were shown to play an important role for
student achievement in developing countries as well, but the magnitude of
their effect is more debated than in industrialized countries. For example,
using vote-counting method2, Velez et al. (1993) found that in many studies
in Latin America student achievement showed positive association with
parental education or occupation, family income, access to books, attitudes
toward study, previous cognitive achievement, access to TV, and IQ / abi-
lity. Family size and student’s age were shown to be negative predictors.
Since the vote counting cannot provide the effect sizes (and many primary
studies do not supply effect sizes, Fuller & Clarke, 1994), the magnitude of
these findings cannot be estimated.
Heyneman and Loxley (1982) found that in four developing countries
student background variables explained only a small percentage of total
variance in science achievement: 2.7% in India, 6% in Thailand, 8% in Iran,
and 8% in Chile. When taken as a percentage of variance that could be
explained by all examined variables, student background accounted for
8.7% of explainable variance in India, 18.8% in Thailand, 47% in Iran and
23.5% in Chile. These numbers point to a modest influence of student
background factors on student achievement.
Overall, student background factors are very important in both indus-
trialized and developing countries. This conclusion is well established in
industrialized countries, and more debated in developing countries.

Input-Output Research
In many input-output studies, large data sets were analyzed and achievement
test scores mostly taken as a measure of school output (Hanushek, 1989).
The most frequently examined school inputs were measures of teacher cha-
2 This method simply counts the number of studies that found statistically significant posi-
tive effect of a variable, regardless of the effect size, sample size, or other study characteristics.
Jelena Teodorović 12

racteristics (experience, education, etc.), school infrastructure and services

(facilities, class sizes, etc.), and per-student expenditures (Hanushek, 1989).
While most researchers recognize that in industrialized countries, student
background factors relate with school outputs more strongly than all scho-
oling factors taken together, the magnitude of the effects of school inputs is
nonetheless debated.
For example, Darling-Hammond (2000) analyzed state-level data con-
trolling for student poverty and limited-English proficiency status and found
that the percentage of teachers with full certification and a major in their
field explained approximately half of total variability in student math and
reading achievement. Payne and Biddle (1999) determined that, after con-
trolling for district-level child poverty, the average level of curricular in-
struction and the percentage of non-white persons, district-level annual per-
student funding had a significant influence on student achievement. Fergu-
son (1991) concluded that, after controlling for district-level socio-economic
status, teacher test scores, teacher experience, and master’s degree each ex-
plained 20–25%, 10% and 5% of inter-district variation in student achieve-
ment, respectively. Goldhaber and Brewer (1997) found that, after control-
ling for students’ parental education, family income, sex, race, and family
structure, teachers with certifications and bachelor and master degrees in
mathematics produced significantly higher student test scores than teachers
whose certifications and bachelor and master degrees were not in
mathematics.
On the other hand, Rivkin et al. (2000) found that teacher experience
beyond first several years and master’s degree were not significant predi-
ctors of student achievement. In single linear regressions, PISA 2000 found
that, on average, across OECD countries, input-output variables (teacher
shortages, quality of the schools’ educational resources, and quality of scho-
ols physical infrastructure), played very little role in explaining student achi-
evement across schools (OECD, 2001).
Contradictory results are also present in the meta-analyses of input-
output research. Hanushek (1989, 1997) used a vote-counting method and,
across many studies, found no consistent relationship between school inputs
(per-pupil expenditure, teacher salary, teacher education, teacher experien-
ce, and teacher / pupil ratio), and school outputs. On the other hand, Hedges
et al. (1994) disputed Hanushek’s findings on the grounds that a vote-coun-
ting method has a very low power to detect significant effects. The authors
then used a combined significance test to show that there was a significant
relationship between each of the above-mentioned variables and school
outputs.
13 School effectiveness: literature review

The problem with both primary studies and, by extension, meta-ana-

lyses that contain them, is that many are methodologically of poor quality
(Hanushek, 1997). First, if output measures are not properly controlled for
student socioeconomic status or prior achievement, the number of positive
findings will be inflated. Second, if studies use aggregated data instead of
individual student data, the school input effects will usually be misestimated
(this effect is known as aggregation bias). Therefore, the results of the above-
-mentioned primary studies (except for Goldhaber and Brewer’s) are not en-
tirely credible because they used few control variables and/or aggregated da-
ta. In order to improve quality of studies, important student background vari-
ables need to be controlled for and proper units of analysis need to be used.
A third methodological shortcoming of past research, noted by many
authors, is that studies vary dramatically in the choice of variables included
in regression analyses. As a result, the effect of a particular variable on stu-
dent achievement can vary from study to study depending on the presence of
other variables in the analysis, particularly if the included and excluded vari-
ables are collinear3. This can result in contradictory findings and contribute
to confusion about which variables are important for student achievement.
Two other related and more substantive issues with input-output stu-
dies were also mentioned by Hanushek (1989) and Scheerens (2000). First,
existing measures of school inputs used in the majority of input-output stu-
dies are likely inadequate to fully capture the true effects of schools (Hanu-
shek, 1989). For example, a teacher’s degree probably does not fully mea-
sure all teacher’s effects on student achievement, as teachers with the same
degree probably teach differently, assign different homework and differently
motivate the students. Second, most input-output studies assume a simple
and direct relationship between inputs and outputs. In reality however, inte-
ractions between various inputs and their interactions with school processes
and contexts, most likely act together to influence the outputs (Scheerens,
2000). In other words, particular teaching styles may be more effective for
high-SES students, or less effective if undertaken in schools with poorly
provided teaching tools. To the extent that these interactions among vari-
ables are disregarded in the input-output research, the effect of school inputs
will be miscalculated.

3 For example, in their review of the impact of teacher qualifications on student achieve-
ment, Wayne and Youngs (2003) reported on five longitudinal studies that found a positive relati-
onship between teacher test scores and student achievement and two longitudinal studies that found
a negative relationship. The seemingly inconsistent findings became less inconsistent once it was
revealed that the two studies that found a negative effect of teacher test scores on student achie-
vement controlled for the rating of the teacher’s college, which presumably captured a similar
aspect of teacher quality as teacher test scores.
Jelena Teodorović 14

Findings of the input-output studies undertaken in developing countries

are less controversial. The reviews of the studies that controlled for student
background (Farrell & Oliveira, 1993; Fuller & Clarke, 1994; Hanushek,
1995; Velez et al., 1993) used a vote-counting method to show that in many
studies, the existing measures of school inputs, such as basic infrastructure,
textbook availability, teacher education, per-student expenditure, and school
facilities, consistently exhibited a significant impact on student achievement
in developing countries. Since all the meta-analyses were of studies that
controlled for student background, their findings are less likely to be infla-
ted. However, studies undertaken in developing countries likely suffer from
similar above-mentioned methodological shortcomings, namely omission of
potentially important variables from regression equations. As a consequence
of such differences, the coefficient estimates are likely to vary from study to
study, and findings are likely to be mixed.
Harbison and Hanushek (1992) provided some estimation of the effects
of school inputs in poor developing countries: supplying all school facilities
to schools that had none increased achievement by 0.09–0.13 standard devi-
ations and supplying a package of writing materials to students increased
achievement by 0.30–0.50 standard deviations.
Heyneman and Loxley (1982) also provided some estimates of school
effects: in developing countries schooling variables (all examined together)
explained two to three times more achievement variance than in industriali-
zed countries (after controlling for same student background factors). The
proportion of total variance that was explained by schooling variables was
28% in India, 26% in Thailand, 9% in Iran, and 20% in Chile, while compa-
rable percentages in industrialized countries ranged from 6% to 16%. To-
gether with the school track variables (general, vocational, or academic) and
school program variables (e.g., focusing on social science), the schooling
variables explained 91.3% of explainable variance in India, 81.2% in Thai-
land, 53% in Iran, and 76.5% in Chile, while comparable percentages in in-
dustrialized countries were around 50% or less.
An explanation for the differences between the findings on student
background and school inputs in industrialized and developing countries is
given by Farrell and Oliveira (1993) and Scheerens (2000). Farrell and Oli-
veira suspect that industrialized countries are close to the limits of the per-
fectibility of the technology of schooling, so that even modest additional
gains in achievement require costly and difficult educational efforts. Simila-
rly, Scheerens attributes a larger impact of school inputs in developing coun-
tries, compared to industrialized countries, to the greater variance in both
school inputs and outputs. The opposite is then true for industrialized coun-
15 School effectiveness: literature review

tries: The smaller influence of school inputs and larger influence of student
background in industrialized countries are probably explained by the smaller
variability in school inputs and/or, less likely, greater variability in students’
home educational environment and support.
Overall, input-output variables seem to be less relevant for student
achievement in industrialized countries and more relevant in developing
countries. Notwithstanding these general trends, input-output research in
both sets of countries is likely to benefit from improved control of student
background factors, an appropriate level of data aggregation, and, most
crucially, the inclusion of additional variables that impact student achi-
evement.

Effective-schools Research
In the third body of research examined in this literature review, researchers
mostly used surveys, case studies and field studies to research organization,
form, and content of high-end outlier schools and identify the factors that
make them unusually effective (Levin & Lockheed, 1991, Scheerens, 2000).
A summary of the five recent reviews of effective-schools studies points out
that, for industrialized countries, consensus is greatest in the research com-
munity on the importance of the following factors: (a) achievement orienta-
tion (related to high expectations), (b) co-operation, (c) strong educational
leadership, and (d) frequent monitoring of progress (Scheerens, 2000).
Orderly climate also appears to be important. However, it seems that studies
employing qualitative data and mixed studies find more support for the
contribution of effective-schools factors than studies that rely solely on
quantitative data. Both sets of studies, however, suffer from some methodo-
logical shortcomings, which results in a lack of consensus on the relevance
of effective-schools variables for student achievement.
For example, in the seminal study “Fifteen Thousand Hours”, Rutter et
al. (1979) concluded that, after taking into account individual student cha-
racteristics, such as parental occupation and verbal reasoning, school cha-
racteristics most associated with student attainment on public academic
exams were good classroom management, high expectations of academic
success, modeling of good behavior by teachers, positive feedback, well-
conducted lessons, staff cohesion in academic and disciplinary matters, and
pleasant working environment (last two are effective-schools factors). The
authors also found that the peer effect (larger presence of more able child-
ren) had an impact on achievement over and beyond what would be expe-
cted from individual characteristics. However, since the authors combined
Jelena Teodorović 16

the scores for all variables into one composite score, they did not assess the
relative importance of individual schooling factors.
In the International School Effectiveness Research Project (ISERP),
Reynolds et al. (2002) examined high- and low-achieving school outliers
enrolling either low-SES or middle-SES children in nine industrialized
countries. The authors used both quantitative and qualitative data to find
that greatest differences between less and more effective schools were in
child experiences in school during the day, instructional style, principal
leadership, expectations for students, school goals, and inter-staff relations
(the last four are effective-schools factors). The study also found that scho-
ol-level variables appeared to be more salient for low-SES schools, while
instructional variables were more important for middle-SES schools, which
presumably already had in place those baseline school-level factors (e.g.,
orderly climate).
Sammons et al. (1998) conducted case studies of six outlier schools
that were categorized as effective, ineffective, and mixed in a study of 94
London secondary schools4. The features of effective schools were high
expectations and emphasis on academic achievement, encouragement of pa-
rental involvement, a principal’s leadership, a strong management team,
quality teaching in the school, and emphasis on homework.
Encouragement of parental involvement appears to play an especially
prominent role in student achievement, probably because it taps into paren-
tal and home educational resources that have proven to have a considerable
influence on student achievement. The majority of 29 controlled studies of
cooperative efforts by parents and educators (Walberg & Paik, 2000) and 51
studies that examined family and community effects on student achievement
(Henderson & Mapp, 2002) showed that parental involvement in school
functioning and programs that actively engaged families in supporting stu-
dent learning at home were associated with better school outputs.
However, some quantitative studies refute this. For example, in their
review of five studies, Ellett et al. (1997) found little association between
school environment and organizational variables (e.g., the nature of decision
making), and student achievement. Similarly, van der Werf (1997) studied
560 low-, medium-, and high-achieving schools using one-way ANOVA to
find that the only one school characteristic – school leadership – was consis-
tently (negatively) associated with student achievement. PISA 2000 also
found that teachers’ morale and commitment, school autonomy, school se-
4 Schools were grouped into three categories by the HLM method that adjusted school mean
scores for student background characteristics and prior achievement, and simultaneously used
students and schools as units of analysis.
17 School effectiveness: literature review

lectivity, and teacher autonomy each explained around 0.5–2% of total vari-
ance in single linear regressions. Only principals’ perceptions of student-
related factors affecting school climate (e.g., absenteeism) explained 5.8%
of the variance (OECD, 2001).
However, the above-mentioned research is problematic. For example, it
is possible that even the educational leadership variable in the van der Werf
study would not have been significantly associated with achievement had
the other variables been simultaneously tested. On the other hand, the
reverse is also plausible: Additional effective-schools variables might have
appeared significant if they acted together with other effective-schools or
instructional-effectiveness variables5. The same holds true for PISA 2000,
since the importance of each effective-schools variable was assessed indivi-
dually. The possibility that such suppression effect occurs can be minimized
by simultaneously examining all relevant, theoretically-implicated variables.
Several quantitative meta-analyses of available studies concluded that
the contribution of effective-schools variables to student achievement was
small to modest (the correlation coefficients ranged from -0.02 to 0.20
without controlling for student background (Scheerens, 2000)). If the stu-
dent background had been included, the correlation coefficients would have
likely been even smaller, since student background variables would have
captured a considerable portion of the relationship between student achie-
vement and effective-schools factors. Wang et al. (1994) also found that va-
riables such as district policies or school organization were associated with
student achievement only moderately in comparison with instructional-ef-
fectiveness variables (see next section). Also, international comparative stu-
dies pointed out that the consistency of effective-schools factors was very
low across various countries (Scheerens, 2000).
The definition and operationalization of the effective-schools factor
categories may vary widely between studies. A restricted range of variation
also may exist in school organization, form, and content, which precludes
quantitative studies from finding significant effects of effective-schools
variables, but does not impact qualitative researchers who study high-end
outliers. Similarly, the failure to consider important effective-schools varia-
bles, variables from two other research paradigms and student background
factors may inflate the findings of studies that focus only on certain ef-

5 A study by Gutman and Midgley (as cited in Henderson & Mapp, 2002) shows this is pos-
sible: High grades were obtained by students who experienced a combination of either high parent
involvement and high teacher support or high parent involvement and student sense of belonging at
school, but not by students who experienced only one of the three examined variables. This is
known as suppression effect.
Jelena Teodorović 18

fective-schools factors. The inconsistency of the findings between the coun-

tries could be due to the different cultural contexts, which foster or oppose a
particular effective-schools factor, such as parental involvement (Fuller &
Clarke, 1994). Last, but not least, effective-schools factors may be too remo-
ved from students to considerably impact student achievement, unlike more
proximal factors, such as teachers (Wang et al., 1994).
Effective-schools factors have been rarely examined in developing
countries. Fuller and Clarke (1994) found that, in over 100 studies, effec-
tive-schools factors were examined only three to four times. In Latin Ame-
rica and the Caribbean, the effective-schools factors were explored rarely
(Velez et al., 1993). In over 50 studies reviewed by Farrell and Oliveira
(1993), four out of seven analyses found the quality of the principal to be
significantly correlated with student achievement.
Studies that did explore effective-schools factors in developing nations
suffer from several problems mentioned before. An example comes from a
study of Gonakelle school in Sri Lanka (Little & Sivasithambaram, 1991).
The authors attributed the school’s high mathematics achievement over
others in the district to interplay between enthusiastic, local teachers,
teacher stability, and the strong leadership of an active principal. However,
considering that the initial average achievement of the Gonakelle students
was much higher than the district average, and that no other control varia-
bles were examined, the above-mentioned effective-schools factors could
not be solely credited for Gonakelle student achievement.
In Colombia, a massive project called Escuela Nueva was implemented
in rural schools beginning in 1975. The hallmarks of Escuela Nueva were
strong teacher-parent-community relationships, active learning curriculum,
critical thinking, cooperation, self-instructional materials, individualized
learning, teacher training program, monthly teacher support groups, joint
teacher and student management of school, and frequent visits from supervi-
sors (Rugh & Bossert, 1998). Several evaluations found varied success of
Escuela Nueva students in Spanish and mathematics (Rugh & Bossert,
1998). However, the exact contribution of each of the factors that comprised
Escuela Nueva schools cannot be properly assessed if student background
factors and factors that potentially impact student success (e.g., dropout
rates) are not examined simultaneously.
In summary, the relevance of effective-schools factors in industrialized
countries is debated. They were not adequately explored in developing
countries. Future effective-schools research needs to be more specific in the
definition and operationalization of effective-schools factors, examine a
range of variation in school organization, form, and content, examine cultu-
19 School effectiveness: literature review

ral factors or factors at higher administrative levels, utilize better analytical

methods, and include additional variables that impact student achievement.

Instructional-effectiveness Research
The final body of research examined in this paper – the instructional-ef-
fectiveness paradigm – has mostly focused on teacher behaviors and prac-
tices in the classroom. Instructional-effectiveness studies, mostly undertaken
in industrialized countries, were often based on experimental design (Sche-
erens, 2000; Walberg & Paik, 2000).
Several important reviews of instructional-effectiveness research
(Brophy & Good, 1986; Wang et al., 1993; Creemers, 1994; Scheerens,
2000; Walberg & Paik, 2000) singled out the following teaching behaviors
as the ones most consistently associated with student achievement: empha-
sizing academic instruction; maximizing efficient time on task; actively
teaching (vs. allowing individual, unsupervised study by students); adjusting
the difficulty and cognitive level of tasks and questions to the students;
structuring, outlining, and reviewing lessons; questioning, testing and pro-
viding homework; prompting and providing feedback; ensuring clear cor-
respondence between covered material and tests (so-called opportunity to
learn); monitoring for completion and accuracy in supervised independent
seatwork and homework; teaching of learning strategies; providing cor-
rective instruction; preparing in advance; being flexible, clear and enthusias-
tic; having high expectations; maintaining an orderly atmosphere; having
quality academic and social interactions with students6.
The results of three meta-analyses estimated that instructional-ef-
fectiveness variables had modest to large effects on student achievement
(Scheerens, 2000). For example, transformed correlation coefficients for stu-
dent achievement and opportunity to learn, co-operative learning, feedback,
and reinforcement were 0.15, 0.27, 0.48, and 0.58, respectively. However,
because student background variables were not controlled for, it is likely
that these correlation coefficients were somewhat inflated.
Even though it appears that instructional-effectiveness factors play an
important role in industrialized countries, several problems with instructi-
onal-effectiveness research still exist. First, some of the experimental re-
search was done in a short period of time, and it is not clear how powerful

6 It should be noted that some of the findings were dependent on the context, such as grade
level or student SES / ability. For example, teacher’s focus on basic skills and encouragement of
overt student participation was especially effective in the early grades, while more drill, review,
and practice appeared effective for low-SES students (Brophy & Good, 1986).
Jelena Teodorović 20

the findings are in long-term educational situations. Second, experimental

research is also susceptible to the Hawthorne effect, that is, any innovation
can artificially increase the effects of the experiment simply due to its no-
velty (Weiss, 1998). Third, non-experimental studies on instructional ef-
fectiveness might artificially show large results if they omitted other theore-
tically-implicated variables from consideration. Fourth, instructional-effecti-
veness variables likely exert differential effects in different contexts (dif-
ferent student composition, grade levels, subject matter, etc.).
The available evidence suggests that some instructional-effectiveness
factors, although rarely explored, may play important roles in developing
countries. For example, significant positive associations were found in seve-
ral reviews between student achievement and instructional time and frequen-
cy of homework (Fuller & Clarke, 1994), time on task and homework prac-
tices (Velez et al., 1993), the length of the instructional program, homework
frequency, teacher’s expectations of student performance, and teacher’s time
spent on class preparation (Farrell & Oliveira, 1993). However, these
findings need to be corroborated as they were based on a small number of
analyses. Due to the vote-counting method applied in the reviews, the effect
sizes were not available.
While consensus seems to be forming in industrialized countries that
instructional-effectiveness factors are important for student achievement,
research on instructional effectiveness in both the industrialized and deve-
loping world would benefit from studies that simultaneously include other
relevant student background and school effectiveness factors, use precisely
specified and operationalized factors, and show substantial range of varia-
tion in instructional practices.

Synthesis of Findings from Research Paradigms

In summary, three research paradigms have been mostly used in isolation
from each other. Table 1 summarizes the most common findings of each of
the three major research paradigms, as well as the common findings about
student background factors that most impact achievement.
In both industrialized and developing countries student background
factors play a large role in student achievement (more so in industrialized
countries). Input-output variables seem to be more important in developing
countries. Instructional effectiveness factors seem to have an important role
in both sets of countries, while the impact of effective-schools factors seems
to be most debated. The implications of these findings are the following: (1)
While educational system cannot be held responsible for the impact of
21 School effectiveness: literature review

student background factors on achievement, it can nonetheless aim to mo-

dify its extent (e.g., it can encourage and facilitate parental involvement in
school work); (2) Education policymakers and practitioners should strive to
improve instructional effectiveness factors, as teaching and learning process
appears to have the largest impact on achievement of all school effecti-
veness factors that are in realm of the educational system.

Table 1: School Effectiveness Factors

Student Background Input-Output Effective Schools Instructional
Effectiveness
 parental educational  teacher  strong educational  clarity of presentation
attainment qualifications leadership  flexibility
 parental (experience,  emphasis on  enthusiasm
occupational status salary, education, acquiring basic skills  clear correspondence
 family wealth sex)  orderly and secure between covered
 cultural possessions  school environment material and tests
at home (books, art) infrastructure and  high expectations of  teaching skills until
 communication on services (facilities, student achievement mastery is achieved
social issues and availability of  frequent assessment  time on task
aspects of culture textbooks, class of student progress  graded homework
 family structure size, administrative  co-operation  direct, structured
 student age expenses)  strong parental teaching
 student gender  average per-student outreach  advance preparation
 parental involvement expenditures of teachers
in school work  teaching of learning
strategies
 classroom
management
 teacher and student
academic and social
interactions

Subsequent paper will solidify these conclusions by showcasing more recent

school effectiveness studies that use appropriate methodology and concep-
tual framework for identification of the most important school effectiveness
factors.

Note. This paper is the result of the Projects which are financially supported
by Serbian Ministry of Science and Technological Development (2006-
2010): “Education for knowledge-based society (No 149001).

References
Bloom, B. (1976): Human characteristics and school learning. New York: McGraw Hill.
Brophy, J. & T. Good (1986): Teacher behavior and student achievement; in M. Wittrock
(ed.): Handbook of research on teaching (340-370). New York: Macmillan.
Jelena Teodorović 22

Carroll, J.B. (1963): A model of school learning, Teachers College Record, 64, 723-733.
Coleman, J.S., E.Q. Campbell, C.J. Hobson, J. McPartland, A.M. Mood & F.D. Weinfeld
(1966): Equality of educational opportunity. Washington: U.S. Government Printing
Office.
Creemers, B.P.M. (1994): The effective classroom. London: Cassell.
Darling-Hammond, L. (2000): Teacher quality and student achievement: a review of state po-
licy evidence, Education Policy Analysis Archives, 8(1), 1-52.
Ellett, C.D., C.S. Logan, J.G. Claudet, K.S. Loups, B.L. Johnson & S.W. Chauvin (1997): Scho-
ol learning environment, organizational structures and effectiveness - a synthesis of re-
search in 442 schools, International Journal of Educational Research, 27(41), 273-282.
Farrell, J.P. & J.B. Oliveira (1993): Teachers in developing countries: improving effective-
ness and managing costs. Washington: The World Bank.
Ferguson, R.F. (1991): Paying for public education: new evidence on how and why money
matters, Harvard Journal on Legislation, 28(2), 465-498.
Fuller, B. & P. Clarke (1994): Raising school effects while ignoring culture? Local conditions
and the influence of classroom tools, rules, and pedagogy, Review of Educational
Research, 64(1), 119-157.
Goldhaber, D.D. & D.J. Brewer (1997): Why don’t schools and teachers seem to matter?
Assessing the impact of unobservable on educational productivity, The Journal of
Human Resources, 32(3), 505-520.
Grissmer, D.W., S.N. Kirby, M. Berends & S. Williamson (1994): Student achievement and
the changing American family. Santa Monica: Rand Corporation.
Hanushek, E.A. (1989): The impact of differential expenditures on school performance, Edu-
cational Researcher, 18(4), 45-51, 62.
Hanushek, E.A. (1994): Money might matter somewhere: a response to Hedges, Laine, and
Greenwald, Educational Researcher, 23(4), 5-8.
Hanushek, E.A. (1997): Assessing the effects of school resources on student performance: an
update, Educational Evaluation and Policy Analysis, 19, 141-164.
Harbison, R.W. & E.A. Hanushek, (1992): Educational performance of the poor: lessons
from rural Northeast Brazil. New York: Oxford University Press.
Hedges, L.V., R. Laine & R. Greenwald (1994): Does money matter? A meta-analysis of
studies of the effects of differential school inputs on student outcomes, Education
Researcher, 23, 5-14.
Henderson, A.T. & K.L. Mapp (2002): A new wave of evidence - the impact of school, family
and community connections on student achievement. Austin: Southwest Educational
Development Laboratory.
Heyneman, S.P. & W.A. Loxley (1982): Influences on academic achievement across high and
low income countries – a re-analysis of IEA data, Sociology of Education, 55(1), 13-21.
Levin, H.M. & M.E. Lockheed (1991): Effective schools in developing countries. Washing-
ton: The World Bank.
Little, A. & R. Sivasithambaram (1991): ;in H.M. Levin & M.E. Lockheed (eds.): Effective
schools in developing countries (96-121). Washington: The World Bank.
OECD (2001): Knowledge and skills for life: First results from PISA 2000. Paris: OECD.
Payne, K.J. & B.J. Biddle (1999): Poor school funding, child poverty, and mathematics achi-
evement, Educational Researcher, 28 (6), 4-13.
Reynolds, D., B. Creemers, S. Stringfield, C. Teddlie & E. Schaffer (2002): World class scho-
ols: international perspectives on school effectiveness. London: Routledge Falmer.
Rivkin, S.G., E.A. Hanushek & J.F. Kain (2000): Teachers, schools and academic achieve-
ment. Cambridge, MA: National Bureau of Economic Research.
Rugh, A. & H. Bossert (1998): Involving communities: participation in the delivery of edu-
cation programs. Washington: Creative Associates International.
Rutter, M., B.Maughan, P. Mortimore, J. Ouston & А. Smith (1979): Fifteen thousand hours:
secondary schools and their effects on children. London: Open Books.
23 School effectiveness: literature review

Sammons, P., S. Thomas, P. Mortimore, A. Walker, R. Cairns & J. Bausor (1998): Under-
standing differences in academic effectiveness: practitioners’ views, School Effective-
ness and School Improvement, 9(3), 286–309.
Scheerens, J. (2000): Improving school effectiveness (Fundamentals of Educational Planning
No. 68). Paris: UNESCO/International Institute for Educational Planning.
Van der Werf, G. (1997): Differences in school and instruction characteristics between high-,
average-, and low-effective schools, School Effectiveness & School Improvement, 8(4),
430-448.
Velez, E., E. Schiefelbein & J. Valenzuela (1993): Factors affecting achievement in primary
education. Washington: The World Bank.
Walberg, H.J. & S.J. Paik (2000): Effective educational practices. Geneva: International
Bureau of Education.
Wang, M.C., G.D. Haertel & H.J. Walberg (1993): Toward a knowledge base for school lear-
ning, Review of Educational Research, 63(3), 249-294.
Wayne, A.J. & P. Youngs (2003): Teacher characteristics and student achievement gains: a
review, Review of Educational Research, 73(1), 89-122.
Weiss, C. (1998): Evaluation (2nd ed.). Upper Saddle River, NJ: Prentice Hall.

Примљено 25.03.2009; прихваћено за штампу 27.04.2009.

Јелена Теодоровић
ШКОЛСКА ЕФИКАСНОСТ: ПРЕГЛЕД ЛИТЕРАТУРЕ
Апстракт
У овом раду читаоцу се нуди први део прегледа истраживања о школској ефи-
касности. Циљ овог прегледа јесте да понуди јаснију слику у погледу тога да
ли, које и у којој мери, варијабле наставника и школе утичу на постигнуће уче-
ника, с обзиром да тренутно не постоји шири и прихваћени консензус у вези са
тим, упркос обиља разних истраживања школске ефикасности. Након увода,
следи део о фрагментираним истраживачким парадигмама. Четири дела који
следе описују и критикују налазе из ових парадигми, наиме, порекло ученика,
input-output, ефикасне школе и истраживања наставне ефикасности. Рад се за-
вршава делом о синтези налаза, који показују да су варијабле које се односе на
порекло ученика најважније за њихово постигнуће, а следе варијабле које се
односе на наставу и наставнике (у веома сиромашним земљама у развоју, input-
-output фактори су такође релевантни за успех ученика). Други део приказа ће
представити нека новија истраживања о школској ефикасности која користе
одговарајућу методологију и појмовни оквир за идентификовање најважнијих
фактора школске ефикасности.
Кључне речи: школска ефикасност, фрагментиране истраживачке парадигме,
порекло ученика, инпут-оутпут, ефикасне школе, наставна ефикасност.
Jelena Teodorović 24

Елена Теодорович
ЭФФЕКТИВНОСТЬ В ШКОЛЕ: ОБЗОР ЛИТЕРАТУРЫ
Резюме
В предлагаемой работе излагается первая часть обзора исследований эффекти-
вности в школе. Цель данного обзора – дать како можно более ясное представ-
ление о том, влияют ли вариаблы преподавателей и школы на постижение уча-
щихся, и если да, то какие и в какой степени. Дело в том, что сегодня нет об-
щепринятого или хотя бы широко распространенного мнения по этому вопро-
су, несмотря на обилие разных исследований эффективности в школе. После
введения автор переходит к рассмотрению фрагментированных исследователь-
ских парадигм. Последующих четыре части работы посвящены описанию и
критике исследовательских результатов, полученных в рамках упомянутых па-
радигм, в частности таких, как происхождение учащихся, input-output, эффек-
тивные школы и исследования эффективности в обучении. Заключительная
часть работы предлагает синтез результатов, показывающих, что вариаблы,
связанные с происхождением учащихся, имеют решающее значение для их
постижений, тогда как на втором месте находятся вариаблы, связанные с обу-
чением и с преподавателями (в предельно бедных развивающихся странах фак-
торы input-output также значимы для успеваемости учащихся). Вторая часть
обзора представит некоторые современные исследования об эффективности в
школе, использующие адекватную методологию и понятийную рамку для выя-
вления важнейших факторов эффективности в школе.
Ключевые слова: эффективность в школе, фрагментированные исследователь-
ские парадигмы, происхождение учащихся, input-output, эффективные школы,
эффективность обучения.