Achievement Inequality and the Institutional Structure of ...hermanvandewerfhorst.socsci.uva.nl/ARS2010.pdf · Achievement Inequality and the Institutional Structure of Educational

SO36CH20-VanDeWerfhorst ARI 3 June 2010 1:12

Achievement Inequalityand the Institutional Structureof Educational Systems:A Comparative PerspectiveHerman G. Van de Werfhorst and Jonathan J.B. MijsDepartment of Sociology and Anthropology, University of Amsterdam,Oudezijds Achterburgwal 185, 1012 DK Amsterdam, The Netherlands;email: [email protected]

Annu. Rev. Sociol. 2010. 36:407–28

First published online as a Review in Advance onApril 20, 2010

The Annual Review of Sociology is online atsoc.annualreviews.org

This article’s doi:10.1146/annurev.soc.012809.102538

Copyright c© 2010 by Annual Reviews.All rights reserved

0360-0572/10/0811-0407$20.00

Key Words

tracking, stratification, standardization, PISA, TIMSS

Abstract

We review the comparative literature on the impact of national-leveleducational institutions on inequality in student achievement. We fo-cus on two types of institutions that characterize the educational systemof a country: the system of school-type differentiation (between-schooltracking) and the level of standardization (e.g., with regard to centralexaminations and school autonomy). Two types of inequality are ex-amined: inequality in terms of dispersion of student test scores andinequality of opportunity by social background and race/ethnicity. Weconclude from this literature, which mostly uses PISA, TIMSS, and/orPIRLS data, that inequalities are magnified by national-level trackinginstitutions and that standardization decreases inequality. Methodolog-ical issues are discussed, and possible avenues for further research aresuggested.

407

Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


INTRODUCTION

Thanks to the vast progress in data collec-tion and availability, we are witnessing therise of an elaborate, multidisciplinary literatureon cross-national variation in student perfor-mance in primary and secondary schools. In-ternational data projects, such as the Programfor International Student Assessment (PISA),the Progress in International Reading LiteracyStudy (PIRLS), and the Trends in InternationalMathematics and Science Study (TIMSS), havepaved the way for a rich body of articles and offi-cial reports on international variation in averagestudent performance, in the dispersion in per-formance, and in the influence of social originand race/ethnicity on school performance.

An important research question that is ad-dressed in this literature concerns the extentto which national educational institutions affectinequality in learning among students. Coun-tries differ strongly in the organization of theireducational systems, and it is important to knowwhether particular educational institutions areconducive to enlarging inequalities amongstudents.

We focus on two types of inequality: in-equality in learning as measured by the dis-persion in test scores (which we label inequal-ity as dispersion) and inequality of educationalopportunity in terms of the influence of so-cial class and race/ethnicity on students’ testscores. These two aspects of inequality areconceptually distinct—an educational systemcould be relatively equal in terms of disper-sion but unequal in terms of opportunities—yet, as we argue, theoretically and empiricallylinked.

With respect to educational institutions, wefocus on the dimensions of differentiation inschool types and tracks in secondary schoolingsystems and nationwide standardization. Wesee these characteristics of educational systemsas institutions, as they reflect “laws, informalrules and conventions which give a durablestructure to social interactions among membersof a population” (Bowles 2004) in the form of(sub)national regulations on the way education

is organized. Differentiation and standardiza-tion are both related to selection and allocationprocesses within and between schools.

We study differentiation primarily in termsof the external differentiation in separateschool types and often schools. In the classi-fication of types of curriculum differentiationof LeTendre et al. (2003), this review is mainlyconcerned with school-type differentiation anddifferentiation in terms of the geographicallocation of schools. Studies focusing on within-school-type ability grouping (i.e., internal dif-ferentiation) are not systematically reviewed (cf.Hopper 1968, LeTendre et al. 2003, Oakes2005 [1985]). The reason for this omission isnot that less institutionalized ability groupingwould be irrelevant for inequality. Indeed, in-equality in learning has often been observed inthe American tracking system (Alexander et al.1978; Ayalon & Gamoran 2000; Gamoran& Berends 1987; Gamoran & Mare 1989;Hallinan 1994a, 1996b; Lucas 1999, 2001),and tracking is prevalent in all educationalsystems in one form or another. Internaldifferentiation, however, is hard to capture incross-national research, and few scholars haveattempted to do so.

Standardization comes in different forms aswell. First, the educational curriculum couldbe standardized across schools, with govern-ments deciding what is to be taught in schoolsand which levels should have been achievedat which grades. The opposite side of stan-dardization in this respect is school autonomy.Second, examinations can be standardized, as isthe case in, for example, the Netherlands andEngland. Third, standardization can refer tothe human and financial resources available toschools, e.g., in the form of teacher training andschool budgets.

As some have noted, inequality (both interms of dispersion and opportunities) is in-teresting particularly because of the potentialtrade-off between equality and efficiency oflearning (Gamoran & Mare 1989, Hanushek& Wossmann 2005, Micklewright & Schnepf2007, Wossmann 2008a). Educational systemsthat are well equipped to maximize average

408 Van de Werfhorst · Mijs

Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


student performance may be ill suited to guar-antee equality of learning. We present thesetrade-offs in a broader perspective of what wesee as the four core tasks of schooling. In thenext section we discuss an institutional perspec-tive of how national educational institutionsmay help or hinder the realization of those fourcore tasks. Then we discuss the data and re-search designs most commonly adopted by theempirical studies examined here, and addresssome methodological issues noted in the field.We also review the literature on inequality asdispersion and inequality of opportunity, be-fore drawing conclusions and presenting someoutlines for further research.

NATIONAL EDUCATIONALINSTITUTIONS AND CORETASKS OF SCHOOLING

To evaluate the relationship between educa-tional institutions and inequality, it is importantto position this relationship within a broadercontext of institutions and their effects. There-fore, we specify commonly agreed core tasks ofeducation in contemporary Western societies.It is relevant to study those core tasks simul-taneously to see whether a potential undesiredeffect of a particular educational institution (i.e.,being detrimental for realizing a core task) hasother desirable effects too.

We distinguish between four core tasks ofschooling: (a) to offer/promote equality of op-portunity; (b) to efficiently select and sort stu-dents on their abilities and interests; (c) to pro-vide skills relevant for the labor market; and(d ) to provide commitment to and skills rele-vant for active citizenship.

The first core task of schooling is to promoteequal educational opportunities to children ofdifferent backgrounds (the equal opportunitytask). Although we can develop separate mea-sures for equality of opportunity and equalityas dispersion, we should recognize three rea-sons why the two are linked. First, inequality ofposition can be linked to inequality of oppor-tunities intergenerationally, as large inequali-ties in parental educational attainment lead to

strong differences in their children’s chances inschool (Breen & Jonsson 2007). Second, onecould maintain that equality of educational op-portunity in terms of attained level is promotedif equality of position is granted at earlier stagesin the educational career. Third, empirically itappears that equal societies in terms of disper-sions are also more equal in terms of oppor-tunities (Boudon 1974, Duru-Bellat & Suchaut2005, Kenworthy 2008).

The second core task of schooling is to effi-ciently sort students according to their talentsand interests (the efficiency task). Students arenot equally equipped with different talents, andfurthermore show interest in diverse kinds ofwork and living. Education can help to sortstudents in finding their career goals. Theeducational system should group students sothat those with high learning skills have theopportunity to reach higher levels of school-ing than those with lower learning skills. Atthe same time, those with other talents oughtto have the chance to optimize their opportu-nities in the domains that more closely matchtheir aptitudes. The total production of knowl-edge and skills is then deemed optimized (givena particular budget for education).

The emphasis on efficient learning by opti-mizing standardized test scores originates froman understanding of the role of education thathas been developed largely from human cap-ital theory. Human capital theory lends it-self very well to education policy aimed atenhancing educational participation and im-proving academic achievement, as increasinginputs automatically lead to higher outputs(e.g., Hanushek 2006, Hanushek et al. 2008).Despite the evidence of a positive relation-ship between education and economic growth(but see Ramirez et al. 2006), the neoclassi-cal model of education has been widely crit-icized (Bowles & Gintis 2000, 2002; Wolf2002; Wolff 2006). However, given the noneco-nomic causal outcomes of schooling relatedto democratic citizenship, friendship networks,crime, and health (e.g., Becker 2007, Dee 2004,Hillygus 2005, Kubitschek & Hallinan 1998),even in the context of disputed causal effects

www.annualreviews.org • Achievement Inequality in Countries 409

Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


Table 1 Tasks of schooling and educational institutions

Tasks of schooling

Educational institutionsPromote equality of

opportunity Sort efficientlyPrepare foremployment

Prepare for activecitizenship

DifferentiationStandardization

of schooling on economic returns policy mak-ers may still want to maximize learning inschools.

A third core task is that education shouldprepare students for the labor market (the la-bor market task). This task implies that educa-tion provides skills that are productive for work,thereby aiding graduates in optimizing their la-bor market opportunities and employers in op-timizing production. One way to promote thelabor market task of schooling is by promot-ing efficiency: The level of skill is optimized,which thereby benefits the economy. Yet labormarket preparation may imply the optimizationnot only of the average level of skill, but also ofthe particular distribution of a type of skill, i.e.,vocationally specific or general. Educationalsystems differ widely in the emphasis placedon vocational or general skills. The Germaneducational system can, for example, be seenas strongly vocationally specific, whereas theAmerican and Scandinavian systems can be seenas more vigorously emphasizing general skills.

The fourth central task of educational insti-tutions is to stimulate citizens to become ac-tively involved in society (the active citizen-ship task). Educational institutions, it follows,should be seen as more than sites of productionof work skills. Their output should include theformation of critical citizens who can activelytake part in social and political life (Ten Dam& Volman 2003, 2007; Terwel 2005; Torney-Purta et al. 2001). This policy target impliesthat the formation of active citizenship cannotbe delegated fully to private caregivers. School-ing could have an active role in it, as it can helppromote equality in terms of civic competenceand commitment. Civic skills overlap partiallywith the skills demanded by the labor market.

For instance, general cognitive skills contributeto labor market productivity as well as to beinginformed about politics and, thereby, to politi-cal participation. Partly, though, civic skills andcommitment are created in subjects specificallydevoted to knowledge of politics, of formal in-stitutions, and of current affairs, subjects thatmay not contribute to learning skills relevantfor work.

We may think of ways in which differentkinds of educational institutions affect the like-lihood that the four core tasks of education canbe achieved. In this review, we focus on twoeducational institutional variables: the extentof differentiation in school types and the levelof national standardization.1 For heuristic rea-sons we plot these two dimensions of educa-tional systems against the four central tasks ofschooling systems. This heuristic framework al-lows us to examine which relationships exist be-tween educational institutions and the tasks ofschooling (see Table 1). When, within a row,a particular educational institution affects theachievement of one particular central task ofeducation positively, and at the same time theachievement of another central task negatively,we have detected a policy trade-off.

If we inspect the first two of these tasks,equality and efficiency, we encounter a cen-tral policy trade-off in educational design.In systems where students are selected anddifferentiated into separate schools as early asage 10 (Germany) or age 12 (the Netherlands),inequality may be higher than in systems

1One could add that countries vary in the extent to whichstudents have the opportunity to move between tracks(Gamoran 1992, Hallinan 1996b, Kerckhoff 2001), but fewscholars have examined this variability cross-nationally.


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


without external differentiation (e.g., Swe-den or the United States), both with regardto dispersion of achievement scores andsocial-structural effects on achievement. Yet,although inequality may be higher in differenti-ated systems, efficiency of learning may also behigher. When all students benefit from beingplaced in homogeneous classes, or when high-performing students benefit more than the low-ability students lose, the average performancemay be higher in strongly stratified educationalsystems. Some of the studies reviewed here haveexplicitly examined this trade-off empirically.

The same equality-efficiency trade-off canalso be detected with regard to the standard-ization of an educational system. Standardizededucational systems can be helpful in reducinginequalities by implementing central examina-tions that temper parental influence. Yet a stan-dardized system with regard to national curric-ula may impede competition between schools,which could lead to lower achievements ofstudents.

Another trade-off is the one between labormarket preparation and equality of opportu-nity. A stratified educational system (in particu-lar, one with a strong vocational sector) clearlyhelps youngsters in the transition process fromthe educational system to work (Arum & Shavit1995, Breen 2005, Muller & Gangl 2003, Shavit& Muller 1998). Yet there is a significant socialclass effect on choice of school type. If studentsenrolled in vocational tracks have fewer oppor-tunities to access tertiary education, stronglyvocationally oriented systems may enlarge so-cial class differences in the attainment of atertiary-level degree (OECD 2008).

A less well-known trade-off is that betweenlabor market preparation and citizenship edu-cation. It may be that educational systems thatperform well in the preparation for the labormarket perform worse when it comes to thecitizenship task of schooling. This could be thecase if early selection in the educational system,and a strong vocational orientation, both leadto improved labor market signaling but at thesame time increase variation in citizenship com-petences. A well-known view on social justice is

that principles of justice depend on the domainof life we are looking at (Miller 1999, Walzer1983). If we subscribe to the view that we shouldabide by an equality principle in issues related topolitics and that a merit principle should governthe labor market (cf. Miller 1999), we then eas-ily see that educational systems are functioningprecisely at the crossroads of these principles.Schooling, that is, should aim at both increasingthe visibility of the merits/skills obtained, whichis promoted by differentiation, and at equalizingthe kind of knowledge and skills that promotedemocratic equality, on which differentiationcould have detrimental effects.

DATA AND METHODOLOGY

Data

A vast and growing number of studies take aninternationally comparative view to the study ofeducational inequality. In this section we brieflydescribe the three most popularly studied datasets: the PIRLS, TIMSS, and PISA data. All ofthese data sets are collected in a wide range ofcountries, both Western and non-Western.2

The Progress in International Reading Lit-eracy Study (PIRLS) concerns students infourth grade of primary school, i.e., childrenof nine or ten years of age. Students are testedon reading skills and are asked to answer ques-tions on their situation at home and at school.Also, their parents and teachers are asked toprovide information concerning the child’s de-velopment, their own role in upbringing andeducating, and the school environment. Finally,the school board provides information on theschools’ organization and staff and on the char-acteristics of the student population. This infor-mation provides insight into differences in testscores as well as potentially influential factors

2For further information on the countries covered, werefer to the Web site of the International Associationfor the Evaluation of Educational Achievement (IEA)which runs the TIMSS and PIRLS data sets (http://www.iea.nl), and the Organisation for Economic Co-operationand Development (OECD) that collects the PISA data(http://www.pisa.oecd.org).


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


such as school environment, student character-istics, and family background.

To enable cross-country comparison, testscores are standardized to an international aver-age of 500 with a standard deviation of 100. Thedata are collected through a two-stage strati-fied sampling design: Schools are selected in thefirst round of sampling, and classes within thoseschools are selected in the second round. Thetarget is to select a minimum of 150 schools forall participating countries, of which one classper school is sampled.

The first PIRLS was conducted in 2001 anda second in 2006, and the intention is to main-tain this five-year cycle. Thirty-five countriesparticipate in PIRLS, with 150,000 respondentsin 2001. The test is financed jointly by the par-ticipating countries and the World Bank andis performed under the supervision of the In-ternational Association for the Evaluation ofEducation Achievement (IEA), an organizationof 62 countries that has been active since 1959.The same organization also coordinates theTIMSS.

The Trends in International Mathematicsand Science Study (TIMSS) tests children ingrade seven or eight. The questions, on math-ematics and science, are based heavily on thestandards of international curricula—the com-mon ground of the participating countries’curricula. The test scores (average 500, stan-dard deviation 100) are supplemented by back-ground questions for students and questions onfactors specific to schools, which are addressedby the school board. The sample is drawn inlike manner as that of PIRLS: At least 150schools per country are selected, of which onegrade seven and one grade eight class is ran-domly picked. The first TIMSS was conductedin 1995, and the test was repeated every fouryears, in 1999, 2003, and 2007.

The Program for International StudentAssessment (PISA) tests students 15 years ofage on their knowledge and skills in reading,mathematics, and science. Scores are standard-ized in like manner as that of PIRLS, albeitwith differences in what is measured: WhereasPIRLS tests only reading and TIMSS is based

on an international standard (curriculum), PISAis aimed toward assessing general skills andcompetencies related to real-world situations.

The data collection of PISA, through a two-stage stratified sampling, is organized some-what differently from that of PIRLS andTIMSS, as the target group is an age grouprather than a grade level. This choice may haveimplications for inequality as countries differ inthe extent to which students repeat classes incase of underperformance. In countries wheregrade repetition is more common, a relativelylarger share of students of one age cohort is ed-ucated in lower grades than in countries whererepetition is rare (Raudenbush & Kim 2002).This may affect their performance negatively,potentially leading to stronger social class ef-fects if repetition is unevenly distributed acrosssocial classes.

After schools have been randomly selected inthe first stage, individual students are selected inthe second. PISA is an initiative of the Organi-sation for Economic Co-operation and Devel-opment (OECD) and has been repeated everythree years since 2000. In that year 43 coun-tries participated; 41 did so in 2003, 57 in 2006,and 69 in 2009. The number of respondents is250,000 and above.

Research Designs

The literature discussed in this review can be di-vided roughly according to two methodologicalapproaches: studies that analyze cross-sectionaldata, linking educational inequality in a coun-try to specific characteristics of that coun-try’s educational system, and studies adopt-ing a difference-in-difference (DiD) model.The latter studies have been designed to dealwith the typical endogeneity problem of cross-sectional surveys; the enrollment in a track isendogenous on school performance (and socialclass). DiD models compare inequality mea-sured in two data sets at different points intime—commonly, one data set based on pri-mary education (i.e., before selection) and oneon secondary education (i.e., after selection intracks). If differentiation has a negative effect on


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


equality, inequality is assumed to increase moreacross educational transitions in those countriesthat have a large degree of differentiation. Thefirst data set (i.e., primary school) serves as apoint of reference for the analysis of the sec-ond data set, effectively holding constant anyunobserved differences between countries.

Although the above-mentioned property ofDiD models surely is one that appeals, theapproach suffers from one important weak-ness: comparing different data sets assumesthat educational ability or performance is mea-sured in like manner, whereas in fact it is not(Micklewright & Schnepf 2007). Ammermuller(2005) formalizes this point by stating thatthe measure derived from the second dataset contains an error component when usedas a second measure of the first data set.Ammermuller’s formalization, however, makesit possible to formulate assumptions underwhich a DiD model is worth using. The mostimportant of these assumptions, in research oninequality of educational opportunity, is thatthere is no variation between countries in thecorrelation between the error component andunmeasured aspects of family background. Thismay not be a very plausible assumption, as coun-tries are likely to vary systematically in theway that children are sorted on the basis ofnoncognitive characteristics. For instance, inunstandardized educational systems it is plau-sible that social background matters more fortrack placement relative to standardized sys-tems, precisely for reasons other than cognitiveones.

There are other techniques for modelingthe unobserved variation between countries incross-sectional data, by using a dummy vari-able for countries in mixed models (Brunello& Checchi 2007). This leads us to regard bothapproaches as important sources for our review.

Methodological Issues

Several methodological issues have beenraised (see Porter & Gamoran 2002b for acomprehensive overview). We discuss threemethodological issues that warrant particular

attention for our review on institutions andinequality.

First, there has been some debate as to theextent to which standardized tests can be usedto compare educational systems. The OECDclaims that PISA tests knowledge and skillsfor life and is not meant to measure whetherstudents have mastered a particular curricu-lum (OECD 2004, Kirsch et al. 2002). Yet, asGoldstein (2004) notes, PISA claims to evaluatethe performance of educational systems, whichseems at odds with the curriculum-independentfocus of PISA.

Moreover, an evaluation of educational sys-tems is difficult if one depends on cross-sectional data like PISA, TIMSS, and PIRLS. Innational educational studies it is commonplaceto use longitudinal data to examine the contri-bution of schools to learning. One needs infor-mation on earlier demonstrated ability to drawconclusions with regard to the causal impactof schooling on learning.3 Also in a compara-tive framework one would prefer longitudinaldata to examine whether educational systemsaffect (inequality in) learning. Inequality of ed-ucational opportunity results from two sortsof inequality: inequality in achievement (oftencalled primary effects; Boudon 1974, Eriksonet al. 2005, Goldthorpe 1996) and inequalityin educational choices conditional on achieve-ment (called secondary effects). Separatingprimary from secondary effects requires longi-tudinal data, where educational choice at a cer-tain point in the school career can be modeledto be conditional on achievement at an earlierpoint in time.4 Especially if one is interestedin cross-national variations in inequality, onewould be particularly interested to see whetherparticular educational institutional characteris-tics affect mostly primary or secondary effects.

3Note that, even if longitudinal data are available, an achieve-ment growth may be attributed to schooling but could alsobe attributable to other (e.g., family) influences.4The separation of primary and secondary effects is, however,also somewhat problematic, as a student’s achievement maydepend on the choices he or she is expecting to make in thefuture.


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


This has not been possible in the comparativesurveys to date.

Second, the DiD models used to simulatea longitudinal research design are not with-out problems. The analysis of country-leveldata prohibits the disentanglement of a coun-try’s variance into between-school and within-school components. Additionally, a comparisonof two surveys assumes that the tests are compa-rable, which rests on the assumption of one un-derlying distribution of, say, mathematical pro-ficiency (assuming no measurement error). Yetprecisely this feature of achievement tests hasbeen disputed.

That brings us to the third methodologi-cal issue—the validity of the assumptions un-derlying the psychometric tests implementedin the surveys. Are these assumptions plausi-ble given the findings on cross-national vari-ation in average scores and dispersion in testresults? The standardized tests in the varioussurveys have been developed on the basis ofItem Response Theory (IRT). There are twoassumptions in IRT that are relevant for ourstudy of institutional effects. First, IRT assumesthat there is an underlying true distributionof a student’s proficiency in a particular sub-ject (OECD 2005, Yamamoto & Kulick 2000).Second, the underlying true distribution has nofixed range but could, theoretically, reach in-finity. The distribution is assessed by repeatedmeasurement, offering various tests consistingof different items with different levels of diffi-culty. Under the assumption that there is a truedistribution of proficiency (or ability) indepen-dent of the test, it is possible that extremelybright students have a proficiency level abovethe threshold of the most difficult item. Thesestudents may be “assumed to be capable of atleast the achievements described for the high-est level” (OECD 2005, p. 256). Furthermore,if the proficiency level of students is below thelowest level defined, a student’s proficiency is“lower than that which PISA can reliably assess”(OECD 2005). These quotes illustrate that thePISA tests are seen as measuring an underlyingdistribution that exists independent of the testand that the true proficiency level of a student

may substantially exceed the highest definedlevel, illustrating an unbounded upper tail of thedistribution.

Several authors have criticized the assump-tion that such a true distribution of achieve-ment or ability exists independently of the test.Atkinson (1975, p. 89; see also Mayer 1960 andMicklewright & Schnepf 2007, p. 133) notedthat “the distribution [of ability] depends on themeasurement rod used and cannot be definedindependently of it.” Given that each surveyuses its own test, strictly speaking each survey“aims to measure somewhat different things”(Brown & Micklewright 2004, p. 42). Brownet al. (2006) examined the country medians anddispersions and found that country-level corre-lations between different domains were higherwithin surveys (e.g., reading and mathematicswithin PISA) than were correlations within do-mains between surveys (e.g., mathematics inPISA and mathematics in TIMSS). This findingmade the authors cautious about interpretingthe different surveys as measuring one under-lying concept of proficiency in reading, mathe-matics, or science. Another study, however, ar-gues that the country-level correlations on thedifferent tests are so large that one could speakof a general intelligence (g) factor underlyingall tests (Rindermann 2007).

The assumption of infinity in the underly-ing proficiency has not been discussed in theliterature. We do not think this assumption istoo important; for the understanding of aver-ages and inequalities in achievement, a test withan upper (and lower) bound is also valuable.However, the empirical results on the nonex-istence of a trade-off between equality and ef-ficiency may be related to the fact that thereis an upper limit in achievement. Similar toeducational attainment, which more evidentlyhas an upper limit, higher averages seem to co-incide with lower dispersions (e.g., Hauser &Featherman 1976, Rijken 1999, Thomas et al.2001). Decreasing dispersions with increasingaverage achievement test scores may possiblybe due to the fact that there is an upper limitin the distribution of achievement, similar toattainment.


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


Despite these methodological concerns, thedata are of high quality. As Porter & Gamoran(2002b, p. 8) conclude in the volume that theyedited about cross-national student achieve-ment data, “the level of methodological qualityis high, and therefore the findings of large-scalestudies are worth taking seriously.”

EMPIRICAL FINDINGS I:INEQUALITY AS DISPERSION

To what extent do countries vary in the dis-persion in achievement tests, and how/to whatextent is the level of dispersion related to edu-cational institutional characteristics?

Hanushek & Wossmann (2005) examine in-equality as dispersion through a DiD approach.The authors analyzed learning outcomes in pri-mary education (grade four) using PIRLS andTIMSS data and those in secondary education(grade eight and age 15) using TIMSS and PISAdata, respectively. They found, for each mea-sure of dispersion used, that once the variancein primary school is controlled for, the variancein learning increased more in those countriesin which students make the transition from pri-mary school to a highly differentiated secondaryschool, relative to students who stay in a com-prehensive system. In other words, controllingfor the fact that countries, for unknown reasons,have different levels of variance in achievementin primary school, early selection leads to anincrease in inequality at the secondary schoollevel.

Huang (2009) approaches the matter in adifferent way. His study takes as the inde-pendent variable classroom homogeneity, mea-sured as the between-school component of thetotal variance in students’ mathematics scores.This, he argues, is at the core of the theoreticalexplanation of why differentiation contributesto efficiency of learning. Teachers, the expla-nation holds, can teach more effectively whentheir students are of comparable ability levels(cf. Hallinan 1994b). In a study of TIMSS datafor 24 countries, Huang shows that this ex-planation does not hold. The effects of class-room homogeneity on inequality as dispersion

are strong and indicate that high-performingstudents stand to gain at the expense of low-performing students.

However, findings of other studies arenot in line with the hypothesis of increasingvariations among students with increasingdifferentiation. Micklewright & Schnepf(2007) study inequality as dispersion using thecross-sectional TIMSS, PISA, and PIRLS datasets. Dispersion is measured by differencesbetween achievement scores of students indifferent parts of the distribution (5th, 50th,and 95th percentiles). Micklewright & Schnepfshow that even in a low-dispersion countrylike France, the difference between P95 andP5 is 5.5 times as large as the average gain ofone year of schooling. In countries with higherdispersions, this factor goes up to 10.7 inGermany and 13.3 in the United States. It maythus be concluded that dispersion in learning issubstantial and that variations among countriesare to be taken seriously.

Although Micklewright & Schnepf (2007)find strong cross-national variation betweencountries in terms of dispersion, our ownanalysis of their findings shows no positive re-lationship between differentiation institutionsand dispersion. We calculated the correlationbetween a country’s level of dispersion andvarious indicators of differentiation: the num-ber of tracks available to a typical 14-year-oldstudent, the length of the tracked curriculumas a proportion of the total length of secondaryeducation, and the percentage of studentsenrolled in vocational education at uppersecondary and lower secondary levels (takenfrom the OECD’s Education at a Glancereports, and Brunello & Checchi 2007). Thesecorrelations were all around −0.20, indicatinglower dispersions when differentiation in thesystem increases. However, none of thesecorrelations turned out to be statistically sig-nificant (with Ns between 15 and 18 countries).Also Duru-Bellat & Suchaut (2005) find nosignificant statistical relationship betweentracking at age 14 and dispersion in testscores, which they examined by looking at theproportion of students classified in the top and


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


bottom categories of the standardized PISAranking.

Another study that refutes the hypothesisthat external differentiation leads to larger dis-persions in learning was done by Vandenberge(2006). He examines dispersion in the resid-uals of a regression equation predicting per-formance, controlling for social background.However, unlike other studies, Vandenbergeuses enrollment in the vocational track at age15 as an indicator of differentiation. The lat-ter indicator is unfortunate because placementin a vocational track at age 15 is unlikely tobe a good representation of the level of dif-ferentiation of the system. In Vandenberge’sstudy, Germany, for example, scores extremelylow on the indicator of differentiation becauseGerman vocational apprenticeships are offeredafter the age of 15. Given the strongly selec-tive nature of the German educational system,with selection around the age of 10 into separateschools, it may be more sensible to measure dif-ferentiation by age of first selection, regardlessof vocational enrollment. Indeed, Vandenbergefinds more evidence for an impact of differenti-ation with regard to the effect of interschoolsegregation, which is measured by the stan-dard deviation in performance across schoolswithin countries. However, as a concept, inter-school segregation confounds segregation froma source independent of the educational struc-ture (i.e., neighborhood segregation) with seg-regation imposed on the system through thestructure of selection and school organization.Thus, we are left unsettled with regard to theeffect of educational institutions. In summary,when it comes to the relationship between ex-ternal educational differentiation and inequalityas dispersion, the evidence is mixed.

Some of the studies mentioned above havealso analyzed the existence of a trade-off be-tween equality and efficiency. It appears thatmore dispersion goes hand in hand with a lowermedian score. Thus, no support is found for thehypothesized trade-off between the mean andvariance of learning achievement. It is not thecase that more efficient learning can be obtainedby institutions that allow for greater inequality

among students (Brown et al. 2006, Hanushek& Wossmann 2005, Micklewright & Schnepf2007). Furthermore, differentiation may reduceefficiency in another way because it poses a se-rious barrier for entry to higher education (VanElk et al. 2009, Marginson et al. 2007). Enroll-ment in tertiary education tends to be substan-tially lower in countries with strongly externallytracked systems.

In a study of seven Central and EasternEuropean (CEE) countries, Ammermuller et al.(2005) find that all four of the well-performingCEE countries (Hungary, Czech Republic,Slovakia, and Slovenia) have tracked educa-tional systems and higher levels of inequality,whereas the three worst-performing countries(Lithuania, Latvia, and Romania) all havecomprehensive educational systems and lowerlevels of inequality. Thus, among CEE coun-tries, there is evidence of a trade-off betweenefficiency and equality in school performance.

Besides these cross-national studies, somescholars have studied institutional changewithin countries from a differentiated to a com-prehensive system. Their research shows, for anumber of countries, that a development to-ward more comprehensive schooling increasesthe average performance and reduces inequal-ities (Duru-Bellat & Kieffer 2000, Gamoran1996, Gamoran & Weinstein 1998, Meghir &Palme 2005, Pekkarinen et al. 2009).

If we think of the mechanisms that explainwhy educational differentiation increases vari-ability between students but does not increasethe average performance substantially, we maylearn from the single-country studies on schoolcompositional effects. Although effects ofschool (or class) composition exist independentof the system of differentiation, tracking insti-tutions affect the within- and between-schoolvariation in school achievement and social class.In strongly differentiated systems, variabilitybetween schools is relatively high, and within-school variability relatively low (OECD 2004).

A few findings are relevant to explain thecross-national findings discussed above. First,tracking affects average performance morefor high-ability (high-track) students than for


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


low-ability (low-track) students, for whom theeffects of tracking are often negative (Hallinan1988, Hallinan & Kubitschek 1999, Hattie2002, Huang 2009, Kulik & Kulik 1982, Marsh1984). Also, low-ability students benefit morefrom mixed-ability grouping (Dobbelsteenet al. 2002, Thrupp et al. 2002, Van den Eeden& Terwel 1994, Willms 1986, Zimmer &Toma 2000). Second, even for high-abilitystudents the gains of tracking are not so great,often around 0.10–0.15 standard deviations inexperimental research designs. Third, the over-all effect of tracking on student performanceis very small and often negative (Figlio & Page2002, Hattie 2002, Michaelowa & Bourdon2006, Thrupp et al. 2002). There is thus littleevidence that tracking positively affects effi-cient learning, although it increases variabilityin achievement between students of differentlevels of demonstrated ability (Betts & Shkolnik2000, Hattie 2002, Huang 2009, Zimmer &Toma 2000). One explanation suggested forthe limited gains from tracking is that teachersdo not adjust their methods of instructionwhen they teach homogeneous classes, andwhen teachers do, it may particularly slowdown the instruction to low-ability studentsin a way that impedes their future learning(Gamoran & Berends 1987, Hattie 2002).

EMPIRICAL FINDINGS II:INEQUALITY OF OPPORTUNITY

Differentiation and FamilyBackground Effects

Since the 1960s, educational systems have beentypified with regard to their impact on in-equalities (Hopper 1968, Turner 1960). Somecross-national empirical studies on achieve-ment focused on institutional and structuralfactors other than educational that affectedinequalities in learning (e.g., Baker et al. 2002,Heyneman & Loxley 1983). Other early stud-ies have, however, focused on early selectionin comparison to comprehensive schooling,pointing to larger effects of social backgroundin systems where students are selected at an

earlier age (Comber & Keeves 1973; Husen1967, 1973). More recent studies on the impactof educational differentiation on equality of op-portunity by family background confirm theseearlier findings: Inequalities are magnified byexternal educational differentiation. Followingis an in-depth discussion of these recentstudies. Fortunately, all of the cross-nationalstudent achievement surveys collect informa-tion on family background, although morehomogenization of measurement instrumentshas been called for (Buchmann 2002).

Schutz et al. (2008) focus on differentia-tion as a potential explanation for cross-nationaldifferences in inequality of opportunity. Thenumber of books in the household, which theauthors take to be the best reflection of fam-ily income, has a stronger effect on studentperformance in countries that track studentsin different school types on the basis of abil-ity. Note that sociologists tend to argue thatcultural possessions such as books are a re-flection more of cultural than of financial re-sources available in the household, and culturalclimate tends to have a stronger effect on stu-dent achievement than financial resources have(Marks 2005, Park 2008a). Studying institu-tional effects more comprehensively than otherstudies have done, Marks (2005) shows that theexplanatory power of social class on literacyis higher in countries with lower participationrates in university education, more educationaltracks at age 15, an earlier age of first selec-tion in the schooling system, a higher between-school variance in learning outcomes, and moreincome inequality (although the latter relation-ship is modest). Thus, there is clear evidencethat differentiation, measured in various ways,increases inequalities in learning by social class.Horn (2009) comes to the same conclusion inhis cross-national study of 29 countries usingPISA 2003 data, and Bauer & Riphahn (2006)confirm these findings as well for the 26 cantonsof Switzerland that vary in timing of tracking.

Wossmann (2008a) makes use of TIMSS1995 data to assess the effect of familybackground (parental education and num-ber of books in the household) on students’


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


mathematics achievement. His study showsthat in the United States and in all 17 Eu-ropean countries included in his study, fam-ily background is correlated with a student’sperformance in mathematics. The effect variesfrom relatively weak in Portugal, France, andWallonia (Belgium), to strong in England,Germany, and Greece. When comparing themathematics scores of students from highly ed-ucated parents to those with only primary edu-cation, the differences are highest in the UnitedStates—higher than in any European country.

Importantly, Wossmann (2008a) and Schutzet al. (2008) show that there is no relationshipbetween social inequality of learning and av-erage school performance, which refutes thehypothesis of a possible trade-off between ef-ficiency and equality. Elsewhere Wossmann(2008b) argued that a trade-off between effi-ciency and equality of opportunity, if at alllikely, is to be found only at the higher stagesof education. From the life-cycle model of hu-man capital formation, which shows that earlylearning enhances later learning (e.g., Heckman2000), it follows that it would be more efficientto offer further education to those with highlevels of early performance. This would en-hance inequality in learning. At the early stageof (pre)schooling, however, more gains can beexpected for children from disadvantaged back-grounds, which implicates a complementarity,rather than a trade-off, between efficiency andequity.

Ammermuller (2005) has conducted a DiDapproach to study inequality of educational op-portunity for 14 countries using PIRLS (2001)and PISA (2000) data. For all family back-ground variables, this study finds that the ef-fects increase most strongly in those countriesthat have either a highly differentiated educa-tional system or a large private school sector.The yearly amount of instruction at school re-duces the effect of family background, whereasthe level of autonomy for schools increases theeffect of parents’ educational stance on theirchildren’s reading skills.

Brunello & Checchi’s (2007) research is themost comprehensive study on the relationship

between differentiation and inequality of op-portunity by family background. The authorsused multiple data sets, on students as wellas on (young) adults, to analyze the effectof family background on educational and onlabor market opportunity. Differentiationis operationalized in two ways: length ofschool-type differentiation within the edu-cational system and percentage of studentsenrolled in secondary vocational education. Inaddition to the standardized test scores usedin most research, Brunello & Checchi includeindicators of educational attainment, accessto tertiary education, school dropout rates,language and mathematical skills, as well as in-dicators of early labor market experience suchas employment, training, and income. Theirdata sources are the European CommunityHousehold Panel (ECHP), the InternationalSocial Survey Project (ISSP), the InternationalAdult Literacy Survey (IALS), and PISA 2003.

The authors show how, for a number ofoutcomes, the effect of family background in-creases with length of tracking. This concernsthe effect of family background on educationalattainment, school dropout rates, access to andenrollment in tertiary education, and job in-come. With regard to mathematics, however,the outcomes for young adults using IALSdata contradict the hypothesis that differenti-ation increases inequality of opportunity; fam-ily background’s positive effect on test scoresdecreases as tracking length increases. The au-thors offer two possible explanations for the dis-crepancy between their findings for mid-teenstudents and young adults. First, time spent inthe labor market could compensate for the neg-ative effect of tracking on equality of opportu-nity. IALS respondents often have some labormarket experience, and those in countries withearly tracking tend to have more experience asthey are less likely to enroll in tertiary educa-tion. A second explanation is that the full effectof tracking can better be assessed at a later age(IALS) than at age 13 or 15.

With regard to lifelong learning, the authorsfind also that the positive effect of family back-ground decreases as differentiation increases. In


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


other words, although students from highly ed-ucated parents are more likely to receive a formof training anywhere, this effect is stronger incountries with a comprehensive system of edu-cation than in those with highly differentiatededucational systems. Importantly, Brunello &Checchi conclude that the length of school-type differentiation increases inequality of op-portunity but that the vocational orientationdoes not.

Differentiation andRace/Ethnicity Effects

In addition to the effects of family background,race/ethnicity effects are sometimes found todepend on the educational institutional struc-ture. School tracking has been related to eth-nic inequality in European cities and coun-tries (Crul & Holdaway 2009; Crul & Schnei-der 2007, 2009; Crul & Vermeulen 2003).Israeli studies have shown that tracking im-pedes the educational opportunities of Arab andJewish minorities relative to the majority Jew-ish ethnic group (Shavit 1984, 1990). Entorf &Lauk (2008), furthermore, showed that differ-entiated educational systems tend to magnifypreviously existing inequalities in educationalperformance between migrant students of lowsocioeconomic background and students frommore privileged families. Peer effects in high-ability tracks favor students with an initial ad-vantage, whereas disadvantaged students expe-rience negative peer effects in the lower tracks(a finding confirmed by ethnographic accounts;e.g., Paulle 2005). Immigrant students, the au-thors report, would benefit from a more diversestudent population in terms of educational abil-ity within, rather than between, schools.

Race/ethnicity has also been associated withtrack mobility, although only from a single-country perspective. Hallinan (1996a) finds thatstudents of color in the United States, in gen-eral, are disproportionately assigned to lowertracks in both English and mathematics andare more likely than white students to dropout of these tracks. Studying track mobility inthe Netherlands, Kalmijn & Kraaykamp (2003)

constructed an event-history model for sec-ondary school careers and showed that studentsof ethnic minorities were less likely than nonmi-nority students to experience downward trackmobility, although they were more likely todrop out of school before completion.

In the North American context, race iscorrelated strongly with indicators of stu-dent achievement. Studies show that AfricanAmerican students are disproportionatelyassigned to lower tracks (Darling-Hammond1994; Hallinan 1991, 1992; Lucas 1999; Oakes1990). When student achievement is controlledfor, however, the racial effect is ambiguous.Some studies find that race differences intrack assignment diminish or disappear (e.g.,Hallinan 1994a, Pallas et al. 1994), whereasothers indicate that minorities are favoredin the assignment to higher-ability groups(Gamoran & Mare 1989, Stanat & Christensen2006, Van de Werfhorst & Van Tubergen2007). An unsettled matter, despite receivingparticular attention in the literature, is whethereducational tracking has negative ethnicityeffects on inequality of educational opportunityindependent of social class effects. In additionto the previously cited studies, Alexander &Cook (1982) find no evidence for such an effect,although Hallinan (1994b) in a later study does.

Standardization and FamilyBackground Effects

It is important to examine the repercussionsof standardization alongside differentiation, asstandardization makes the process of groupingstudents transparent and objectifies its criteria.Indicators of standardization, such as centralexaminations, a national curriculum, and stan-dardized school resources (regarding budgets,staffing decisions, teacher training, establishingsalaries), reduce the influence of social originon student performance (although more soin TIMSS than in PISA), whereas schoolautonomy increases the influence of socialorigin (Horn 2009; Muller & Schiller 2000;Park 2008b; Schutz et al. 2008; Wossmann2003a,b, 2005). Additionally, standardized


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


educational systems are said to reflect thatqualifications represent the same skill levelthroughout the country, implying strongersignaling to the labor market (Shavit & Muller1998). However, standardization is not signif-icantly related to intergenerational mobilityin educational attainment, as has been shownwith IALS data (Pfeffer 2008).

An interesting study to examine in moredetail is Stevenson & Baker’s (1991, p. 1) in-vestigation of what they call the implementedcurriculum—“that portion of the curriculumthat is taught to students in the classroom”—with an interest in explaining variability byaddressing the level of standardization (statecontrol over the curriculum). In their analysisof SIMS data (Second International Math-ematics Study, the predecessor of TIMSS)for 15 educational systems, they find thatstandardization decreases the effect of studentand teacher characteristics on how muchmathematics is taught in class. Students instandardized school systems, that is, are morelikely to get the same education, whereas schoolsystems with more school, local, or provincialautonomy exhibit higher variability within andacross schools in what students are taught.Their study is criticized by Westbury & Hsu(1996), who argue that it is not standardizationbut within-country differences in school types(external differentiation) or within-schooltracks (internal differentiation) that drive theresults.

With regard to efficiency, school auton-omy boosts average performance only in coun-tries with central examinations (Dronkers &Robert 2008, Fuchs & Wossmann 2007). So,in the context of central examinations warrant-ing school accountability, there is a potentialtrade-off between equality and efficiency, forautonomy tends not only to enhance averageperformance, but also to increase the levels ofinequality of educational opportunity. How-ever, in the context of more limited levels ofexternal accountability, we may tentatively con-clude that there is no such trade-off, as the effi-ciency gains of school autonomy are much lesspronounced.

Studying educational differentiation andstandardization jointly is relevant for tworeasons. First, as both indicators may be re-lated, one needs to control for standardizationto assess the impact of differentiation, and viceversa. Pfeffer (2008) examines both institutionalcharacteristics with regard to educational mo-bility between parents and children with regardto their highest attained level of schooling. Hisresults show that differentiation hinders edu-cational mobility but that standardization hasno effect at all. Also, Horn (2009) analyzes theimpact of differentiation and standardizationsimultaneously using PISA data and shows thatinequality of opportunity increases with differ-entiation and school autonomy and that uncon-trolled effects are largely the same as controlledeffects due to the low correlation betweenstandardization and differentiation indicators.

Second, it has been argued that there isan interaction effect between differentiationand standardization. In a comparison betweenIsrael and the United States, Ayalon &Gamoran (2000) demonstrate that similar edu-cational reforms (i.e., differentiation) can leadto different outcomes due to differences in thedegree of the country’s level of standardization.Israel’s standardized curriculum and nationalexaminations appear to be an incentive forachievement among teachers and students inall levels of schooling, whereas their absence inthe United States reinforces inequality withoutraising average scores. In other words, in a stan-dardized school setting with limited autonomyof schools with regard to standards, curriculumdevelopment, and teacher training, the effectof tracking on inequalities may be temperedrelative to decentralized systems. Such aneffect is also suggested by Broaded (1997) in hisstudy of Taiwan, where he found very limitedinfluence of family background on educationalaspirations and senior high school placement.His results indicate that the standardizedcurriculum of the Taiwanese educationalsystem, as well as its practice of standardizedexamination, provide the institutional contextin which tracking helps realize the efficiencytarget without negatively affecting inequality of


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


Table 2 Summary of empirical findings on tasks of schooling and educational institutions

Tasks of schooling

Educational institutionsPromote equality of

opportunity Sort efficientlyPrepare foremployment

Prepare for activecitizenship

Differentiation − − ± ?Standardization + + + ?

+: Evidence points to the task benefiting from strengthening this institutional variable.− : Evidence points to the task being impeded by strengthening this institutional variable.± : Mixed or weak evidence regarding the relationship between institution and task.?: Underinvestigated relationships between institution and task.

opportunity. These findings should, however,be taken as tentative findings because the studyis limited to just one institutional setting.

CONCLUSIONS

We can draw several conclusions regarding theimpact of national educational institutions oninequality in student achievement. In Table 2we return to the grid that cross-classifies ed-ucational institutions (differentiation and stan-dardization) by four core tasks of schooling. Wehave now put pluses and minuses in the boxes,where an institutional feature promotes (+) orimpedes (−) the realization of one of the coretasks.

A first conclusion is that equality in achieve-ment is negatively affected by differentiationinstitutions. Both the variability among stu-dents and the dependence of achievementon social class and race/ethnicity are higherin educational systems that track students indifferent school types and school locationsrelative to systems that offer comprehensiveschools. This effect is sizeable and is confirmedby national studies on the mechanisms under-lying tracking effects. Moreover, with regardto inequality of opportunity, this finding isdisputed rarely, although no consensus is foundwith regard to inequality as dispersion. A posi-tive relationship is consistently found betweenstandardization and equality of opportunity.Central examinations, especially, tend to havea positive effect on equality.

Second, the evidence with regard to ef-ficiency in learning, operationalized by the

average performance in a country, shows thateducational differentiation leads to lower,rather than higher, average achievement ina number of subjects. These two conclusionsimply that there is no evidence that averageperformance could be improved by allowingfor higher inequality by means of educationaldifferentiation. Hence, similar to the conclu-sions in the single-country tracking studies,the comparative literature refutes a trade-offbetween equality and efficiency. Thus, there isno evidence that average performance could beenhanced by allowing for a higher dispersionby means of tracking. This lack of evidence fora trade-off may result from a ceiling effect intest scores. Just like educational attainment,achievement has an upper limit, and higher av-erages may go together with lower dispersionsas a result. With regard to standardization, it istrue that average performance is increased bymore standardization.

Although beyond the scope of our review,we may tentatively draw some conclusions withregard to the labor market and civic outcomesof schooling. This provides for a more compre-hensive picture with regard to the institutionaleffects on core tasks of schooling. With regardto employment outcomes, the evidence ismixed. The effect of education on labor marketoutcomes tends to be stronger in stronglystratified systems owing to restricted access totertiary education and clearer signaling (Breen2005, Shavit & Muller 1998). Yet, to the extentthat tracking decreases average performanceand does not offer equal opportunity to allstudents, countries may not optimize their


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


human capital potential to the extent thatwould have been possible by offering com-prehensive schooling. Standardization seemsto reduce the noise in the signal providedby schooling for work, so there is reason toconclude that standardization enhances thelabor market task of schooling.

With regard to civic skills, the evidence, un-fortunately, is rather meager. Although differ-entiated systems have been claimed to magnifyinequalities in civic skills (Terwel 2005), andsingle-country studies have shown differencesacross tracks in the kinds of civic competenciesacquired (Damico et al. 1996; Niemi & Junn1998; Ten Dam & Volman 2003, 2007), thereare no studies that have compared the influenceof educational institutional characteristics suchas tracking on citizenship outcomes in a cross-national perspective.

The total impact of differentiation is likelylarger than the comparative literature is able toshow, for two reasons. First, in systems with-out external differentiation, forms of trackingalso exist that are related to inequalities. Moregenerally, educational systems are character-ized by various sorts of segregation of stu-dents of different ability levels, social classes,and ethnic/racial groups that are not capturedby national-level indicators of differentiation.The United States is a case in point: Becauseof the common approach of taking national-level indicators, its school system is often clas-sified in cross-country research as comprehen-sive, although a vast body of American researchclearly shows otherwise. It is important that fu-ture comparative research take up this issue oftracking within school types.

Second, given that externally differentiatinginstitutions reduce the number of students

eligible for college, initial inequalities in learn-ing at mid-teen age are likely to be magnified infurther educational attainment. Some studieshave shown that school type has a tremendous(and realistic) effect on college expectations andaspirations in differentiating educational sys-tems (Buchmann & Dalton 2002, Buchmann& Park 2009). Hence, it seems to be the casethat differentiated educational systems channeleducational careers at a very early stage,in comparison to systems without externaldifferentiation. To gain more insight into theseprocesses, future research may want to examineschool continuation patterns conditional uponachievement in a comparative framework. Suchresearch would shed light on how institutionalfactors influence the relative size of primaryand secondary effects of social and ethnic/racialbackground. This distinction is of the utmostimportance for educational policy, as reducingprimary effects requires policy measures differ-ent from those of reducing secondary effects.Obviously, longitudinal school career data areneeded for such an analysis and require that theauthorities of cross-national student achieve-ment data build a longitudinal design into theirsurveys.

Another line of research that deserves fur-ther attention is the interaction between stan-dardization and differentiation to test the claimmade by Ayalon & Gamoran (2000) that stan-dardization can reduce the negative effects ofdifferentiation on equality. It is plausible thatstandardization of educational systems dimin-ishes the negative effects of external differentia-tion on inequality, as allocation to school typesis based on objective and transparent criteria,reducing the effects of social origin on educa-tional decision making.

SUMMARY POINTS

1. Countries with a more strongly differentiated (school-type tracked) educational systemtend to have higher levels of inequality of educational opportunity by social class andrace/ethnicity.

2. Countries with a more standardized educational system have lower levels of inequalityof opportunity compared to those with unstandardized systems.


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


3. There is no clear pattern in the relationship between external differentiation and inequal-ity in terms of dispersion in learning.

4. There is no evidence that average achievement is higher in countries with larger disper-sions in achievement tests. This implies that there is no evidence for the existence of atrade-off between equality and efficiency in this regard.

FUTURE ISSUES

1. More research is needed on the interplay between standardization and differentia-tion. Does differentiation lead to higher levels of inequality in countries with lowstandardization?

2. Future research could also study educational institutional effects on the level and variationin social and political participation of youth and young adults. Is inequality in theseoutcomes larger in more strongly differentiating educational systems?

DISCLOSURE STATEMENT

The authors are not aware of any affiliations, memberships, funding, or financial holdings thatmight be perceived as affecting the objectivity of this review.

ACKNOWLEDGMENTS

We thank Jaap Dronkers and Sjoerd Karsten for their comments on an earlier version of thispaper.

LITERATURE CITED

Alexander KL, Cook MA. 1982. Curricula and coursework: a surprise ending to a familiar story. Am. Sociol.Rev. 47:626–40

Alexander KL, Cook MA, McDill EL. 1978. Curriculum tracking and educational stratification: some furtherevidence. Am. Sociol. Rev. 43:47–66

Ammermuller A. 2005. Educational opportunities and the role of institutions. ZEW Discuss. Pap. 05-44, Cent.Eur. Econ. Res., Mannheim

Ammermuller A, Heijke H, Wossmann L. 2005. Schooling quality in Eastern Europe: educational productionduring transition. Econ. Educ. Rev. 24:579–99

Arum R, Shavit Y. 1995. Secondary vocational education and the transition from school to work. Sociol. Educ.68:187–204

Atkinson AB. 1975. The Economics of Inequality. Oxford: ClarendonAyalon H, Gamoran A. 2000. Stratification in academic secondary programs and educational inequality in

Israel and the United States. Comp. Educ. Rev. 44:54–80Baker DP, Goesling B, LeTendre GK. 2002. Socioeconomic status, school quality, and national economic

development: a cross-national analysis of the “Heyneman & Loxley effect” on mathematics and scienceachievement. Comp. Educ. Rev. 46:291–312

Bauer P, Riphahn RT. 2006. Timing of school tracking as a determinant of intergenerational transmission ofeducation. Econ. Lett. 91:90–97

Becker GS. 2007. Health as human capital: synthesis and extensions. Oxford Econ. Pap. 59:379–410


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


Betts JR, Shkolnik JL. 2000. The effects of ability grouping on student achievement and resource allocationin secondary school. Econ. Educ. Rev. 19:1–15

Boudon R. 1974. Education, Opportunity, and Social Inequality. New York: WileyBowles S. 2004. Microeconomics. Behavior, Institutions, and Evolution. Princeton, NJ: Princeton Univ. PressBowles S, Gintis H. 2000. Does schooling raise earnings by making people smarter? In Meritocracy and Economic

Inequality, ed. K Arrow, S Bowles, S Durlauf, pp. 118–36. Princeton, NJ: Princeton Univ. PressBowles S, Gintis H. 2002. Schooling in capitalist America revisited. Sociol. Educ. 75:1–18Breen R. 2005. Explaining cross-national variation in youth unemployment: market and institutional factors.

Eur. Sociol. Rev. 21:125–34Breen R, Jonsson JO. 2007. Explaining change in social fluidity: educational equalization and educational

expansion in twentieth-century Sweden. Am. J. Sociol. 112:1775–810Broaded CM. 1997. The limits and possibilities of tracking: some evidence from Taiwan. Sociol. Educ. 70:36–53Brown G, Micklewright J. 2004. Using International Surveys of Achievement and Literacy: A View from the Outside.

Montreal: UNESCO Inst. Stat.Brown G, Micklewright J, Schnepf SV, Waldmann R. 2006. International surveys of educational achievement:

how robust are the findings? J. R. Stat. Soc. A 170:623–46Brunello G, Checchi D. 2007. Does school tracking affect equality of opportunity? New international evidence.

Econ. Policy 22:781–861Buchmann C. 2002. Measuring family background in international studies of education: conceptual issues and

methodological challenges. See Porter & Gamoran 2002a, pp. 150–97Buchmann C, Dalton B. 2002. Interpersonal influences and educational aspirations in 12 countries: the im-

portance of institutional context. Sociol. Educ. 75:99–122Buchmann C, Park H. 2009. Stratification and the formation of expectations in highly-differentiated educa-

tional systems. Res. Soc. Stratif. Mobil. 27:245–67Comber LC, Keeves J. 1973. Science Education in Nineteen Countries. Stockholm: Almqvist & WiksellCrul M, Holdaway J. 2009. Children of immigrants in schools in New York and Amsterdam: the factors

shaping attainment. Teach. Coll. Rec. 111:1476–507Crul M, Schneider J. 2007. Integration of Turkish second-generation men and women in Germany and the

Netherlands. The impact of differences in academic tracking systems. In Educating Immigrant Youth:Pathways to Employment and Citizenship in International Perspective, ed. M Crul, J Holdaway, C Roberts.Spec. Issue Teach. Coll. Rec.

Crul M, Schneider J. 2009. Children of Turkish immigrants in Germany and the Netherlands: the impact ofdifferences in vocational and academic tracking systems. Teach. Coll. Rec. 111:1508–27

Crul M, Vermeulen H. 2003. The second generation in Europe. Introduction to the special issue. Int. Migr.Rev. 37:965–86

Damico AJ, Conway MM, Damico SB. 1996. Democratic education: the associational life of African-American andwhite high school students. Presented at Annu. Meet. Am. Educ. Res. Assoc., April, New York City

Darling-Hammond L. 1994. Performance-based assessment and educational equity. Harvard Educ. Rev. 64:5–30

Dee TS. 2004. Are there civic returns to education? J. Public Econ. 88:1697–720Dobbelsteen S, Levin J, Oosterbeek H. 2002. The causal effect of class size on scholastic achievement: distin-

guishing the pure class size effect from the effect of changes in class composition. Oxford Bull. Econ. Stat.64:17–38

Dronkers J, Robert P. 2008. Differences in scholastic achievement of public, private government-dependent,and private independent schools. A cross-national analysis. Educ. Policy 22:541–77

Duru-Bellat M, Kieffer A. 2000. Inequalities in educational opportunities in France: educational expansion,democratization or shifting barriers? J. Educ. Policy 15:333–52

Duru-Bellat M, Suchaut B. 2005. Organisation and context, efficiency and equity of educational systems: whatPISA tells us. Eur. Educ. Res. J. 4:181–94

Entorf H, Lauk M. 2008. Peer effects, social multipliers and migrants at school: an international comparison.J. Ethnic Migr. Stud. 34:633–54

Erikson R, Goldthorpe JH, Jackson M, Yaish M, Cox DR. 2005. On class differentials in educational attain-ment. Proc. Natl. Acad. Sci. USA 102:9730–33


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


Figlio DN, Page ME. 2002. School choice and the distributional effects of ability tracking: Does separationincrease inequality? J. Urban Econ. 51:497–514

Fuchs T, Wossmann L. 2007. What accounts for international differences in student performance? A re-examination using PISA data. Empir. Econ. 32:433–64

Gamoran A. 1992. The variable effects of high school tracking. Am. Sociol. Rev. 57:812–28Gamoran A. 1996. Curriculum standardization and equality of opportunity in Scottish secondary education:

1984–90. Sociol. Educ. 69:1–21Gamoran A, Berends M. 1987. The effects of stratification in secondary schools: synthesis of survey and

ethnographic research. Rev. Educ. Res. 57:415–35Gamoran A, Mare RD. 1989. Secondary school tracking and educational inequality: compensation, reinforce-

ment, or neutrality? Am. J. Sociol. 94:1146–83Gamoran A, Weinstein M. 1998. Differentiation and opportunity in restructured schools. Am. J. Educ.

106:385–415Goldstein H. 2004. International comparisons of student attainment: some issues arising from the PISA study.

Assess. Educ. 11:319–30Goldthorpe JH. 1996. Class analysis and the reorientation of class theory: the case of persisting differentials

in educational attainment. Br. J. Sociol. 47:481–505Hallinan MT. 1988. Equality of educational opportunity. Annu. Rev. Sociol. 14:249–68Hallinan MT. 1991. School differences in tracking structures and track assignments. J. Res. Adolesc. 1:251–75Hallinan MT. 1992. Organization of students for instruction in the middle school. Sociol. Educ. 65:114–27Hallinan MT. 1994a. School differences in tracking effects on achievement. Soc. Forces 72:799–820Hallinan MT. 1994b. Tracking: from theory to practice. Sociol. Educ. 67:79–84Hallinan MT. 1996a. Race effects on students’ track mobility in high school. Soc. Psychol. Educ. 1:1–24Hallinan MT. 1996b. Track mobility in secondary school. Soc. Forces 74:983–1002Hallinan MT, Kubitschek WN. 1999. Curriculum differentiation and high school achievement. Soc. Psychol.

Educ. 3:41–62Hanushek EA. 2006. Alternative school policies and the benefits of general cognitive skills. Econ. Educ. Rev.

25:447–62Hanushek EA, Jamison DT, Jamison EA, Wossmann L. 2008. Education and economic growth: It’s not just

going to school, but learning something while there that matters. Educ. Next 8:62–70Hanushek EA, Wosmann L. 2005. Does educational tracking affect performance and inequality? Differences-

in-differences evidence across countries. Econ. J. 116:C63–76Hattie JAC. 2002. Classroom composition and peer effects. Int. J. Educ. Res. 35:449–81Hauser RM, Featherman DL. 1976. Equality of schooling: trends and prospects. Sociol. Educ. 49:99–120Heckman JJ. 2000. Policies to foster human capital. Res. Econ. 54:3–56Heyneman SP, Loxley WA. 1983. The effect of primary-school quality on academic achievement across

twenty-nine high and low-income countries. Am. J. Sociol. 88:1162–94Hillygus DS. 2005. The missing link: exploring the relationship between higher education and political en-

gagement. Polit. Behav. 27:25–47Hopper EI. 1968. A typology for the classification of educational systems. Sociology 2:29–46Horn D. 2009. Age of selection counts: a cross-country analysis of educational institutions. Educ. Res. Eval.

15:343–66Huang M-H. 2009. Classroom homogeneity and the distribution of student math performance: a country-level

fixed-effects analysis. Soc. Sci. Res. 38:781–91Husen T, ed. 1967. International Study of Achievement in Mathematics. A Comparison of Twelve Countries.

Stockholm: Almqvist & WiksellHusen T. 1973. The standard of the elite: some findings from the IEA international survey in mathematics

and science. Acta Sociol. 16:305–23Kalmijn M, Kraaykamp G. 2003. Dropout and downward mobility in the educational career: an event-history

analysis of ethnic schooling differences in the Netherlands. Educ. Res. Eval. 9:265–87Kenworthy L. 2008. Jobs with Equality. New York: Oxford Univ. PressKerckhoff AC. 2001. Education and social stratification processes in comparative perspective. Sociol. Educ.

Extra Issue:3–18


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


Kirsch I, De Jong J, Lafontaine D, McQueen J, Mendelovits J, Monseur C. 2002. Reading for Change: Perfor-mance and Engagement Across Countries: Results from PISA 2000. Paris: OECD

Kubitschek WN, Hallinan MT. 1998. Tracking and students’ friendships. Soc. Psychol. Q. 61:1–15Kulik C-L, Kulik JA. 1982. Effects of ability grouping on secondary school students: a meta-analysis of

evaluation findings. Am. Educ. Res. J. 19:415–28LeTendre GK, Hofer BK, Shimizu H. 2003. What is tracking? Cultural expectations in the United States,

Germany, and Japan. Am. Educ. Res. J. 40:43–89Lucas SR. 1999. Tracking Inequality. Stratification and Mobility in American High Schools. New York: Teach.

Coll. PressLucas SR. 2001. Effectively maintained inequality: education transitions, track mobility, and social background

effects. Am. J. Sociol. 106:1642–90Marginson S, Weko T, Channon N, Luukkonen T, Oberg J. 2007. Thematic Review of Tertiary Education.

Paris: OECDMarks GN. 2005. Cross-national differences and accounting for social class inequalities in education. Int.

Sociol. 20:483–505Marsh HW. 1984. Self-concept, social comparison, and ability grouping: a reply to Kulik and Kulik. Am. Educ.

Res. J. 21:799–806Mayer T. 1960. The distribution of ability and earnings. Rev. Econ. Stat. 42:189–95Meghir C, Palme M. 2005. Educational reform, ability, and family background. Am. Econ. Rev. 95:414–24Michaelowa K, Bourdon J. 2006. The impact of student diversity in secondary schools: an analysis of the international

PISA data and implications for the German education system. HWWI Res. Pap. 3–2, Hamburg Inst. Int.Econ., Hamburg

Micklewright J, Schnepf SV. 2007. Inequality of learning in industrialized countries. In Inequality and PovertyRe-Examined, ed. SP Jenkins, J Micklewright, pp. 129–45. Oxford: Oxford Univ. Press

Miller D. 1999. Principles of Social Justice. Cambridge, MA: Harvard Univ. PressMuller C, Schiller KS. 2000. Leveling the playing field? Students’ educational attainment and states’ perfor-

mance testing. Sociol. Educ. 73:196–218Muller W, Gangl M, eds. 2003. Transitions From Education to Work in Europe—the Integration of Youth into EU

Labor Markets. Oxford: Oxford Univ. PressNiemi RG, Junn J. 1998. Civic Education. What Makes Students Learn? New Haven, CT: Yale Univ. PressOakes J. 1990. Opportunities, achievement and choice: women and minority students in science and mathe-

matics. Rev. Res. Educ. 16:153–222Oakes J. 2005 [1985]. Keeping Track. How Schools Structure Inequality. New Haven, CT: Yale Univ. PressOECD. 2004. Learning for Tomorrow’s World: First Results from PISA 2003. Paris: OECDOECD. 2005. PISA 2003. Tech. Rep., Org. Econ. Coop. Dev., Paris: OECDOECD. 2008. Tertiary Education for the Knowledge Society. Vol. II. Paris: OECDPallas AM, Entwistle DR, Alexander KL, Stukla FM. 1994. Ability-group effects: instructional, social, or

institutional? Sociol. Educ. 67:27–46Park H. 2008a. Home literacy environments and children’s reading performance: a comparative study of 25

countries. Educ. Res. Eval. 14:489–505Park H. 2008b. The varied educational effects of parent-child communication: a comparative study of fourteen

countries. Comp. Educ. Rev. 52:219–43Paulle B. 2005. Anxiety and Intimidation in the Bronx and the Bijlmer: An Ethnographic Comparison of Two Schools.

Amsterdam: Dutch Univ. PressPekkarinen T, Uusitalo R, Kerr S. 2009. School tracking and development of cognitive skills. IZA Discuss. Pap.

4058, Inst. Study Labor, BonnPfeffer FT. 2008. Persistent inequality in educational attainment and its institutional context. Eur. Sociol. Rev.

24:543–65Porter AC, Gamoran A. 2002a. Methodological Advances in Cross-National Surveys of Educational Achievement.

Washington, DC: Natl. Acad. PressPorter AC, Gamoran A. 2002b. Progress and challenges for large-scale studies. See Porter & Gamoran 2002a,

pp. 3–23


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


Ramirez FO, Luo X, Schofer E, Meyer JW. 2006. Student achievement and national economic growth. Am.J. Educ. 113:1–29

Raudenbush SW, Kim J-S. 2002. Statistical issues in analysis of international comparisons of educationalachievement. See Porter & Gamoran 2002a, pp. 267–94

Rijken S. 1999. Educational expansion and status attainment. A cross-national and over-time comparison. ICS Diss.Univ. Utrecht

Rindermann H. 2007. The g-factor of international cognitive ability comparisons: the homogeneity of resultsin PISA, TIMSS, PIRLS and IQ tests across nations. Eur. J. Personal. 21:667–706

Schutz G, Ursprung H, Wossmann L. 2008. Education policy and equality of opportunity. Kyklos 61:279–308Shavit Y. 1984. Tracking and ethnicity in Israeli secondary education. Am. Sociol. Rev. 49:210–20Shavit Y. 1990. Segregation, tracking, and the educational attainment of minorities: Arabs and oriental Jews

in Israel. Am. Sociol. Rev. 55:115–26Shavit Y, Muller W, eds. 1998. From School to Work. A Comparative Study of Educational Qualifications and

Occupational Destinations. Oxford: ClarendonStanat P, Christensen G. 2006. Where Immigrant Students Succeed: A Comparative Review of Performance and

Engagement in PISA 2003. Paris: OECDStevenson DL, Baker David P. 1991. State control of the curriculum and classroom instruction. Sociol. Educ.

64:1–10Ten Dam GTM, Volman MLL. 2003. A life jacket or an art of living: inequality in social competence education.

Curric. Inq. 33:117–37Ten Dam GTM, Volman MLL. 2007. Educating for adulthood or for citizenship: social competence as an

educational goal. Eur. J. Educ. 42:281–98Terwel J. 2005. Curriculum differentiation: multiple perspectives and developments in education. J. Curric.

Stud. 37:653–70Thomas V, Wang Y, Fan X. 2001. Measuring Education Inequality: Gini Coefficients of Education. Washington,

DC: World BankThrupp M, Lauder H, Robinson T. 2002. School composition and peer effects. Int. J. Educ. Res. 37:483–504Torney-Purta J, Lehmann R, Oswald H, Schulz W. 2001. Citizenship and Education in Twenty-Eight Countries.

Civic Knowledge and Engagement at Age Fourteen. Amsterdam: Int. Assoc. Eval. Educ. Achiev.Turner RH. 1960. Sponsored and contest mobility and the school system. Am. Sociol. Rev. 25:855–67Van de Werfhorst HG, Van Tubergen F. 2007. Ethnicity, schooling, and merit in the Netherlands. Ethnicities

7:416–44Van den Eeden P, Terwel J. 1994. Evaluation of a mathematics curriculum: differential effects. Stud. Educ.

Eval. 20:457–75Vandenberge V. 2006. Achievement effectiveness and equity: the role of tracking, grade repetition and inter-

school segregation. Appl. Econ. Lett. 13:685–93Van Elk R, van der Steeg M, Webbink D. 2009. The effect of early tracking on participation in higher education.

CPB Neth. Bur. Econ. Policy Anal. No. 182, The HagueWalzer M. 1983. Spheres of Justice. New York: Basic BooksWestbury I, Hsu C-S. 1996. Structures of curriculum governance and classroom practice in mathematics.

Educ. Eval. Policy Anal. 18:123–39Willms JD. 1986. Social class segregation and its relationship to pupils’ examination results in Scotland.

Am. Sociol. Rev. 51:224–41Wolf A. 2002. Does Education Matter? Myths about Education and Economic Growth. London: PenguinWolff EN. 2006. Does Education Really Help? Skill, Work and Inequality. New York: Oxford Univ. PressWossmann L. 2003a. Central exit exams and student achievement: international evidence. In No Child Left

Behind? The Politics and Practice of Accountability, ed. PE Peterson, MR West, pp. 292–324. Washington,DC: Brookings Inst. Press

Wossmann L. 2003b. Schooling resources, educational institutions and student performance: the internationalevidence. Oxford Bull. Econ. Stat. 65:117–70

Wossmann L. 2005. The effect heterogeneity of central examinations: evidence from TIMSS, TIMSS-repeatand PISA. Educ. Econ. 13:143–69


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.


Wossmann L. 2008b. Efficiency and equity of European education and training policies. Int. Tax Public Financ.15:199–230

Wossmann L. 2008a. How equal are educational opportunities? Family background and student achievementin Europe and the US. Z. Betriebswirtsch. 78:45–70

Yamamoto K, Kulick E. 2000. Scaling methodology and procedures for the TIMSS mathematics and sciencescales. In TIMSS 1999 Tech. Rep., ed. MO Martin, KD Gregory, SE Stemler, pp. 237–63. Chestnut Hill,MA: Int. Study Cent. Boston Coll.

Zimmer RW, Toma EF. 2000. Peer effects in private and public schools across countries. J. Policy Anal. Manag.19:75–92


Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.

SO36-FM ARI 2 June 2010 4:38

Annual Reviewof Sociology

Volume 36, 2010Contents

FrontispieceJohn W. Meyer � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � xiv

Prefatory Chapter

World Society, Institutional Theories, and the ActorJohn W. Meyer � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 1

Theory and Methods

Causal Inference in Sociological ResearchMarkus Gangl � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �21

Causal Mechanisms in the Social SciencesPeter Hedstrom and Petri Ylikoski � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �49

Social Processes

A World of Standards but not a Standard World: Toward a Sociologyof Standards and StandardizationStefan Timmermans and Steven Epstein � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �69

Dynamics of Dyads in Social Networks: Assortative, Relational,and Proximity MechanismsMark T. Rivera, Sara B. Soderstrom, and Brian Uzzi � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �91

From the Sociology of Intellectuals to the Sociology of InterventionsGil Eyal and Larissa Buchholz � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 117

Social Relationships and Health Behavior Across the Life CourseDebra Umberson, Robert Crosnoe, and Corinne Reczek � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 139

Partiality of Memberships in Categories and AudiencesMichael T. Hannan � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 159

v

Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.

SO36-FM ARI 2 June 2010 4:38

Institutions and Culture

What Is Sociological about Music?William G. Roy and Timothy J. Dowd � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 183

Cultural Holes: Beyond Relationality in Social Networks and CultureMark A. Pachucki and Ronald L. Breiger � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 205

Formal Organizations

Organizational Approaches to Inequality: Inertia, Relative Power,and EnvironmentsKevin Stainback, Donald Tomaskovic-Devey, and Sheryl Skaggs � � � � � � � � � � � � � � � � � � � � � � � � 225

Political and Economic Sociology

The Contentiousness of Markets: Politics, Social Movements,and Institutional Change in MarketsBrayden G King and Nicholas A. Pearce � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 249

Conservative and Right-Wing MovementsKathleen M. Blee and Kimberly A. Creasap � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 269

The Political Consequences of Social MovementsEdwin Amenta, Neal Caren, Elizabeth Chiarello, and Yang Su � � � � � � � � � � � � � � � � � � � � � � � � � � 287

Comparative Analyses of Public Attitudes Toward Immigrantsand Immigration Using Multinational Survey Data: A Reviewof Theories and ResearchAlin M. Ceobanu and Xavier Escandell � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 309

Differentiation and Stratification

Income Inequality: New Trends and Research DirectionsLeslie McCall and Christine Percheski � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 329

Socioeconomic Disparities in Health BehaviorsFred C. Pampel, Patrick M. Krueger, and Justin T. Denney � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 349

Gender and Health InequalityJen’nan Ghazal Read and Bridget K. Gorman � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 371

Incarceration and StratificationSara Wakefield and Christopher Uggen � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 387

Achievement Inequality and the Institutional Structure of EducationalSystems: A Comparative PerspectiveHerman G. Van de Werfhorst and Jonathan J.B. Mijs � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 407

vi Contents

Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.

SO36-FM ARI 2 June 2010 4:38

Historical Studies of Social Mobility and StratificationMarco H.D. van Leeuwen and Ineke Maas � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 429

Individual and Society

Race and TrustSandra Susan Smith � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 453

Three Faces of IdentityTimothy J. Owens, Dawn T. Robinson, and Lynn Smith-Lovin � � � � � � � � � � � � � � � � � � � � � � � � � � 477

Policy

The New Homelessness RevisitedBarrett A. Lee, Kimberly A. Tyler, and James D. Wright � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 501

The Decline of Cash Welfare and Implications for Social Policyand PovertySandra K. Danziger � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 523

Indexes

Cumulative Index of Contributing Authors, Volumes 27–36 � � � � � � � � � � � � � � � � � � � � � � � � � � � 547

Cumulative Index of Chapter Titles, Volumes 27–36 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 551

Errata

An online log of corrections to Annual Review of Sociology articles may be found athttp://soc.annualreviews.org/errata.shtml

Contents vii

Ann

u. R

ev. S

ocio

l. 20

10.3

6:40

7-42

8. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by U

nive

rsite

it va

n A

mst

erda

m o

n 07

/15/

10. F

or p

erso

nal u

se o

nly.

Documents

Achievement Inequality and the Institutional Structure of ...hermanvandewerfhorst.socsci.uva.nl/ARS2010.pdf · Achievement Inequality and the Institutional Structure of Educational