This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
Author's personal copy
Social Science Research 40 (2011) 1494–1505
Contents lists available at ScienceDirect
Social Science Research
journal homepage: www.elsevier.com/locate/ssresearch
The diversity dividends of a need-blind and color-blind affirmative
action policy
Sigal Alon
Department of Sociology and Anthropology, Tel Aviv University, Ramat Aviv, Tel Aviv 69978, Israel
a r t i c l e
i n f o
Article history:
Received 14 October 2010
Revised 18 May 2011
Accepted 25 May 2011
Available online 6 June 2011
Keywords:
Higher education
Affirmative action
Diversity
Admission regimes
Disadvantaged populations
a b s t r a c t
In the early to mid-2000s, four flagship Israeli selective universities incorporated a needblind and color-blind affirmative action policy into their admissions practices. The program,
which gives an edge in admission to academically borderline applicants from disadvantaged backgrounds, emphasizes structural disadvantages, such as neighborhood socioeconomic status and high school rigor. The results of this study, based on administrative data
from the four universities, demonstrate that having such a policy in place made the four
institutions, especially the echelons at the most selective departments, more diverse than
they otherwise would have been. The rise in geographic, economic and demographic diversity of a student population suggests that the plan’s focus on structural determinants of
disadvantage yields broad diversity dividends. The paper discusses the relevance of the
findings to the ongoing discussion of admission regimes, diversity and equality of opportunity in the US.
Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction
The continuing public discontent with race-conscious admissions in the US – echoing similar claims in India, Brazil, South
Africa and other countries – and the ban on such preferences in several states have motivated the search for race-neutral
admission alternatives. The past decade brought the implementation of percent plans in Texas, California and Florida, alongside growing interest in propositions based on class preference. However, the shortcomings of these propounded alternatives
(Bowen et al., 2005; Espenshade and Radford, 2009; Kane, 1998; Kurlaender and Grodsky, 2010; Tienda, 2010; The US Commission on Civil Rights, 2002), together with the rising class inequality in higher education in recent decades (Alon, 2009),
suggest that the quest for innovative race-neutral ways to increase demographic, socioeconomic and geographical diversity
at selective postsecondary institutions is not over.
The current study contributes to the rich scholarship on admission regimes and preferences by evaluating a unique policy
of class-based affirmative action used at four flagship Israeli universities since the early to mid-2000s. The program promotes preferential treatment of academically borderline applicants from disadvantaged backgrounds. The most theoretically
attractive aspect of this policy, however, is its distinctive design: neither the students’ financial constraints nor their national
or ethnic origins are considered. The emphasis, rather, is on students’ structural disadvantages, such as neighborhood socioeconomic status and high school rigor.
In Israel, like in other countries, socioeconomic and national/ethnic origin inequalities are intertwined with spatial
inequality. Consequently, this program serves as a natural experiment that tests the extent of this entanglement. Specifically,
it demonstrates whether and how much a need-blind and color-blind affirmative action policy can increase geographic, economic and demographic diversity within a student population, and whether the diversification effects are uniform across
E-mail address: [email protected]
0049-089X/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.ssresearch.2011.05.005
Author's personal copy
S. Alon / Social Science Research 40 (2011) 1494–1505
1495
different levels of selectivity. The conclusions reached here about the effectiveness of an existing program of class-based affirmative action, and the scope of its dividends in terms of diversity, have far-reaching significance for the design of preferential
treatment policies, especially in light of the controversy concerning admission practices in the US. The universal relevance of
these findings is considered, while taking into account the differences between the US and Israeli postsecondary systems, as
well as in the judicial standing of preferential treatment policies.
2. Diversity and alternative admission models
Diversification of selective colleges and universities is necessary because they are bastions of privilege (Alon, 2009; Alon
and Tienda, 2007; Bowen and Bok, 1998; Bowen et al., 2005; Karabel, 2005; The National Association for College Admission
Counseling (NACAC), 2006). US Justice Powell, the pivotal opinion in the Bakke case, outlined the educational foundation for
the ‘‘diversity rationale.’’ He argued that a diverse student body broadens the range of viewpoints collectively held by those
students and subsequently, allows a university to provide an atmosphere that is ‘‘conducive to speculation, experiment and
creation – so essential to the quality of higher education.’’ Since then, ample research has demonstrated the educational benefits of campus diversity (see Milem et al., 2005). A rise in student body diversity is associated with better learning experiences for all students and more opportunities for interaction with students from different backgrounds. This increases
students’ tolerance to a wide range of viewpoints and enhances their cognitive and identity development, preparing them
for better participation in a democratic society.
Justice Powell concluded that race-conscious admissions practices, when narrowly tailored, serve a compelling educational interest. The US Supreme Court ruling in 2003, in the two cases regarding the University of Michigan, enacted this
rationale but made it clear that affirmative action should be a temporary remedy. The practice of affirmative action in
higher education admissions was thus deemed permissible because it is believed to yield educational benefits by assembling a student body with diverse talents and perspectives (Gurin et al., 2002). Yet, despite evidence of the demographic
diversification effects of race-conscious admissions on elite campuses (Alon and Tienda, 2005, 2007; Bowen and Bok,
1998), there is growing opposition to these practices. One of the main criticisms of the preferential treatment of racial
and ethnic minorities in admissions is that they promote applicants who are ‘‘not deserving’’ because they take into account group identity rather than individual circumstances (Clayton and Tangri, 1989). The rising discontent has led to public referenda against the use of race-sensitive admissions, such as Proposition 209 in California, Initiative I-200 in
Washington State, and, most recently, Proposal 2 in Michigan and Arizona Proposal 107. It has also resulted in judicial
bans on racial preferences in higher education admissions in Texas and Florida, as well as numerous lawsuits against affirmative action practices in higher education (Gratz v. Bollinger (2003) and Grutter v. Bollinger (2003) are the most
notable).
States that have banned race-conscious admission practices have since undertaken alternative approaches. Texas, Florida
and California, for example, have adopted various forms of percent plans, programs that admit a fixed percentage of each
school’s class rank distribution. The percent plans were designed to broaden college access across economic, demographic,
geographic and social groups in the absence of race-based preferences. The cumulative evidence from Texas, California, and
Florida indicates that a uniform admission regime cannot increase ethno-racial diversity, even in predominantly minority
and residentially segregated states, without race-sensitive outreach to students at high schools with low college-going rates
and without generous financial support (Grodsky and Kurlaender, 2010; Tienda, 2010; The US Commission on Civil Rights,
2002).
Under the Texas H.B. 588 law, for example, applicants who graduate in the top decile of their senior class are guaranteed
admission to the public postsecondary school of their choice regardless of their standardized test scores. Although the enactment of this Plan in Texas was followed by an increase in racial and ethnic diversity in the state’s flagship institutions, Harris
and Tienda (2010) show that the effect was the result of changes in the size and composition of high school graduation cohorts rather than the policy. Moreover, while the plan was successful in broadening geographic diversity (Long et al., 2010), it
failed to augment socioeconomic diversity (Koffman and Tienda, 2008; Tienda et al., 2010).1 Another problem with the uniform admission regime that Texas put in place is that the University of Texas at Austin (UT-Austin) became saturated with
applicants eligible for automatic admission, thus limiting the university’s ability to craft a class and achieve its institutional
goals (Tienda, 2010).2 Moreover, critics of the law, mostly suburban whites, continue to argue that the plan gives unjust advantage to ‘‘undeserving’’ applicants, claiming that low-ranked students from academically rigorous high schools are often more
qualified than those from low performing high schools who are guaranteed college admission. Thus, despite the initial appeal
of a uniform admission regime, the Texas Top 10% policy has become just as controversial as the race-conscious system it replaced (Tienda, 2010).
As the voices objecting to racial preferences became louder in recent years, another race-neutral alternative to diversifying the collective college student body has been proposed: preference plans based on socioeconomic class, either in addition
1
Likewise, the ban on the consideration of race in admissions in California resulted in a decline in minority enrollment at the most selective University of
California campuses and at some of the more selective California State University schools (Grodsky and Kurlaender, 2010). The talented 20 plan in Florida did
not produce racial and ethnic diversity at the two most selective campuses in the state’s university system, the University of Florida (UF) and Florida State
University (FSU), since it benefitted white and Asian students more than blacks and Hispanics (The US Commission on Civil Rights, 2002).
2
After the Top 10% bill was revised by the Texas legislature in 2009, UT-Austin was required to fill only 75% of its freshman slots with eligible seniors.
Author's personal copy
1496
S. Alon / Social Science Research 40 (2011) 1494–1505
to or as an alternative to race-conscious admissions (Bowen et al., 2005; Kahlenberg, 1996). Since the poor of all races are
seen as more deserving than blacks (Skrentny, 1996) the concept of class-based preference is compelling and enjoy strong
public support (Carnevale and Rose, 2004). Moreover, giving an edge in admission to applicants from modest backgrounds to
the most selective institutions is expected to advance socioeconomic diversity and social mobility. Nonetheless, class-based
preferential policies remain theoretical in the US. Since they have never been implemented (except for sporadic experiments), the evidence on their consequences is based on statistical simulations rather than real initiatives (Bowen et al.,
2005; Espenshade and Radford, 2009; Kane, 1998). The statistical simulations suggest that class-based preferences cannot
replace race-sensitive admissions because income and parental educational attainment are not good substitutes for race (Bowen et al., 2005; Espenshade and Radford, 2009; Kane, 1998). In addition, whereas policies based on racial preferences include the pool of socioeconomically strong minorities, class-based policies are limited, it is argued, by the small pool of
qualified applicants from socioeconomically disadvantaged backgrounds.
Noteworthy is the fact that these simulations use a narrow definition of ‘‘class,’’ based solely on parental education and
income. Kahlenberg (1996), however, in his thorough conceptualization of class preferences, maintains that the fairest and
most apt method for implementing class-based affirmative action should take into account household net worth, the quality
of secondary education, neighborhood influences, and family structure; it is also the most reliable option because the abundance of information required lowers the chances of abusing the system.3 Yet it seems that this multidimensionality, however
justified, complicates the implementation of class preferences tremendously. Administering such a comprehensive plan demands the taxing tasks of determining the relevant indicators, and then collecting, verifying and weighting a wide array of sensitive information.4 Consequently, such a design for class-based preferential policy may not prove practical.
The Israeli affirmative action plan at elite institutions, however, takes a different route for determining class: it targets
individuals from disadvantaged neighborhoods and high schools. The theoretical foundation of this design is rooted in the
long sociological tradition that highlights the effects of social structures, such as neighborhoods and schools, on youth
achievements and educational outcomes (Ainsworth, 2002; Brooks-Gunn et al., 1997; Massey and Denton, 1993; McLeod
and Edwards, 1995; Pebley and Sastry, 2004; Sampson et al., 1997, 1999; South and Crowder, 1999; Wilson, 1987, 1996).
This neighborhood/school-based need-blind and color-blind affirmative action policy capitalizes on the overlap between spatial boundaries and categorical inequality. That is, bad neighborhoods and failing schools are populated with categorically
disadvantaged groups, such as racial and ethnic minorities, recent immigrants, and the poor (Blau, 1977; Massey, 2007).
Thus, to the extent that there is sufficient overlap between systems of inequality, this scheme has the potential to yield
wide-ranging diversity dividends. The next section provides a description of the guidelines and operation of the plan.
3. A need- and color-blind affirmative action policy at selective Israeli universities
The higher education system in Israel is mostly public. As of 2007/2008, there were 75 institutions in two tiers. In the first
tier, there are six research universities. These universities generally rely on a formulaic selection process for bachelor’s
admissions based entirely on an academic index, calculated by taking a weighted mean of an individual’s matriculation diploma grades and psychometric test score (similar to an SAT score).5 The admission cutoff points are not institution-wide
but rather major-specific; thus, within each institution there are departments that vary in their selectivity level.
Even after the spectacular expansion of the Israeli higher education system in 1995,6 underprivileged populations remained under-represented in most of the first-tier universities. In order to address the issues of diversity and access, the four
most selective and internationally recognized universities – Tel-Aviv University, The Hebrew University, Ben-Gurion University
and The Technion – incorporated a comprehensive and standardized program of class-based affirmative action into their admission practices in the early to mid-2000s.7
The implementation of the policy involves three stages. First, socioeconomic eligibility is established outside of the universities. Applicants to all four universities are given the option to complete and submit a standard application for preferential
treatment (substantiated with documents). The form is examined by a centralized nonprofit organization that weights the
information and creates an index of socioeconomic disadvantage, and in turn, reports each applicant’s score to the universities. The score ranges from 0 to 85 and the threshold for socioeconomic eligibility used by the universities is 30 points.
Eligibility is determined according to three parameters, all of which are based on an applicant’s high school years: the structure of opportunity (neighborhood and high school attended); family socioeconomic status (parental education and family
3
Kahlenberg (1996) suggests three methods for implementing class-based affirmative action. The simplest is based on family income alone, while the second
includes the three factors at the basis of the common definition of socioeconomic status: parental income, education, and occupation. The third, and most
sophisticated, alternative takes into account individual as well as structural determinants of inequality.
4
This was demonstrated by the year-long experience of UCLA’s School of Law with class-based admission preferences in 1996 (Sander, 1997).
5
The second tier consists of degree-granting non-research academic colleges, and specialized institutions or branches of foreign universities. At the academic
colleges, admission is based solely on the matriculation diploma, and is therefore substantially less selective than admission to the universities.
6
The total number of undergraduate students tripled from 58,000 in 1990/1991 to 168,000 in 2007/2008, mostly due to the addition of 2nd tier instiutions.
The number of undergraduate students attending the six universities also rose (from 49,000 to 67,000, respectively).
7
The other two institutions, Bar-Ilan University and Haifa University, did not adopt such a program. Both schools are less selective than the other four
universities.
Author's personal copy
S. Alon / Social Science Research 40 (2011) 1494–1505
1497
size); and individual and/or family adverse circumstances.8 The weighting algorithm, not known to the applicants, makes it
almost impossible to pass the 30-point threshold without a structural disadvantage since the two structural factors account
for about 50% of the eligibility score. Thus, applicants from a poor community who attended the only high school in the vicinity
automatically pass the threshold for socioeconomic eligibility. Notably, information about an applicant’s national or ethnic origin and financial situation is neither requested nor considered.
In the second stage, academic eligibility is determined for those who passed the threshold for socioeconomic eligibility.
The eligible pool consists of applicants with borderline achievements, those whose academic index score is around 0.5–
1.0 standard deviations below the major-specific cutoff point of admissibility. The third, and final, stage is the admission decision. It is important to note here that the admission process is non-mechanistic: that is, admission is not guaranteed to applicants that pass the first two stages of eligibility and, as such, the departments have the discretion to pick and choose their
students.9 To prevent the saturation of certain majors with AA-eligible admits, the cap on the share of affirmative action admits
in each department is 5% of each department’s entering class.
The focus on structural inequality facilitates the administration and implementation of the program because an applicant’s place of residence and high school are easily available on public record. It also expands the pool of qualified applicants,
which is one of the key problems cited in the literature of class-based affirmative action. Moreover, by eliminating the need
to verify an applicant’s financial standing, the program is less invasive of privacy, more reliable, and reduces the likelihood of
manipulation. Yet, feasibility is not a goal in itself; the main issue is whether targeting structural determinants of disadvantage yields geographic, demographic and economic diversity dividends. The answer to this question will depend on the level
of overlap between spatial boundaries and categorical inequality in Israel.
4. The overlap between systems of inequality in Israel
Due to the ethnic and socioeconomic diversity of its population, Israel provides the perfect setting for studying categorical
inequality. There are two main demographic cleavages in Israel. The first is along national lines, between Israeli Jews and
Arabs. Seventy-six percent of Israel’s population is comprised of Jews whereas 20% consists of Arabs, the majority of which
are Muslims. The second cleavage exists along ethnic lines within the Jewish population, between Jews of European and
American origin and Jews of Asian and African origin. Immediately after the establishment of the State of Israel in 1948,
the Jewish population of the state grew several folds from numerous and massive waves of immigration from both America
and Europe (mostly Holocaust survivors), and Asia and Africa. In recent years, the most significant immigration wave totaled
one million Jews, most of whom arrived from the Former Soviet Union between 1989 and 1995.10 Yet as time lapses from the
major immigration waves of the 1940s and the 1950s, the share of first- and second-generation immigrants in the population is
declining, especially among the college age population.
Differences in national and ethnic origin shape Israel’s stratification system. There is a clear hierarchy, which has persisted over the past 50 years, in the levels of educational attainment, occupational status, and earnings: Jews of European
and American descent are at the top of the socioeconomic ladder followed by Jews of Asian and African origin, while the Arab
citizens of Israel occupy the bottom echelons of the socioeconomic hierarchy (Cohen and Haberfeld, 1998; Haberfeld and
Cohen, 1998, 2007; Lewin-Epstein and Semyonov, 1986, 1992; Semyonov and Tyree, 1981; Yaish, 2001). In Israel, like in
other countries, this categorical inequality is coupled with spatial inequality, creating an important contour of stratification.
Israel’s development towns are the paragon of the overlap of systems of segregation (Alon, 2004a; Spilerman and Habib,
1976; Yiftachel, 2000). The Israeli government established these geographically peripheral towns in order to absorb, in particular, the massive influx of Jews from Asia and Africa during the 1950s (Khazzoom, 2005). In 2006, more than 90% of these
localities ranked in the bottom half of the localities’ socioeconomic index (ICBS, 2006), a clear indication that the disadvantaged status of these towns has not changed much over the decades.11 Their unbalanced demography has also persisted: more
than half of the residents in these towns in 1995 were of Asian/African origin (first and second generation) as opposed to about a
third in the general population (ICBS, 1995). Such coupling also occurs in the Arab localities, deprived areas with high shares of
poor, unskilled residents. In 2006, 80% of the independent Arab villages and towns ranked in the bottom third of the socioeconomic index (ICBS, 2006). Moreover, most of the Arab population lives in the geographical periphery (about half resides in the
northern part of Israel), far from the main metropolitan areas of core economic activity. The spatial segregation of recent immigrants, who are mostly from the Former Soviet Union, is less pronounced because they, unlike the immigrants before them,
were free to choose where to live. In fact, as of 1995, there is an overrepresentation of recent immigrants in strong localities
(ICBS, 1995). Nevertheless, these new immigrants may still benefit from the affirmative action plan since it is likely that most
8
‘Adverse circumstances’ applies if the applicant is an orphan, an immigrant, divorced, a single parent, suffers from a health disability, experienced the death
of a sibling, or has divorced parents. The definition also includes any applicant that has a parent with a disability or chronic illness.
9
All the applicants are grouped into one pool; the only difference is that eligible students may have a small edge over non-eligible applicants who are
similarly academically situated. Moreover, the exact socioeconomic eligibility score of applicants above the 30-point threshold has no bearing on the admission
decision.
10
This wave represents 15% of the population. There were also two small waves of immigrants from Ethiopia during this period.
11
The index ranges from 1 to 10. Localities are classified according to the socioeconomic level of the population, which is determined by the financial
resources of the residents, housing, standard of living, schooling and education, employment profile, various types of socio-economic distress, and demographic
characteristics.
Author's personal copy
1498
S. Alon / Social Science Research 40 (2011) 1494–1505
do not reside in the affluent parts of these localities, and because recent immigration status is also considered an indicator of
socioeconomic eligibility.
The implications of the spatial stratification of Israel for its postsecondary education system are blatant: in 2006, the
share of university students from affluent localities (ranked in the top third of the socioeconomic index) was about 28%,
but only 6% of students were from localities in the bottom third (ICBS, 2006). Consequently, a policy that spotlights spatial
and school inequality should, theoretically, stretch the diversity dividends to economic status and national and ethnic origin.
Thus, Arabs, Jews of Asian or African origin, and the poor are likely overrepresented in the affirmative action-eligible pool of
students compared to the regular pool. Recent immigrants may also be overrepresented due to their incomplete economic
and social assimilation in Israel. This hypothesis, however, demands empirical testing because even in the development
towns and Arab localities where segregation is uttermost, not all residents are underprivileged or belong to an ethnic or national minority.
This investigation sheds light on the effect of the Israeli affirmative action initiative on the contours of diversification in
selective university campuses. Given that applications and admissions to the universities are major-specific, it is critical to
not only examine the magnitude of the plan’s diversity dividends, but also their distribution across selectivity levels. Hence,
the study asks: Where does the plan have its greatest diversifying influence? Is it on the more selective programs, where
demand and academic requirements are high and, consequently, where disadvantaged groups are most under-represented?
Or, alternatively, is the plan most effective among the least selective majors, where the under-representation of groups is
smallest?
5. Data and methods
5.1. Database and sample
I obtained institutional administrative data from four chief Israeli universities—Tel-Aviv University (TAU), The Hebrew
University (HUJI), The Technion (TEC), and Ben-Gurion University (BGU)—for periods ranging from 10 to 12 consecutive
years (ca. 1997–2008). During this period, the fields of study at the four universities received 1,135,286 applications from
348,869 applicants (each applicant can apply to multiple majors or apply in different years). Of the applicants, 60% were
admitted to at least one of the fields they applied to. Ultimately, 163,610 students enrolled at the four institutions during
this period.
The preferential policy began in 2001 at HUJI; 2003 at TAU;12 2004 at TEC; and 2006 at BGU. Consequently, the data yields
8 years of admission observations under the AA regime at HUJI, six at TAU, five at TEC, and three at BGU.13 The analytical sample
contains around 180,000 observations of applicants and 72,000 of students who enrolled in the above four universities after the
affirmative action plan went into effect (AA regime). For the sake of parsimony, I report aggregated results and do not dwell on
between-institution variations.
5.2. Selectivity
To understand the choices and destinations of the AA applicants and admits, it is critical to consider between-major and
between-institution differences in the selectivity of admissions. Since Israeli applicants apply to specific majors within each
institution and, thus, the admission decisions are major-institution-specific, the basic unit for the measure of selectivity is
the department. Altogether, the four institutions contained about 170 departments, representing, approximately, 50 different
general fields of study. For each department I calculated two selectivity indicators: first, competitiveness, defined as the
department admission rate (44% on average for the period between 1997 and 2008); and, second, academic rigor, determined
by the mean test score of admits (the overall 1997–2008 average is 636; to put this in perspective, the national average of all
test-takers in 2009 was 564).14 Next, I created an index of selectivity by calculating the sum of the standardized scores of the
admission rate and mean test score of each department (ranges from 4.26 to 3.44). The departments were then classified into
selectivity quintiles. Table 1 presents general statistics regarding this classification.
The results underscore the selective and competitive nature of most departments at the four flagship Israeli universities.
Classified as most selective are departments that, on average, admitted 23% of applicants and whose mean test score was 698.
The results in the bottom panel of Table 1 indicate that 43% of first-time applicants in 1997–2008 sent an application to these
12
TAU’s faculties of law and social sciences have experimented with the affirmative action program since 1997. However, due to the small scale of this local
initiative (168 students), I deleted these observations and do not consider these years under the AA regime.
13
The analyses are limited to students under age 28 due to the different admissions criteria for older students and the fact that the affirmative action plan has
the same age cap. Also excluded are 230 students who benefited from a relatively new program at TAU, which does not require the submission of test scores for
applicants from disadvantaged schools who ranked at the top of their class. In addition, 290 students of Ethiopian origin who benefit from other types of
preferential treatment are omitted from the analyses. International students and students enrolled in various programs that combine military service and
academic studies are also not included in the sample.
14
These two measures do not completely overlap (correlation of 0.56). For instance, there are departments with very low admission rates but that draw
from an applicant pool with relatively low academic credentials (e.g., nursing) and vice versa (e.g., physics).
Author's personal copy
1499
S. Alon / Social Science Research 40 (2011) 1494–1505
Table 1
The selectivity index for academic departments (1997–2008).
Panel A. Admission rate and mean test scores
Selectivity level
Average admission rate
Mean test scores
Top quintile (33 departments)
2 quintile (33)
3 quintile (32)
4 quintile (33)
Bottom quintile (33)
23
35
38
54
72
698
673
636
602
573
Total
44
636
Panel B. The distribution of first-time applicants, admits, and students across selectivity level
Selectivity level
Applicants
Share of admits
Students
Top quintile
2 quintile
3 quintile
4 quintile
Bottom quintile
43
25
16
12
4
19
24
23
23
11
20
24
23
23
10
N
370430
216632
158155
Note. Selectivity level is missing for about 1% of the applicants.
majors.15 Among the most selective departments are the vocational/professional fields, many of which are in the medical sciences16 or in the engineering/computer sciences (but also architecture and psychology). It is important to note that these most
selective professional programs are longer (between 4 and 6 years) than the typical 3-year bachelor’s programs. The bottom
quintile has an admission rate of 72%, on average, and mean test score of 573 (similar to the national average test score); only
4% of applicants in 1997–2008 sought admission to these departments. All departments classified as least selective are the
humanities.
5.3. Variables
Appendix A provides definitions for all variables.
AA eligibility. This is a dichotomous outcome that indicates whether a student applied for preferential treatment and
passed the 30-point threshold.
Locality. I identified several theoretically relevant indicators based on student locality of residence. They are: whether the
locality is classified as a development town; its ranking on the localities socioeconomic index; and its geographic region.17
Origin. I created five mutually exclusive indicators for student origin: (a) second generation Asia–Africa: native-born Israeli Jews with at least one parent from Asia or Africa, making up 12% of the students in the AA regime; (b) second generation
Europe–America: native-born Israeli Jews with both parents from Europe or America, or only one parent from Europe or
America if the other parent is a native-born Israeli (15%); (c) Israel: third-generation immigrants, that is, native-born Israeli
Jews with two native-born parents (36%); (d) first-generation Jewish immigrants from Asia–Africa or Europe–America (17%);
and (e) Arab (8%). Twelve percent of the sample lacked information about their origin, mainly because applicants in one of
the institutions were not asked about their parents’ place of birth. Yet, since I was able to identify Arabs and the origin of
first-generation immigrants in this institution, I decided to include its observations in the analytical sample, but report
the share of missing data in the origin variable. Consequently, the results may underestimate the share of second-generation
students.
Economic need. This indicator is based on students’ receipt of institutional need-based grants during their first year of university studies (there are no direct aid provisions by the state). This is a relative measure of pecuniary constraints: within
each institution in a given year, the allocation of grants is solely need-based, and, as a result, distinguishes between the more
and less needy students. About 18% of the students in the AA regime received need-based financial aid. Because this information comes from administrative records and is not self-reported, it is an attractive measure of student’s financial constraints. It is worth emphasizing that in Israel, unlike in the US, decisions regarding eligibility for affirmative action and
financial aid are orthogonal: the former is strictly need-blind and made by an external non-profit organization, while the
latter is made by universities only after a student has been admitted.18 Also noteworthy is the low and uniform tuition at
15
Since applicants can apply to multiple majors and students can have dual majors, each individual is classified according to the most selective major she
applied to, was admitted to, and enrolled in.
16
Included are medicine, dentistry, pharmacology, occupational therapy, communication disabilities, and physiotherapy.
17
The database does not contain neighborhood- and school-level information.
18
Racial preference policies in the US expand educational opportunities for under-represented minorities not only because of the edge that they provide in
admissions, but also because they imply better financial aid packages for minorities (Alon, 2001, 2004b, 2010).
Author's personal copy
1500
S. Alon / Social Science Research 40 (2011) 1494–1505
Table 2
First-time applicants, admits and students at four Israeli universities by AA status, AA regime.
Total
N
N
N
N
applicants
applicants: SES score P 30
admits
students
178713
178713
101703
72191
AA status
Share of AA request (%)
Regular
AA
167785
173105
98630
69540
10928
5608
3073
2651
Share of AA-eligible (%)
6.1
3.1
3.0
3.7
Note. The AA plan started in 2001 in HUJI; 2003 in TAU; 2004 in TEC; and 2006 in BGU.
all Israeli public institutions,19 which renders a unique opportunity to measure the non-pecuniary effect of affirmative action
initiatives on economic diversity.
5.4. Analytical strategy
In order to estimate the program’s diversification effect on the overall composition of the student body, and the distribution of this effect across levels of selectivity, I compare the diversity of the AA-eligible and the regular student pools during
the AA regime.20 The diversity in the regular pool captures the counterfactual diversity level, that is, the level of diversity that
would exist if the affirmative action plan were not in place. The analyses take into account the two factors that determine the
magnitude of the overall diversification effect on student body composition: first, the prevalence of AA-eligible students in the
relevant student body; and second, the dissimilarity between the AA-eligible and regular pools in terms of key characteristics.
6. Results
6.1. The prevalence of AA-eligible students in the student body
The results in Table 2 show that since the affirmative action plan began, there were 179,000 applicants to the four universities. Six percent of applicants sought preferential treatment (10,928 applicants had a score on the socioeconomic index),
but only half of them met the 30-point threshold for eligibility (3% of applicants). About 55% of AA-eligible applicants were
admitted compared to 57% of the regular applicants. Eventually, in the relevant years, 3.7% of the student body was AA-eligible (N = 2651).
Yet the interesting question is: how do the two groups compare in their field of study preferences as applicants (in terms
of department selectivity), and their ultimate destinations as students? An examination of the distribution of the two pools
(AA-eligible and regular) across quintiles of departmental selectivity (Table 3) shows that the major choice sets of AA-eligible
and regular applicants are quite similar. One in two of the AA-eligible applicants included a most selective department in
their application (comprising 3.4% of the applicant pool in this tier), while less than 2% chose a least selective major (comprising 1.6% of the applicant pool in this tier). The widespread attempts of this group to access the more selective fields of
study indicate its high ambition. However, more AA-eligible applicants are turned down from these departments than are
regular ones. Eventually, the share of AA-eligible admits at the most (and least) selective departments is the smallest (only
2.4%) while their share is largest among admits to the second, third and fourth quintiles of selectivity (3.3–3.5%). Given a
higher yield rate among the AA admits at the most selective departments compared to regular admits (plausibly another sign
of their ambition), their final share at this tier rises to 3.2%. Their share is the highest at the forth quintile (4.4%). In sum, the
preferential policy has the potential to enhance diversity in all selectivity echelons, depending, of course, on the dissimilarities in key attributes between students admitted without preferential treatment and those who received an edge. The following analysis assesses these dissimilarities.
6.2. Dissimilarity between pools in key characteristics and the plan’s diversity dividends
Table 4 provides evidence regarding differences in spatial location, ethnic origin and economic status between the groups.
The results capture a remarkable effect of spatial diversity, yet it should be noted that the real spatial diversity is probably
larger since the data do not have neighborhood-level information, which is the level targeted by the plan. Nine percent of the
AA-eligible students came from development towns, compared to only 2% in the regular pool. Overall, this represents an in19
The tuition in 2009, for example, was 9351 NIS, which is approximately $2500.
Since longitudinal data is available, a before and after comparison seems theoretically attractive for assessing the plan’s diversity dividends. However, the
temporal trends in group representation during the period of investigation undermine the feasibility of such a strategy. Most notable is the sharp decline in the
share of first- and second-generation immigrants within the student body (either from Asia–Africa or Europe–America) and the increase in the share of
students with native-born parents. This reflects a natural demographic trend, resulting from the lapsed time from the major immigration waves to Israel.
20
Author's personal copy
1501
S. Alon / Social Science Research 40 (2011) 1494–1505
Table 3
The distribution of first-time applicants, admits, and students across selectivity level, by AA status, AA regime.
Selectivity level
AA (%)
Regular (%)
Share of AA-eligible
Applicants
Top quintile (33 majors)
2 quintile (33)
3 quintile (32)
4 quintile (33)
Bottom quintile (33)
49.3
26.4
13.5
9.2
1.6
45.3
26.1
14
11.4
3.2
3.4
3.2
3.1
2.6
1.6
Admits
Top quintile
2 quintile
3 quintile
4 quintile
Bottom quintile
15.6
26.5
23.7
27.1
7.2
20.4
24.7
21.9
23.8
9.2
2.4
3.3
3.3
3.5
2.4
Students
Top quintile
2 quintile
3 quintile
4 quintile
Bottom quintile
19.1
26.2
21.1
27.4
6.2
21.9
25.1
21.4
22.9
8.8
3.2
3.8
3.6
4.4
2.6
Note. Selectivity level is missing for about 1% of the applicants.
Table 4
Characteristics of students at the four institutions by status of AA-eligibility, AA regime.
Entire student body
Locality characteristicsa
Development town
Locality SES cluster
Bottom
Middle
Top
Geographic region
Jerusalem
North
Haifa
Center
Tel-Aviv
South
Origin
Asia–Africa 2nd gen.b
Europe–America 2nd gen.b
Israelb
New immigrants – Total
Asia–Africa 1st gen.
Europe–America 1st gen.
Missing
Arab
Missing
Economic need
Need-based financial aid recipient
N
AA (%)
REG (%)
Total (%)
Growth ratec
(percentage points)
9.1
2.3
2.5
9
12.3
73.8
13.9
100%
5.4
64.6
30
100%
5.6
64.9
29.5
100%
4
0
19
22.3
8.1
17.7
15
18.1
100%
14
9.5
12.2
29
22.8
12.5
100%
14.2
9.9
12.1
28.6
22.5
12.7
100%
1
4
21.3
6.8
23.7
19.7
4.6
13.7
1.4
25.2
3.4
100%
11.4
15.1
36.2
17
1.4
12.6
3.1
7.8
12.5
100%
11.8
14.8
35.7
17.1
1.5
12.6
3
8.4
12.2
100%
4
50.3
2651
18.2
69,540
19.4
72,191
7
Index of dissimilarity
0.28
2
0.23
1
1
1
2
0.27
2
1
1
7
0
8
a
Information on locality SES cluster and geographic region is missing for 25% and 17% of the students, respectively, but there
are no differences between AA-eligible and regular students.
b
Based on data from three institutions.
c
The growth rate is calculated as [(%group_x TOT %group_x REG)/%group_x REG] * 100.
Author's personal copy
1502
S. Alon / Social Science Research 40 (2011) 1494–1505
Table 5
The diversity dividends of the preferential policy by selectivity quintiles, AA regime.
Departmental selectivity tiers
Top tier (top quintile)
Growth rate
(percentage points)
Development town
Locality SES cluster
Bottom
Geographic region
North
South
Origin
Asia–Africa 2nd gen
Asia–Africa 1st gen
Arab
Economic need
Need-based financial aid recipient
Index of
dissimilarity
15
Second tier (2 + 3 quintiles)
Bottom tier (4 + 5 quintiles)
Growth rate
(percentage points)
Growth rate
(percentage points)
Index of
dissimilarity
11
0.31
6
9
0.27
6
0.27
5
4
0.25
2
0.25
5
1
0.34
Index of
dissimilarity
0.17
4
1
0.25
0.29
3
11
8
4
0
9
2
10
9
9
6
6
crease of 9 percentage points in the share of students from development towns in the student body.21 Similarly, the share of
students from localities ranked in the bottom third of the localities SES index among the AA-eligible pool is more than double
that in the regular pool (12% vs. 5%, respectively), indicating an increase of 4 percentage points. The comparisons of the other
localities are less straightforward because not all the neighborhoods and schools in those localities are classified as disadvantaged. Even so, AA-eligible students were more likely to hail from localities in the middle third of the localities SES index, and
less likely to be residents of strong localities (in the top third), than students admitted without preferences. I also calculated the
index of dissimilarity, which measures the evenness of the distributions of the AA and regular students. The index of dissimilarity
for the locality’s SES cluster is 0.28, implying that about a third of one of the two groups would have to move to different localities in order to produce a distribution that matches that of the other group. The index of dissimilarity for geographic areas is
0.23; most of the difference in geographic diversity stems from the high share of AA students from remote areas (most gains due
to students from the northern part of Israel with some gains from the south).
Table 4 also juxtaposes the ethnic composition of the two groups. The index of dissimilarity in origin is 0.27, most of
which stems from the higher share of students classified as first- or second-generation Asia–Africa or Arabs in the AA-eligible
pool. About 21% of the AA-eligible students are Israeli-born of Asian or African origin, while only 11% of the regular students
are (growth rate of 4 percentage points). Conversely, the share of regular admission students of European or American descent is more than double the share among the AA-eligible students (15% vs. 7%). The program also infuses the student body
with Arab students: 25% of the students who receive preferential treatment are Arabs, despite comprising only 8% of the general student body. Furthermore, the AA-eligible pool contains a higher share of new immigrants than the regular pool; most
of this gain is due to immigrants of Asian or African origin. Given the color-blind nature of the preferential policy this diversification effect is quite remarkable. Finally, the strictly need-blind plan also taps into students with economic constraints:
50% of the AA-eligible students received need-based grants in their freshman year compared to just 18% of the regular students (growth rate of 7 percentage points).
One of the most challenging aspects of any preferential policy is its ability to diversify bastions of privilege. To examine
whether the diversification effect reaches the most selective departments, I replicated the analyses presented in Table 4 for
different selectivity tiers. For the sake of parsimony, Table 5 only reports the two indices of diversification (the growth rate
and the index of dissimilarity) for three tiers of selectivity: the top tier (top quintile), the second tier (second and third
quintiles), and the bottom tier (fourth and bottom quintiles).22 The results demonstrate that the plan’s diversity dividends
are the largest in the most selective departments because the baseline share of disadvantaged groups in the top tier is very
low. For example, in the top tier, the share of students from development towns in the regular pool is lowest (1.3%, result not
shown), as expected, but the growth rate of such students is the largest (15 percentage points compared to 11 and 9 in the
middle and bottom tiers). Likewise, the indices of dissimilarity for the distributions of localities and origin are the largest in
the top tier. The growth rate of students receiving financial aid is also higher in the top tier (9 percentage points compare to 6
points in the other selectivity tier). In sum, the Israeli color- and need-blind preferential policy infuses the student body of the
four flagship universities, including the bastions of privilege within them, with under-represented and disadvantaged
populations.
21
The growth rate of under-represented groups is calculated as follows: ((group share in the student body – group share without AA students)/group share
without AA students) 100.
22
The detailed results are available from the author upon request.
Author's personal copy
S. Alon / Social Science Research 40 (2011) 1494–1505
1503
7. Conclusions
The Israeli policy for preferential treatment is a fascinating hybrid of race-conscious admissions, class-based preferences,
and percent plans. It is class-based, but, like race-conscious admissions and percent plans, stresses group affiliation rather
than individual traits. The Israeli design shares its focus on spatial segregation and school inequality with percent plans
but targets these aspects more directly. Moreover, as in other forms of affirmative action, admission is not guaranteed; thus,
institutions enjoy discretion in crafting a class and the issue of institutional saturation, a major weakness of some forms of
percent plans, is circumvented.
The results of this study demonstrate that a need- and color-blind affirmative action policy can increase the geographic,
economic and demographic diversity of a student population. Having such a policy in place at four selective Israeli universities made the entire student body, even the echelons at the most selective departments, more diverse than they otherwise
would have been. All in all, the plan’s focus on structural determinants of disadvantage, such as geographic and school segregation, yields broad diversity dividends.
These insights can enrich the ongoing discussion of admission regimes and equality of opportunity in other countries.
Obviously, between-country differences in postsecondary systems preclude a straightforward duplication of this plan to
other nations. Nonetheless, the concept of basing preferences on school/neighborhood affiliation is worth serious consideration, especially by elite postsecondary institutions in the US, which are struggling to find race-neutral ways to increase diversity. Class-based preferential treatment at these institutions is imperative in order to offset the widening
class gaps (Alon, 2009; see also Bowen et al., 2005; Carnevale and Rose, 2004; Espenshade and Radford, 2009; Kahlenberg, 1996), and neighborhood/school-based preference plans seem the most practical and feasible alternative. Wideranging diversity dividends from such policies are likely not only because spatial segregation has a central role in the
formation of life chances in the US, but also because it overlaps with other systems of inequality (Massey, 2007). Such
a policy can replace the banned racial preference programs or supplement them in states where they are still permissible. (In the former case it should be noted that, by default, any race-neutral policy cannot produce the same level of
demographic diversity as racial preferences, as the data from simulations and the implementation of percent plans can
attest.23)
Another issue to consider is the likely public reaction to such a policy in the US. Despite greatly facilitating the administration of such a plan, the focus on group indicators subjects it to the same criticisms often directed at race-conscious policies: reverse discrimination and creaming. Nonetheless, class-based preferences may stir less antagonism than plans that
emphasize race because, similar to percent plans, they do not rely on ascribed traits. To illustrate, Skrentny (1996) contends
that there is no opposition to group preferences per se, but rather to race-conscious admissions in particular because blacks,
unlike, say, veterans, are perceived as undeserving of the preferences.
Furthermore, since they apply to people of all races, such programs are probably in legally safe waters in states that have
banned racial preferences. Bleich and Conrad (2010) reviewed several legal cases in these states and suggest that while quotas and mechanistic racial affirmative action methods are prohibited, more limited consideration of race that does not focus
on individual characteristics is usually allowed. They argue (p. 32) that ‘‘Potentially permissible programs in these untested
waters would include . . . admissions programs that promote diversity using other criteria, such as native language or neighborhood demographics, that ultimately will have an effect on the racial composition of the public work force or student
body.’’ They conclude that significant opportunities exist for creative minds to promote under-represented groups while still
respecting constitutional regimes that ban racial preferences.
Implementing such plans in elite institutions in the US will require additional tailoring to ensure their feasibility and success. For example, the non-mechanistic nature of the admission process and the legality of such programs will be jeopardized
if schools do not erect a wall between the stages of determining SES eligibility and the admission decisions, as is the case in
the Israeli design. A possible solution would be to incorporate structural preferences into a comprehensive full-file review of
applicants; that is, applicants’ academic achievements should be interpreted together with their background information,
particularly place of residence and school quality. Furthermore, since class-based preferences will expand the share of
low-income admits, financial aid budgets will have to be augmented in order to facilitate enrollment at elite and expensive
institutions, and ensure persistence and attainment of a bachelor’s degree. Consequently, the actual diversity dividends of
such programs will also depend on the magnitude of the financial resources dedicated to them. Nonetheless, given the appeal of a color-blind and need-blind affirmative action program, and its potential to diversify bastions of privilege, this is
likely a worthwhile investment.
Acknowledgments
This research was supported by Grants #200800120 and #200900169 from the Spencer Foundation. I thank Erez Garnai
for research assistance.
23
The US Court of Appeals for the Fifth Circuit recently reinforced this point when they ruled to allow the University of Texas (UT) to consider race in
admissions in addition to the Top 10% plan in order to boost the level of diversity in its student body. The majority opinion supported UT’s contention that the
percent plan is not a substitute for race-conscious admission policies.
Author's personal copy
1504
S. Alon / Social Science Research 40 (2011) 1494–1505
Appendix A. Variables’ definitions
Variable
Definition
AA-eligible
Locality: Development town
Whether a student applied for preferential treatment and passed the 30-point threshold
The student’s locality is classified as a Development town. These towns were established
in the 1950s to absorb the massive immigration waves of Jews from Asia and Africa
The classification of local authorities by the socioeconomic level of the population is
based on: financial resources of the residents; housing; standard of living; schooling and
education; employment profile; various types of socio-economic distress; and
demographic characteristics. The index ranges from 1 to 10
Locality ranked at the bottom of the index (ranks 1–3)
Locality ranked at the middle of the index (ranks 4–7)
Locality ranked at the top of the index (ranks 8–10)
The classification of local authorities by their geographic location
Locality: SES cluster
Low-cluster
Mid-cluster
Top-cluster
Locality: Geographic region
Jerusalem
North
Haifa
Center
Tel-Aviv
South
Origin
Asia–Africa 2nd gen.
Europe–America 2nd gen.
Israel
New immigrants – Total
1st gen. – Asia–Africa
1st gen. – Europe–America
Arab
Need-based financial aid
recipient
The student’s most selective
department
Native-born Israeli Jews to at least one parent from Asia–Africa
Native-born Israeli Jews to two parents from Europe–America or one parent of
European–American origin the second native-born Israeli
Native-born Jews to both native-born parents
First generation immigrants
First generation immigrants from Asia–Africa
First generation immigrants from Europe–America
Arab
Student received need-based financial-aid in the first year.
The department’s index of selectivity is the sum of the standardized scores of admission
rate and admits’ mean test scores. Each applicant/admit/student is classified to the most
selective major she applied to; admitted to; and enrolled in
References
Ainsworth, James W., 2002. Why does it take a village? The mediation of neighborhood effects on educational achievement. Social Forces 81 (1), 117–152.
Alon, Sigal, 2001. Racial, Ethnic and Socioeconomic Disparities in College Destinations, 1982 and 1992. Office of Population Research Working Paper No.
2001-02.
Alon, Sigal, 2004a. The gender stratification of employment hardship: queuing, opportunity structure and economic cycles. Research in Social Stratification
and Mobility 20, 115–143.
Alon, Sigal, 2004b. Affirmative action, financial aid and access to postsecondary education in the US. In: Maor, Anat (Ed.), Affirmative Action in Israel. Tel
Aviv, Ramot (in Hebrew).
Alon, Sigal, 2009. The evolution of class inequality in higher education: competition, exclusion and adaptation. American Sociological Review 74, 731–755.
Alon, Sigal, 2010. Racial differences in test preparation strategies. Social Forces 89 (2), 463–474.
Alon, Sigal, Tienda, Marta, 2005. Assessing the mismatch hypothesis: differences in college graduation rates by institutional selectivity. Sociology of
Education 78, 294–315.
Alon, Sigal, Tienda, Marta, 2007. Diversity, opportunity and the shifting meritocracy in higher education. American Sociological Review 72 (4), 487–511.
Blau, Peter M., 1977. Inequality and Heterogeneity: A Primitive Theory of Social Structure. Free Press, New York.
Bleich, J., Conrad, M., 2010. Interpreting proposition 209 and California’s developing case law on affirmative action. In: Grodsky, E., Kurlaender, M. (Eds.),
Equal Opportunity in Higher Education: The Past and Future of California’s Proposition 209. Harvard Education Press, Cambridge, pp. 11–32.
Bowen, William G., Bok, Derek, 1998. The Shape of the River: Long-Term Consequences of Considering Race in College and University Admission. Princeton
University Press, Princeton, NJ.
Bowen, William G., Kurzweil, Martin A., Tobin, Eugene M., 2005. Equity and Excellence in American Higher Education. University of Virginia Press,
Charlottesville, VA.
Brooks-Gunn, Jeanne, Duncan, Greg J., Leventhal, Tama, Lawrence Aber, J., 1997. Lessons learned and future directions for research on the neighborhoods in
which children live. In: Brooks-Gunn, Jeanne, Duncan, Greg, Aber, J. Lawrence (Eds.), Neighborhood Poverty: Context and Consequences for Children,
vol. 1. Russell Sage, New York, NY, pp. 279–298.
Carnevale, Anthony P., Rose, Stephen J., 2004. Socioeconomic status, race/ethnicity, and selective college admissions. In: Kahlenberg, R.D. (Ed.), America
Untapped Resources. The Century Foundation Press, New York, pp. 101–56.
Clayton, Susan D., Tangri, Sandra S., 1989. The justice of affirmative action. In: Blanchard, F., Crosby, F. (Eds.), Affirmative Action in Perspective. SpringerVerlag, New York, pp. 177–192.
Author's personal copy
S. Alon / Social Science Research 40 (2011) 1494–1505
1505
Cohen, Yinon, Haberfeld, Yitchak, 1998. Second generation Jewish immigrants in Israel: have the ethnic gaps in schooling and earnings declined? Ethnic and
Racial Studies 21, 507–528.
Espenshade, Thomas J., Radford, Alexandria Walton, 2009. No Longer Separate, Not Yet Equal: Race and Class in Elite College Admission and Campus Life.
Princeton University Press, Princeton, NJ.
Gratz v. Bollinger, 2003, 539 US 244.
Grodsky, Eric, Kurlaender, Michal, 2010. The demography of higher education in the wake of affirmative action. In: Grodsky, E., Kurlaender, M. (Eds.), Equal
Opportunity in Higher Education: The Past and Future of California’s Proposition 209. Harvard Education Press, Cambridge, pp. 33–58.
Grutter v. Bollinger, 2003, 539 US 306.
Gurin, Patricia, Dey, Eric L., Hurtado, Sylvia, Gurin, Gerald, 2002 Diversity and Higher Education: theory and Impact on Educational Outcomes. Harvard
Educational Review 72(3), 330–366.
Haberfeld, Yitchak, Cohen, Yinon, 1998. Earnings of native-born Jewish and Arab men in Israel, 1987–1993. Research in Social Stratification and Mobility 16,
69–88.
Haberfeld, Yitchak, Cohen, Yinon, 2007. Gender, ethnic, and national earnings gaps in Israel: the role of rising inequality. Social Science Research 36, 654–
672.
Harris, Angel, Tienda, Marta, 2010. Minority higher education pipeline: consequences of changes in college admissions policy in Texas. Annals of the
American Academy of Political and Social Science 627 (1), 60–81.
Israel Central Bureau of Statistics (ICBS), 1995. Table C1: Population in Jewish localities, mixed localities and statistical areas, by continent of origin, period
of immigration and sex. In: Demographic Characteristics of the Population in Localities and Statistical Areas. Census of Population and Housing
Publications.
Israel Central Bureau of Statistics (ICBS), 2006. Table 1: Local councils and municipalities – rank, cluster membership, population, variable values,
standardized values and ranking for the variables used in the computation of the index. In: Characterization and Classification of Local Authorities by
the Socio-Economic Level of the Population. Census of Population and Housing Publications.
Kahlenberg, Richard D., 1996. The Remedy: Class, Race, and Affirmative Action. BasicBooks, New York.
Kane, Thomas J., 1998. Misconceptions in the debate over affirmative action in college admissions. In: Orefield, G., Miller, E. (Eds.), Chilling Admissions: The
Affirmative Action Crisis and the Search for Alternatives. Harvard Education Press, Boston, MA, pp. 17–31.
Karabel, Jerome, 2005. The Chosen: The Hidden History of Admission and Exclusion at Harvard, Yale, and Princeton. Houghton Mifflin, Boston.
Khazzoom, Aziza, 2005. Did the Israeli state engineer segregation? On the placement of Jewish immigrants in development towns in the 1950s. Social Forces
84 (1), 115–134.
Koffman, Dawn, Tienda, Marta, 2008. Missing in Application: Admission Guarantees and Socioeconomic Diversity. THEOP Working Paper. <http://
theop.princeton.edu/reports/wp/ApplicantSocialClass.pdf> (retrieved 22.12.10)
Kurlaender, Michal, Grodsky, Eric, 2010. Toward equal opportunity in higher education. In: Grodsky, E., Kurlaender, M. (Eds.), Equal Opportunity in Higher
Education: The Past and Future of California’s Proposition 209. Harvard Education Press, Cambridge, pp. 217–226.
Lewin-Epstein, Noah, Semyonov, Moshe, 1986. Ethnic group mobility in the Israeli labor market. American Sociological Association 51 (3), 342–352.
Lewin-Epstein, Noah, Semyonov, Moshe, 1992. Local labor markets, ethnic segregation, and income inequality. Social Forces 70 (4), 1101–1119.
Long, Mark C., Saenz, Victor, Tienda, Marta, 2010. Policy transparency and college enrollment: did the Texas Top 10% law broaden access to the public
flagships? Annals of the American Academy of Political and Social Science 627 (1), 82–105.
Massey, Douglas S., 2007. Categorically Unequal: The American Stratification System. Russell Sage Foundation, New York.
Massey, Douglas S., Denton, Nancy, 1993. American Apartheid: Segregation and the Making of the Underclass. Harvard University Press, Cambridge.
McLeod, Jane D., Edwards, Kevan, 1995. Contextual determinants of children’s responses to poverty. Social Forces 73, 1487–1516.
Milem, Jeffrey F., Chang, Mitchel J., Antonio, Anthony Lising, 2005. Making Diversity Work on Campus: A Research-Based Perspective. The Association of
American colleges and Universities. <http://siher.stanford.edu/AntonioMilemChang_makingdiversitywork.pdf> (retrieved 17.05.11).
Pebley, Anne R., Sastry, Narayan, 2004. Neighborhoods, poverty and children’s well-being: a review. In: Neckerman, Kathryn (Ed.), Social Inequality. Russell
Sage Foundation, New York, NY, pp. 119–145.
Sampson, Robert J., Raudenbush, Stephen W., Earls, Felton, 1997. Neighborhoods and violent crime: a multilevel study of collective efficacy. Science 277
(5328), 918–924.
Sampson, Robert J., Morenoff, Jeffrey D., Earls, Felton, 1999. Beyond social capital: spatial dynamics of collective efficacy for children. American Sociological
Review 64, 633–660.
Sander, Richard H., 1997. Experimenting with class-based affirmative action 47. Journal of Legal Education, 472–503.
Semyonov, Moshe, Tyree, Andrea, 1981. Community segregation and the costs of ethnic subordination. Social Forces 59 (3), 649–666.
South, Scott J., Crowder, Kyle D., 1999. Neighborhood effects on family formation: concentrated poverty and beyond. American Sociological Review 64, 113–
132.
Skrentny, John David, 1996. The Ironies of Affirmative Action: Politics Culture and Justice in America. University of Chicago Press.
Spilerman, Seymour, Habib, Jack, 1976. Development towns in Israel: the role of community in creating ethnic disparities in labor force characteristics.
American Journal of Sociology 81, 781–812.
The National Association for College Admission Counseling (NACAC), State of College Admission. David A. Hawkins and Melissa Clinedinst, NACAC. <http://
www.nacacnet.org/PublicationsResources/Research/Documents/06StateofAdmission.pdf> (retrieved June 2009).
Tienda, M., 2010. Lessons about college admissions from the Texas top 10 percent law. In: Grodsky, E., Kurlaender, M. (Eds.), Equal Opportunity in Higher
Education The Past and Future of California’s Proposition 209. Harvard Education Press, Cambridge, pp. 195–210.
Tienda, Marta, Alon, Sigal, Niu, Sunny X., 2010. Affirmative action and the Texas top 10% percent admission law: balancing equity and access to higher
education. Sociétés Contemporaines 79, 19–39.
US Commission on Civil Rights (USCCR), 2002. Beyond Percentage Plans: The Challenge of Equal Opportunity in Higher Education. Draft Staff Report. <http://
www.vpcomm.umich.edu/admissions/research/beyondpercentplns.pdf> (retrieved 21.03.11)
Wilson, J. William, 1987. The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy. University Of Chicago Press.
Wilson, J. William, 1996. When Work Disappears: The World of the New Urban Poor. Knopf, New York.
Yaish, Meir, 2001. Class structure in a deeply divided society: class and ethnic inequality in Israel, 1974–1991. British Journal of Sociology 52 (3), 409–439.
Yiftachel, Oren, 2000. Social control, urban planning and ethno-class relations: Mizrahi Jews in Israel’s ‘development towns’. International Journal of Urban
and Regional Research 24 (2), 418–438.