ArticlePDF Available

Abstract

To understand the persistent social class achievement gap, researchers have investigated how educational settings affect lower vs. higher socio-economic status (SES) students’ performance. We move beyond the question of actual performance to study its assessment by evaluators. We hypothesized that even in the absence of performance differences, assessment’s function of selection (i.e., compare, rank and track students) leads evaluators to create a SES achievement gap. In two experiments (N = 196; N = 259), participants had to assess a test supposedly produced by a high- or a low-SES student, and used assessment for selection (i.e. normative grading) or learning (i.e. formative comments). Results showed that evaluators using assessment for selection found more mistakes if the test was attributed to a low- rather than a high-SES student, a difference reduced in the assessment for learning condition. The third and fourth experiments (N = 374; N = 306) directly manipulated the function of assessment to investigate whether the production of the social class achievement gap was facilitated by the function of selection to a greater extent than the educational function. Results of Experiment 3 supported this hypothesis. The effect did not reach significance for Experiment 4, but an internal meta-analysis confirmed that assessment used for selection led evaluators to create a SES achievement gap more than assessment used for learning, thereby contributing to the reproduction of social inequalities.
Running head: SELECTION AND SOCIAL CLASS
!
1
The Function of Selection of Assessment Leads Evaluators to Artificially Create the Social
Class Achievement Gap
Accepted for publication in: Journal of Educational Psychology
Frédérique Autin
Anatolia Batruch
Fabrizio Butera
Université de Lausanne
Author Note
Frédérique Autin, Anatolia Batruch, Fabrizio Butera, UNILaPS, Institut de Psychologie,
Faculté des Sciences Sociales et Politiques, Université de Lausanne, Lausanne, Switzerland.
Frédérique Autin is now at CeRCA, CNRS - Université de Poitiers, France. This work was
supported by the Swiss National Science Foundation (grant CRSII1_141872), and was
conducted during Frédérique Autin’s postdoctorate at University of Lausanne under the
supervision of Fabrizio Butera. We wish to thank the members of the Sinergia project “The
struggle for competence in academic selection: Social psychological influences on
Competence Threat” and Benoit Dompnier for their help with the development of materials
and their comments on the research plan and results, and Nicolas Sommet and Suzanne Faber
for their help during data collection. Correspondence concerning this article should be
addressed to Fabrizio Butera, Université de Lausanne - IP-SSP, Géopolis. CH 1015 -
Lausanne, Switzerland. E-mail: fabrizio.butera@unil.ch
Running head: SELECTION AND SOCIAL CLASS
!
2
Abstract
To understand the persistent social class achievement gap, researchers have investigated how
educational settings affect lower vs. higher socio-economic status (SES) students’
performance. We move beyond the question of actual performance to study its assessment by
evaluators. We hypothesized that even in the absence of performance differences,
assessment’s function of selection (i.e., compare, rank and track students) leads evaluators to
create a SES achievement gap. In two experiments (N = 196; N = 259), participants had to
assess a test supposedly produced by a high- or a low-SES student, and used assessment for
selection (i.e. normative grading) or learning (i.e. formative comments). Results showed that
evaluators using assessment for selection found more mistakes if the test was attributed to a
low- rather than a high-SES student, a difference reduced in the assessment for learning
condition. The third and fourth experiments (N = 374; N = 306) directly manipulated the
function of assessment to investigate whether the production of the social class achievement
gap was facilitated by the function of selection to a greater extent than the educational
function. Results of Experiment 3 supported this hypothesis. The effect did not reach
significance for Experiment 4, but an internal meta-analysis confirmed that assessment used
for selection led evaluators to create a SES achievement gap more than assessment used for
learning, thereby contributing to the reproduction of social inequalities.
Keywords: social class achievement gap, educational institutions, function of
selection, evaluator, assessment practices
Running head: SELECTION AND SOCIAL CLASS
!
3
Educational Impact and Implications Statement
Evaluators’ knowledge about students’ social class can bias their assessment, in favor of
privileged students. The present research suggests that assessment in itself does not trigger
such a bias, nor are teachers biased in themselves; rather it is the function given to assessment
that can trigger or prevent discriminatory assessment. This research found that a social class
gap in evaluation appears when assessment is used for selective purposes (i.e., gauging merit
and sorting students) to a greater extent that when it is used for educational purposes (i.e.,
fostering learning). The findings indicate that to ensure equality in educational institutions,
closer attention should be paid to the role and meaning of assessment.
Running head: SELECTION AND SOCIAL CLASS
!
4
The Function of Selection of Assessment Leads Evaluators to Artificially Create the Social
Class Achievement Gap
In most industrialized countries, educational institutions have developed with the goal
of establishing a fair society in which social positions are ascribed based on individual merit,
irrespective of individuals’ social class (Bell, 1973; Duru-Bellat, 2006; Turner, 1961). And
yet, a wealth of empirical evidence questions the fact that the educational system truly
provides equal opportunities and fosters social mobility. For example, international testing
such as the Program for International Student Assessment (PISA) consistently shows that,
across many countries (65 involved in 2012), low socio-economic status (SES) students are
more likely to underperform compared to high-SES students (OECD, 2006, 2013a). To
explain the persistent social class achievement gap, some scholars pointed to the way
educational institutions function (Bourdieu & Passeron, 1977; Croizet, Goudeau, Marot, &
Millet, 2017; Stephens, Markus, & Phillips, 2014). A steadily growing research stream in
social and educational psychology has investigated how educational settings create a set of
barriers that hinder the success of low-SES students while supporting the performance of
high-SES students (e.g., stereotype threat, cultural mismatch; Croizet & Claire, 1998;
Stephens, Fryberg, Markus, Johnson, & Covarrubias, 2012). In the present article, we propose
to move beyond the question of the processes affecting students’ performance, and address
the processes that contribute to the social class achievement gap via evaluators. We argue that
the use of assessment practices with a focus on selection (i.e., compare and rank students to
guide them toward different social positions) can lead evaluators to create a social class
achievement gap, even in the absence of objective performance differences.
Educational Institutions and Students’ Performance
Many analyses of educational institutions suggest that they play a role in perpetuating
social inequalities (e.g., Bourdieu & Passeron, 1977; Croizet & Millet, 2012; Fine & Burns,
Running head: SELECTION AND SOCIAL CLASS
!
5
2003; Stephens et al., 2014). The way institutions operate can be understood as a social
product that not only conveys cultural ideas, values and beliefs (Fiske, Kitayama, Markus, &
Nisbett, 1998; Markus & Hamedani, 2007), but also carries traces of power relations between
social groups that participate in the creation, maintenance and justification of inequalities
(Adams, Biernat, Branscombe, Crandall, & Wrightsman, 2008; Jackman, 1994). Educational
institutions have been created around values, norms regarding language use, bodily posture,
self models and forms of knowledge that are close to those of the middle and upper classes
(Bourdieu & Passeron, 1977; Croizet, et al., 2017; Stephens, et al., 2014). One consequence is
that students from low status groups suffer harmful effects in these institutions while the
experience of individuals from dominant groups is improved (Goudeau & Croizet, 2017; Jury
et al., 2017).
In line with these ideas, research has identified a set of characteristics of educational
settings that leads low-SES students to underperform and foster the performance of high-SES
students. The evaluative dimension of educational settings, by making lower social class
students’ stereotype of incompetence salient (Cozzarelli, Wilkinson, & Tagler, 2001; Durante
& Fiske, 2017), contributes to the SES performance gap (Croizet & Claire, 1998; Croizet &
Dutrévis, 2004; Désert, Préaux, & Jund, 2009; Harrison, Stevens, Monty, & Coakley, 2006;
Spencer & Castano, 2007). Another line of research argues that the performance gap is fueled
by the norms of independence (i.e., express yourself, follow your own path) institutionalized
in American universities, that match the middle or upper class students’ upbringing, but
mismatch the more interdependent socialization of lower class students1 (i.e., be responsive to
others, work with them and contribute to a community; Stephens et al., 2012; Stephens,
Hamedani, & Destin, 2014).
These lines of research are important because they document how educational settings
are often organized in a way that leads lower SES students to be outperformed by higher SES
Running head: SELECTION AND SOCIAL CLASS
!
6
students (via stereotype threat, cultural mismatch). As a consequence, they mitigate the
interpretation of the social class achievement gap in terms of essentialized differences
between students of different social class, and pave the way to interventions aimed at
reducing the effects of those barriers (see Dittmann & Stephens, 2017; Jury et al., 2017). For
instance, Harackiewicz, Canning, Tibbets, Giffen, Blair, Rouse, and Hyde (2014) have shown
that an intervention asking students to write about their most important values serves as a
buffer against social identity threat and reduces the social class achievement gap (see also
Tibbetts et al., 2016). These results may lead one to think that if the barriers affecting lower
SES students’ performance were removed then educational institutions would offer real
equality of opportunity. Yet, we propose that even in the absence of actual performance
differences, other processes are at work to maintain the social class achievement gap. Thus,
we now turn to a set of studies that point to sources of inequalities that go beyond students’
performance.
Educational Institutions and Evaluators’ Behavior
A parallel line of research has pointed out that teachers’ evaluation of performance can
be biased by their knowledge of a student’s social background (e.g., see Malouff &
Thorsteinsson, 2016). For example, Sprietsma (2013) asked German teachers to grade essays
of unknown fourth-graders. Typical German or Turkish names were randomly assigned to the
same essays. The essays received lower grades when the teachers thought that students with a
migrant background, compared to native students, had produced them. It should be noted that
students with a migrant background tend to be socio-economically disadvantaged compared
to native students (OECD, 2013a). Rangvid (2015) used large-scale data registers to compare
teacher scores and external exam scores. Disparities between these scores indicate bias in
teachers grading. The study showed notably that pupils with low-educated parents (an aspect
Running head: SELECTION AND SOCIAL CLASS
!
7
of lower social class backgrounds) receive lower teacher scores than pupils with high-
educated parents with similar external scores.
The above results strikingly reveal that, even if actual performance is identical, the
outcome of assessment is influenced by the students’ social background. However, in these
studies, discrimination in grading is usually interpreted as an effect of teachers’ bias: They
hold prejudiced expectations based on the students’ social backgrounds, which affect their
behavior, even if this discrimination is not intentional. Without underplaying the impact of
expectations, we propose that biased assessment cannot be isolated from the sociocultural
context in which this behavior is produced. When assessing, teachers act as agents of an
institution that conveys specific values and norms and promote specific practices; we thus
contend that biased assessment can be interpreted as the product of the way educational
institutions are structured and operate.
Two Functions of Educational Institutions
Modernized industrial societies are faced with a paradox: How to reconcile an
endorsement of equality of all humans as a fundamental value, with being stratified (i.e.,
different occupations give unequal access to symbolic and material resources). To find
justifiable ways to rank individuals, most Western societies opted for an ascription of social
positions based on a characteristic seemingly naturally distributed across individuals:
individual merit (Bisseret, 1974; Carson, 2007). Educational institutions became the place
where individual differences can be detected, gauged and certified, to give access to the
corresponding social positions. The paradox between equality and stratification is thus
embodied in educational institutions, which are expected to serve two different functions in
society: an educational function, ensuring equality of opportunity, and a function of selection,
sorting individuals by merit (Darnon, Dompnier, Delmas, Pulfrey, & Butera, 2009;
Dornbusch, Glasgow, & Lin, 1996).
Running head: SELECTION AND SOCIAL CLASS
!
8
Educational Function
Following the Universal Declaration of Human Rights stating that “everyone has the
right to education”, most Western societies implemented compulsory elementary education
and free access to public schools. The unrestricted access to education serves as a safeguard
for equality of opportunity. The educational function of educational institutions refers to their
role in equipping all students with knowledge, skills and capacities for learning and helping
them develop their potential. Having all individuals master basic knowledge and competence
ensures that they can all take part in society (Dubet, 2004; Forquin, 1992; Parsons, 1959).
Moreover, the democratization of knowledge is expected to expand opportunities and ensure
that no talent is wasted; accordingly, the educational function is perceived as promoting social
mobility (Bowen, Kurzweil, Tobin, & Pichler, 2005; Duru-Bellat, 2008).
Function of Selection
If mass education offers to all the opportunity to show their potential, then education
institutions fulfill a function of selection by sorting individuals into different educational paths
and ultimately different occupations. The 2012 PISA survey established that in all 64
countries, educational institutions implement some form of selection practices (OECD,
2013b), which include school admission, transfer, grade repetition, tracking into academic or
vocational programs, grouping across and within classes and combinations of these. To
illustrate, tracking into different programs occurs in 75% of the countries. Across countries,
43% of the 15-year-old students are in academically selective schools, 75% attend schools
that use between-classes ability grouping and 49% within-classes grouping. The time and
rigidity of this stratification varies between countries; nevertheless, all selection practices
have consequences for the students’ educational trajectory and, at each step of selection, a
reduced proportion of the population moves to the most valued tracks.
Running head: SELECTION AND SOCIAL CLASS
!
9
The different educational lanes are conceptualized as a way to develop the students’
potential, meet their needs and assure that they are in the right place (Chmielewski, 2014;
LeTendre, Hofer, & Shimizu, 2003). Indeed, the function of selection is intertwined with the
meritocratic ideal. Educational institutions are viewed as a neutral context to detect and
measure the qualities of students, and select the most deserving (Carson, 2007; Lemann,
1999). Accordingly, educational institutions have been described as a social filter (Arrow,
1973), or sorting machines (Domina, Penner, & Penner, 2017), since academic credentials
have become the supposedly fair basis for ascribing positions in the occupational hierarchy.
To fulfill both their functions – education and selection – educational institutions rely
notably on assessment. As a consequence, the distinction between different functions exists in
the theorization of assessment. Assessment can serve an educational role of promotion of
learning, and a social role of estimation of merit, ranking and certification, namely the
function of selection (Filer, 2002; Taras, 2005, 2009; Torrance & Pryor, 1998). We contend
that beyond the consequences for students’ learning, the two functions underlying assessment
can have consequences in terms of social inequalities. This contention builds on research
focusing on the function of assessment and inequalities in students’ performance. Giving a
formative framing to evaluation, by stating that critical feedback reveals the teacher’s belief in
the students’ potential, improved low-status students’ performance (Yeager et al., 2014). On
the contrary, reminding students of the function of selection increased their belief in the utility
of outperforming others to succeed at college and consequently increased their endorsement
of such a performance-approach goal (Jury, Darnon, Dompnier, & Butera, 2017). Endorsing
this goal predicted better grades but only for higher social class students (Darnon, Jury, &
Aelenei, 2017). More directly related to the present research, assessment presented as a tool
for selecting the best students elicited a SES achievement gap on an exam, a gap that was
closed when assessment was presented as a tool for learning (Smeding, Darnon, Souchal,
Running head: SELECTION AND SOCIAL CLASS
!
10
Toczek-Capelle, & Butera, 2013). Similarly, simply reminding students of the function of
selection of university led lower SES students to underperform compared to higher SES
students (Jury, Smeding, & Darnon, 2015). We argue that, as the institutional function of
assessment impacts the student’s construal of the performance setting, it can likewise
influence the evaluator’s construal of the assessment setting. We propose that assessment for
selection reflects a meritocratic ethos while assessment for learning reflects an egalitarian
ethos, which beget different consequences for the reproduction of inequalities. Specifically,
we propose that assessment for selection might induce more reproduction of inequalities than
assessment for learning.
Assessment for Selection and Inequality
Although all forms of assessment can serve both educational and selection functions,
grading may especially be relevant for selective purposes, in that it allows normative
assessment. Traditionally, normative assessment, or norm-referenced assessment, is
conceived as allowing one to compare the performance of the person being assessed to that of
other persons (Glaser, 1963). Normative assessment uses indicators such as numerical grades,
letters, percentages, or value judgments (e.g., good, excellent), that can also be used in other
assessment methods, but perfectly serve the purpose of normative assessment, namely
comparison to a standard and across individuals (Rosenholtz & Simpson, 1984; Thorndike,
1913). These indicators summarize performance in a number—or a letter, or a judgment—and
thereby constitute an easily interpretable criterion of relative success or failure (Butler, 1987;
Butler & Nisan, 1986). In industrialized countries, normative grading constitutes the most
widely used method of assessment in educational and professional settings (Knight & Yorke,
2003), and is the main basis for admission to schools and programs (OECD, 2013b).
Beyond the institutional role of grading, this form of evaluation is seen as well suited
to select students who are most deserving. In fact, research has shown that the more
Running head: SELECTION AND SOCIAL CLASS
!
11
individuals believed that the function of educational institutions is to select the best students,
the more they supported the implementation of normative grading; this relationship was
mediated by the belief that grading fulfills equity justice principles (Autin, Batruch, & Butera,
2015). Other research showed that both teachers and students believe that grade distribution is
fair as long as it follows an equity principle (Jasso & Resh, 2002; Resh, 2009; Sabbagh,
Faher-Aladeen, & Resh, 2004). These elements highlight the intertwinement of normative
grading, the function of selection and the meritocratic ethos, which assumes that rewards
should be allocated equitably, on the basis of individual ability and hard work (Son Hing et
al., 2011; Wiederkehr, Bonnot, Krauth-Gruber, & Darnon, 2015). This feature of educational
institutions, however, is not without consequences in terms of inequalities.
Contexts emphasizing meritocratic selection elicit psychological and behavioral
tendencies to justify and maintain social inequalities. For example, believing in meritocracy
decreases perceptions of discrimination in low status groups (McCoy & Major, 2007) and
perceptions of privilege in dominant groups (Knowles & Lowery, 2012). Moreover, the
perceived violation of meritocratic selection is central in the opposition to social policies that
challenge the status quo (Bobocel, Son Hing, Davey, Stanley, & Zanna, 1998; Faniko,
Lorenzi-Cioldi, Buschini, & Chatard, 2012; Zdaniuk & Bobocel, 2011). In education, the
more students and parents believe in school meritocracy the less they are willing to implement
a pedagogical intervention that reduces the SES achievement gap (Darnon, Smeding, &
Redersdorff, 2017).
More directly related to the effect of meritocratic assessment on bias in evaluators,
Castilla and Benard (2010) found that inducing an organizational culture that emphasizes
meritocracy led individuals in a managerial position to favor a male employee over a female
employee who achieved similar performance. Closely related to the matter of academic
assessment, a recent study had pre-service teachers grade a test that was attributed either to a
Running head: SELECTION AND SOCIAL CLASS
!
12
low- or a high-SES student. When the student was presented as being enrolled in a selective
program, pre-service teachers gave a lower grade to the test attributed to a low-SES student
comparatively to a high-SES student. The gap in evaluation was reduced if students were
supposedly enrolled in a less selective program (Batruch, Autin, & Butera, 2017). It is
important to note that we do not suggest that assessment in itself necessarily produces biased
evaluations; instead, we propose that assessment practices that focus on meritocratic
selection, such as normative grading, may lead evaluators to reproduce inequalities in their
evaluations.
Assessment for Learning and Equality
Alternative forms of assessment have long been developed, including formative
assessment (Black & Wiliam, 1998), which can be defined as assessment providing specific
and detailed feedback with the goal of adjusting the teaching and learning activities to the
students’ needs and providing relevant comments on how to overcome difficulties and make
progress (Frey & Schmitt, 2007; Sadler, 1989). Formative assessment is often opposed to
summative assessment, to the extent that the former intervenes during the learning process
and the latter at the end of it (Bloom, Hastings, & Madaus, 1971). However, in the present
research we do not focus on the temporal aspects of formative assessment, but on its function,
that of providing feedback for learning. In particular, among the various existing kinds of
formative assessment, we refer to qualitative feedback that points to specific learning
objectives and suggests ways to improve (Bennett, 2011; Shute, 2008). Formative feedback
provides useful information to students: what the expected outcome is and guidance on how
to attain it (Sadler, 1989); and most importantly, feedback related to the task reduces the focus
on social comparison and enhances the focus on the mastery of the task (Bloom, 1968; Butler,
1987). Because of its focus on the development of competence and knowledge, formative
assessment is in line with educational institution’s educational function.
Running head: SELECTION AND SOCIAL CLASS
!
13
Equality is central to the rationale for implementing formative assessment. It is
presented as a tool to implement a corrective process and reduce the gap between individuals
who are unequal before entering school (Crahay, 2012; Perrenoud, 1995). By enabling
adjustment to meet the students’ needs, formative feedback aims at helping all students to
attain a high level of competence, irrespective of their initial abilities. Formative assessment
thus seems to convey the institutional purpose of education centered on equality. Beyond this
institutional role, research shows that people’s support for formative feedback is related to a
principle of corrective justice that ensures equality of outcomes through adjustment to the
students’ needs (Autin et al., 2015).
Importantly, promoting equality has positive effects on the treatment of groups.
People’s endorsement of egalitarian principles relates to lower levels of stereotyping
(Moskowitz, Gollwitzer, Wasel, & Schaal, 1999; Moskowitz, Salomon, & Taylor, 2000) and
the to support for social policies aimed at reducing social inequalities (Zdaniuk & Bobocel,
2011). Invoking the concept of equality also induces a more favorable implicit evaluation of
an out-group (Zogmaister, Arcuri, & Castelli, 2008). Compared to activating meritocratic
values, activating egalitarian values reduces the accessibility of negative stereotypes (Wyers,
2003) and elicits more positive attitudes toward a low status group (Katz & Hass, 1988). It
also diminishes the extent to which prejudice relates to avoidance activation in response to
low status groups (Wyer, 2010). Thus, we propose that assessment practices oriented toward
educational purposes, such as formative feedback, may prevent evaluators from reproducing
inequalities in their evaluation.
Hypotheses and Overview
In the present research, we argue that institutional practices of assessment constrain
the way individuals in a position of evaluator behave toward students from lower or higher
social class. In two experiments, we test the hypothesize that, compared to assessment for
Running head: SELECTION AND SOCIAL CLASS
!
14
learning, assessment for selection leads evaluators to create a larger achievement gap that
reproduces existing social inequalities (i.e., low-SES students have a lower performance than
high-SES students), even though the actual performance is identical. In a third and fourth
experiment, we test that it is indeed the function of assessment, rather than the form of
assessment (grade vs. comments), that leads evaluators to differentially evaluate students as a
function of their SES.
It should be noted that considering the difficulty to recruit practicing teachers, the
studies were conducted with college students playing the role of teachers. Although this is a
limitation, we decided to first test our hypotheses and paradigm with an accessible population
to be able to conduct well-powered studies. We believe that the long-lasting socialization of
the students in the educational institution implies that they are well aware of its functions and
practices (see Darnon et al., 2009) and therefore able to enact them. After all, we hypothesize
that it is the educational system’s selective function should drive the effects. Moreover
previous studies has shown similar results with students acting as teachers and actual teachers
(Batruch, Autin, Bataillard, & Butera, in press; Rattan, Good, & Dweck, 2012; Simon,
Ditrichs, & Grier, 1995)
Experiment 1
Method
Participants. A total of 220 students from a medium-size French-speaking Swiss
university participated in return for a 10 CHF (10.30 USD) gift card. At the time of the study,
we used a rule of at least 50 participants per cell to determine sample size (Simmons, Nelson,
& Simonsohn, 2011). Data collection stopped at the end of the semester considering that the
sample size requirement had been reached. Data from 17 non-French native speakers and 7
participants who failed the manipulation check were excluded. The analyses including
participants who failed the check are reported in the supplemental materials. The final sample
Running head: SELECTION AND SOCIAL CLASS
!
15
comprised 196 students (117 women, 79 men, Mage = 22.29, SDage = 1.87). Participants were
randomly assigned to one of the experimental conditions in the Assessment method (grading
vs. formative comments) x Target’s SES (low vs. high) between-participants design. The
target’s sex was also manipulated as a control and was not part of our hypotheses. It should be
noted that in Switzerland experiments that do not include physical measures or vulnerable
participants do not need permission from an ethical committee and the experiment was
conducted in compliance with the APA ethical guidelines. Participants were informed about
confidentiality and anonymity of data, right to decline and withdraw without consequences,
and whom to contact in case of questions.
Material and procedure. Students were approached in university cafeterias by one of
two experimenters and asked whether they would take part in a study about assessment tools
used by teachers. Participants received a booklet containing instructions about the assessment
method, a description of the target (i.e., the student who produced the test) followed by a
dictation to be assessed. Participants read a cover story asking them to imagine that they were
a French-language teacher in a secondary school, and to assess a dictation test using a specific
method.
Manipulation of the assessment method. Instructions were based on the specific
properties of the assessment method reviewed above (i.e., grading vs. formative comments).
Participants in the assessment for selection condition read that, as a teacher, they were to use a
method based only on grades. They were to give students grades depending on the number of
mistakes they made. The instructions also referred to the normative aspect of this assessment,
that relates to the social function of certification and ranking (cf. Taras, 2009, 2005).
Participants read that this method allows checking the student’s level and whether he/she met
the requirements. They also read that this method allows assessment of the students’ learning,
Running head: SELECTION AND SOCIAL CLASS
!
16
their standing compared to a norm that defines success and compared to the other students.
This explanation was illustrated with an example of a math test graded with this method.
Participants in the assessment for learning condition read that they were to use a
method based on formative comments only. They were to make comments to help the
students learn from their mistakes. Instruction referred to the educational role of assessment.
Participants read that this method explains to students how to improve and to adapt to
learning situations. They also read that such method allows assessment of the students’
learning and their distance from the learning goals, and propose them strategies to meet these
goals. This description was illustrated with an example of a math test corrected with this
method.
Manipulation of the target’s SES. After reading about the assessment method,
participants were presented with information about a student allegedly belonging to their
class. Participants saw the student file (similar to the official student file in use) and a brief
description of his/her extra-curricular activities. Relevant information about the target’s SES
were presented among neutral information (e.g., date of birth, address, nationality—all targets
were presented as Swiss). SES was manipulated via a series of indicators. The student’s first
name was manipulated using stereotypical names of higher- vs. lower-SES girls and boys
(e.g., “Louis” for a high-SES boy, “Brian” for a low-SES boy, “Charlotte” for a high-SES
girl, and “Cindy” for a low-SES girl), based on Coulmont’s (2011) work on the sociology of
first names. Moreover, parental occupation (mother: director of marketing vs. waitress; father:
architect vs. construction workman), number of siblings (1 vs. 4) and extra-curricular
activities (e.g., local amusement park vs. traveling to London) were also manipulated. Sex
was manipulated through the student’s first name and reported sex.
Dictation test. After reading the relevant information about the target, participants had
to correct a dictation test. They were asked to first underline all the mistakes. Then, in the
Running head: SELECTION AND SOCIAL CLASS
!
17
assessment for selection condition, participants had to give a grade in line with common
practice in Swiss schools, i.e., from 1 to 6, with higher numbers indicating better
performance. In the assessment for learning condition, participants had to write a comment
next to each mistake to explain the student what mistake he/she did and how to improve. The
test contained 11 obvious mistakes (wrong spelling, wrong verb conjugation and wrong
name-adjective agreement) and 6 ambiguous mistakes (two possible conjugations or
spellings)2.
The booklet ended with some manipulation check items—two asking to report
information presented in the description of the target and one asking to rate the socio-
economic background of the target (from 1 highly disadvantaged to 7 highly advantaged)—as
well as socio-demographic questions, including self-reported GPA. Finally, participants were
thanked and debriefed. As we anticipated that the use of formative comments would take
more time than the use of grading, we recorded the time that participants took to complete the
study.
Results
Perceived SES. To determine whether the description of the target affected
participants’ perception of his/her socio-economic status, we analyzed participants’ rating of
the target’s socio-economic background with the full sample except for non-native speakers.
The regression included the Assessment Method (assessment for selection coded -0.5,
assessment for learning coded 0.5), Target’s SES (low-SES coded -0.5, high-SES coded 0.5)
and the interaction as predictors3. As expected, a main effect of Target’s SES was obtained
indicating that low-SES targets were perceived as coming from a more disadvantaged
background (M = 3.84, SD = .87, 95% CI [3.67, 4.01]) than high-SES targets (M = 5.92, SD =
.75, [5.77, 6.07]), b = 2.07, 95% CI [1.85, 2.31], t(197) = 18.14, p < .001, η2p = .63, 95% CI
[.55, .68]) . The main effect of Assessment Method and the interaction did not reach
Running head: SELECTION AND SOCIAL CLASS
!
18
significance, respectively b = 0.02, t(197) = 0.20, p = .84 and b = 0.27, t(197) = 1.19, p = .23.
As the vast majority of participants perceived the targets as belonging to the expected social
class, we decided to exclude the 7 participants mentioned in the Participants section because
they either perceived a low-SES target’s socio-economic background as “advantaged”, or a
high-SES target’s socio-economic background as “average” or “disadvantaged”.
Number of mistakes. A preliminary analysis revealed that participants took more
time to complete the study when they had to use formative comments (M = 22.00, SD = 6.04,
[20.81, 23.19]), compared to grading (M = 14.94, SD = 4.30, [14.07, 15.82]), b = 7.09, [5.59,
8.58], t(192) = 9.37, p < .001, η2p = .31,![.21, .41]. Time was not affected by the target’s SES
or interactions between SES and Assessment Method, respectively b = 0.59, t(192) = 0.78, p
= .43 and b = -0.25, t(192) = -0.16, p = .87. We decided to control for time in the analysis of
the number of mistakes detected in the dictation test, because the time needed to assess a test
could affect the number of mistakes found, but is not a variable of interest here. Following the
recommendations for the inclusion of covariates, we tested for possible interactions between
the covariate Time and the predictors (Judd, McClelland, & Ryan, 2009). We performed a
regression analysis on the total number of mistakes with Assessment Method (assessment for
selection coded -0.5, assessment for learning coded 0.5), Target’s SES (low-SES coded -0.5,
high-SES coded 0.5), Time (centered) and all interaction terms as predictors4.
Results showed a main effect of Time, with participants detecting more mistakes as
they took more time to complete the study, b = 0.17, [0.11, 0.24], t(185) = 5.25, p < .001, η2p
= .13, [.05, .22]. The main effect of Assessment Method also reached significance indicating
that participants detected more mistakes when using assessment for selection (M = 10.16, SD
= 2.53, [9.57, 10.75]) than assessment for learning (M = 8.61, SD = 2.18, [8.11, 9.10]), b = -
1.55, [-2.32, -0.78], t(185) = -3.97, p < .001, η2p = .07, [.02, .16]. The target’s SES also
affected the number of mistakes such that participants found more mistakes in the dictation of
Running head: SELECTION AND SOCIAL CLASS
!
19
low-SES students (M = 9.90, SD = 2.54, [9.33, 10.47]) than in that of high-SES students (M =
8.87, SD = 2.16, [8.34, 9.39]), b = -1.03, [-1.81, -0.26], t(185) = -2.64, p = .009, η2p = .03,
[.00, .10]. Time interacted with Assessment Method, b = -0.17, [-0.29, -0.04], t(185) = -2.55,
p = .01, η2p = .03, [.00, .10] and with the Target’s SES, b = -0.15, [-0.28, -0.02], t(185) = -
2.33, p = .02, η2p = .03, [.00, .09]. The positive relationship between time and the number of
mistakes was stronger in the grading compared with the formative comments condition and
for low-SES students compared with high-SES students. The expected interaction between
SES and method was not significant, b = 0.92, [-0.62, 2.47], t(185) = 1.18, p =.24, η2p = .01,
[.00, .05], Cohen’s d = .19, but in the expected direction, suggesting a greater gap in the
number of mistakes between low and high SES students in the assessment for selection
condition than in the assessment for learning condition.
However, these effects were qualified by a three-way interaction between Time,
Assessment Method and Target’s SES, b = 0.46, [0.20, 0.72], t(185) = 3.52, p < .001, η2p =
.06, [.01, .14]. This interaction, depicted in Figure 1, was unexpected but made sense given
the effect of time, and was therefore decomposed by assessment method. In the assessment
for selection condition, the positive relationship between time and the number of mistakes
was significantly stronger for low-SES targets than for high-SES targets, b = -0.38, [-0.60, -
0.17], t(185) = -3.50, p < .001, η2p = .06, [.01, .14]. In other words, the more participants
spent time assessing a dictation test with grading, the more they found mistakes, especially if
the target was from a low socio-economic background. As a result, participants who took a
moderate and long time to complete the study in the assessment for selection condition found
on average respectively 1.50 (95% CI for b [-2.68, -0.31]), and 3.90 (95% CI for b [-6.22, -
1.59]) more mistakes in the dictation of a low-SES student than in the dictation of a high-SES
student, respectively t(185) = -2.50, p = .01, η2p = .03, [.00, .10], and t(185) = -3.33, p = .001,
η2p = .06, [.01, .13]. In the assessment for learning condition, the positive relationship
Running head: SELECTION AND SOCIAL CLASS
!
20
between time and the number of mistakes did not significantly differ as a function of Target’s
SES, b = 0.08, [-0.07, 0.22], t(185) = 1.07, p = .28, η2p = .01, [.00, .05].
Supplementary analyses.
Grades. To better understand the mechanisms at work in the assessment for selection
condition, we analyzed how mistakes affected the participants’ grading of the test (given the
design, grades were only available for the assessment for selection condition). Grades were
analyzed in a regression with the Target’s SES (low-SES coded -0.5, high-SES coded 0.5),
the number of mistakes (centered) and the interaction term as predictors5. The analysis
revealed a main effect of the number of mistakes, such that the more mistakes were detected
the lower the grade, b = -0.14, [-0.18, -0.09], t(87) = -6.38, p < .001, η2p = .32, [.17, .45]. The
main effect of SES was not significant , b = 0.22, [-0.02, 0.45], t(87) = 1.85, p = .07, η2p = .04,
[.00, .14]. The interaction between the number of mistakes and the target’s SES was
significant, b = 0.09, [0.009, 0.19], t(87) = 2.19, p = .03, η2p = .05, [.00, .16]. As can be seen
in Figure 2, the negative relationship between the number of mistakes and the grade was
stronger for low-SES targets, b = -0.19, [-0.25, -0.13], t(87) = -6.07, p < .001, η2p = .29, [.15,
.43], compared to high-SES targets, b = -0.09, [-0.15, -0.03], t(87) = -2.96, p = .004, η2p = .09,
[.01, .22]. This suggests that mistakes led to a more negative evaluation when they were
produced by low-SES students. Importantly, this negative evaluation resulted in more low-
SES students performing below the passing grade (4, in Swiss schools). In the sample, we
observed that 32.5 % of the low-SES targets received a grade lower than 4 but this proportion
dropped to 16.6 % for the high-SES targets.
Impact of participants’ characteristics. We tested whether participants’ own
characteristics could account for or moderate the results observed on the number of mistakes.
We looked at the effect of participants’ level of competence, gauged by self-reported GPA,
and their own social class, indicated by whether at least one of their parents has a college
Running head: SELECTION AND SOCIAL CLASS
!
21
degree (“continuing generation”) or not (“first generation”). The analyses revealed no
evidence that including the participants’ level of competence or their social class moderated
the observed results (see supplemental material).
Discussion
This first experiment was designed to test the hypothesis that assessment for selection,
more than assessment for learning, would lead evaluators to reproduce existing social
inequalities and find lower performance for low-SES students than for high-SES students,
even though the actual performance was identical. The target’s SES x assessment method
interaction that tested this hypothesis was not significant, although in the expected direction.
Participants found on average 1.49 more mistakes in the low-SES test than in the high-SES
test when using assessment for selection, a SES performance gap reduced to .57 in the
assessment for learning condition. The size of this effect is small (η2p = .01, Cohen’s d = .19)
but we believe it should not be disregarded. Indeed, students in the member countries of the
Organisation for Economic Co-operation and Development (OECD, e.g., USA, Australia,
Latvia, Korea, Germany, Mexico) receive on average 9 years of compulsory education and
can expect to receive 17 years of study over their lifetime (OECD, 2006). During this time,
assessment is a frequent and important part of the students’ experience so small biases could
have a large impact in the long run.
The significant time x SES x assessment method interaction, although unexpected,
indicates that the hypothesized creation of an SES performance gap by participants using
assessment for selection is stronger as participants spend more time on the study. We interpret
this effect as a consequence of the participants’ engagement in the study. It is possible that
those who quickly completed the study paid less attention to the instructions and the test and
were then less affected by the manipulations. The findings observed among those who spend
more time on the study are in line with the idea that evaluators asked to use a traditional,
Running head: SELECTION AND SOCIAL CLASS
!
22
normative form of assessment, artificially produce a performance gap that corresponds to the
existing status asymmetry more than evaluators who use a form of assessment more oriented
toward learning.
It is interesting to note that, when using assessment suited for selection, the mistakes
produced by low-SES students were judged in a more punitive way, as indicated by an
average decrease in grades of .19 points for every mistake made whereas making a mistake
resulted in a loss of .09 points for high-SES students. Ultimately, we observed a rate of low-
SES students below the pass threshold two times higher than the rate of high-SES students.
This effect is consistent with previous research showing that evaluators can redefine their
assessment criteria (i.e., what is a weakness or a strength) in a way that justifies
discriminatory decisions (Norton, Vandello, & Darley, 2004; Uhlmann & Cohen, 2005). For
our participants who used normative grading, mistakes became less of a weakness when
produced by a high- than a low-SES student. This finding further supports the idea that the
practice of grading may lead evaluators to restrain the success of low status students (see also
Batruch et al., 2017).
Supplementary analyses considered the impact of the participants’ competence (i.e.,
self-reported GPA) and social class. In all cases, the interaction between target’s SES,
assessment method and time remained significant and was not further moderated. This rules
out the idea that variations in competence could explain the results. Moreover, participants’
own social class did not affect their behavior toward the target, which suggests that the bias
against the lower SES students does not reflect an intergroup bias (Hewstone, Rubin, &
Willis, 2002). It thus seems that our work does not fall in the scope of intergroup feedback.
This literature showed that evaluators from a majority group do not communicate the same
praise and criticism to majority and minority students (Croft & Schmader, 2012; Crosby &
Monin, 2007). In our research, we hypothesized that participants endorsed their role of agents
Running head: SELECTION AND SOCIAL CLASS
!
23
of the educational institution and acted as such, beyond their own social identity. The results
supported this contention.
The unexpected interaction with the time spent on the task raises questions. The longer
time needed to use formative assessment might indicate the greater cognitive and motivational
costs of such a method, which requires one to identify and explain in simple words the rules
underlying each mistake and to think of ways to improve. It remains possible that, even after
accounting for time, the cost of formative assessment contributed to the lower number of
mistakes found by participants using this method. To rule out this interpretation we conducted
a second study in which we equalized the motivational and cognitive costs of the two
assessment methods.
Experiment 2
Method
Participants. A total of 269 students from a medium-size French-speaking Swiss
university voluntarily took part in the study. Data collection stopped at the end of the semester
considering that we achieved the minimum of 50 participants per cell of the 2 (SES) by 2
(Assessment Method) design. Data from 10 participants were excluded because they were
suspicious (N = 6) or failed the manipulation checks (N = 4) (see supplemental material for
analysis with the full sample). The final sample consisted of 163 women, 93 men, 3
unspecified (Mage = 21.55, SD age = 2.35). Each participant was randomly assigned to one of
the experimental conditions in the Assessment method (for selection vs. for learning) x
Target’s SES (low vs. high) between-participants design. The target’s sex was also
manipulated as a control and was not part of our hypotheses.
Material and procedure. Students were approached in university cafeterias by the
experimenter and asked to take part in a study about assessment tools used by teachers. The
procedure was similar to the one followed in Experiment 1. The main difference was in the
Running head: SELECTION AND SOCIAL CLASS
!
24
instructions about assessment. Participants in the assessment for selection condition read that
they would have to give a grade, while participants in the assessment for learning condition
would have to write formative comments. However, to equalize the motivational and
cognitive costs of the two assessment methods, and hopefully time spent on the task,
participants were asked to only underline the mistakes “for the time being”, and told that they
will be asked to give the grade or write the comments at a later stage (but were actually never
asked to do so). The dictation test was a slightly modified version of the one used in
Experiment 1 and contained 14 obvious mistakes (thus, 3 additional mistakes as compared
with Experiment 1) and 6 ambiguous mistakes. At the end of this task, participants moved on
to the following pages of the booklet. In the last section, participants answered the
manipulation checks and the socio-demographic questions6. Finally, participants were
thanked and debriefed.
Results
Perceived SES. Participants’ perception of the target’s SES was analyzed in a
regression with Assessment Method (assessment for selection coded -0.5, assessment for
learning coded 0.5), Target’s SES (low-SES coded -0.5, high-SES coded 0.5) and the
interaction term as predictors. The analysis was run on the sample that excluded suspicious
participants (N = 263, but 1 missing value). As expected, Target’s SES had a main effect on
ratings, b = 2.16, [1.97, 2.34], t(258) = 22.94, p < .001, η2p = .67, [.61, .72]. The low-SES
targets were perceived as coming from a more disadvantaged background (M = 3.87, SD =
.78, [3.74, 4.00]) than the high-SES targets (M = 6.03, SD = .75, [5.90, 6.16]). The main
effect of Assessment Method and the interaction between Assessment Method and Target’s
SES did not reach significance (b = -0.13, t(258) = -1.35, p = .18 and b = 0.07, t(258) = 0.38,
p = .71). We excluded 4 participants, mentioned in the Participant section, who did not
properly perceive the target’s SES.
Running head: SELECTION AND SOCIAL CLASS
!
25
Time to complete the study. An important goal of this study was to test the
hypothesis without the methodological problem related to the difference in time needed to
perform the two types of assessments that we observed in Experiment 1. We analyzed the
time taken by participants to complete the study in a regression including Assessment
Method, Target’s SES and the interaction as predictors7. The analysis indicated that
participants took a similar amount of time when they used grades (M = 13.51, SD = 3.29,
[12.93, 14.09]) and formative comments (M = 13.15, SD = 3.92, [12.47, 13.83]), b = -0.36,
t(254) = -0.78, p = .43. No SES main effect or interaction reached significance, b = -0.24,
t(254) = -0.54, p = .59 and b = 0.44, t(254) = 0.48, p = .63.
Number of mistakes. We analyzed the number of mistakes detected by the
participants in the dictation test in a regression with Assessment Method, Target’s SES and
the interaction as predictors. Results showed no main effect of Assessment Method, b = -0.01,
[-0.67, 0.65], t(255) = -0.03, p = .98, η2p = .00, [.00, .00] and a main effect of the Target’s
SES, indicating that again participants detected more mistakes in the dictation of low-SES
targets (M = 12.09, SD = 2.66, [11.62, 12.55]) than in that of high-SES targets (M = 11.29, SD
= 2.674, [10.83, 11.75]), b = -.80, [-1.46, -0.14], t(255) = -2.39, p = .02, η2p = .02, [.00, .07].
The predicted Target’s SES x Assessment Method interaction was significant, b = 1.37, [0.06,
2.68], t(255) = 2.05, p = .04, η2p = .02, [.00, .06], Cohen’s d = .26. As shown in Figure 3,
when participants used assessment for selection, they found a greater number of mistakes in
the dictation attributed to low-SES students (M = 12.43, SD = 2.86, [11.77, 13.09]) compared
with high-SES students (M = 10.95, SD = 2.69, [10.29, 11.61]), b = -1.48, [-2.42, -0.55],
t(255) = -3.12, p = .002, η2p = .04, [.01, .09]. This difference was not significant when
participants used assessment for learning (Mlow-SES = 11.74, SDlow-SES = 2.41, [11.09, 12.39];
Mhigh-SES = 11.63, SDhigh-SES = 2.77, [10.97, 12.28]), b = -0.11, [-1.04, 0.81], t(255) = -0.23, p =
Running head: SELECTION AND SOCIAL CLASS
!
26
.81, η2p = .00, [.00, .02]. The participants’ self-reported GPA and social class did not account
for, or moderate these results (see supplemental material for detailed analyses).
Discussion
This study intended to test our hypothesis without the interference of the effect of time.
To do so, we asked all participants to underline the mistakes in the dictation test, and only
mentioned that they would write the formative comments or provide a grade (depending on
the condition) at a later stage. Results showed that after equalizing the motivational and
cognitive costs of assessment, participants took approximately the same amount of time to
complete the task, regardless of the condition. As in Experiment 1, participants found a
greater number of mistakes in the dictation tests of low- as compared with high-SES students.
More importantly, the predicted Assessment Method x Target’s SES interaction was
significant. As expected, a significant social class achievement gap was artificially produced
by participants who used assessment for selection, but not by participants who used
assessment for learning. Participants who used assessment for selection reported on average
1.48 more mistakes in the test of low-SES students than in the test of high-SES students.
Again, the participants’ characteristics did not moderate this effect.
In the theoretical development of our hypothesis, we argued that the impact of
normative grading on the creation of social class inequalities is due to the fact that this
method epitomizes the function of selection of educational institutions. This assumption was
grounded in research showing that, from the perspective of students, assessment oriented
toward selection triggered a greater SES performance gap than assessment for learning
(Smeding et al., 2013) and that, from the perspective of evaluators, adherence to the function
of selection related to more support for grading (Autin et al., 2015). We conducted a third
experiment to directly test the hypothesis that the selective purposes of assessment, usually
conveyed by normative grading, is indeed what underlies evaluators’ tendency to artificially
Running head: SELECTION AND SOCIAL CLASS
!
27
produce performance differences based on students’ SES, whatever the actual form
assessment takes.
Experiment 3
To test this hypothesis, we manipulated the function of assessment, to induce either
selective or educational purposes. We expected that evaluators would create a social class
achievement gap when assessment is framed as a way to select the best students more than
when it is presented as a tool to improve learning. This hypothesis could lead to two possible
effects: The function of selection may potentiate the effect of normative grading in the
creation of the social class gap and the educational function may potentiate the egalitarian
effect of formative comments. In this case, a Function of Assessment x Assessment Method x
Target’s SES interaction should emerge. However, since we conceptualize assessment
methods as tools to fulfill a specific institutional purpose it is also possible that the function of
assessment overrides the effect of assessment tools (i.e., grading vs. formative comments). In
this case, a Function of Assessment x Target’s SES interaction should emerge.
Method
Participants. A total of 501 students from a medium-size French-speaking Swiss
university voluntarily took part in the study in the cafeterias on campus or in class (data were
collected in several classes, resulting in a field-related diversity similar to the data collected in
cafeterias). We decided to double the sample size, and data collection was contingent on the
class we had access to and the end of the semester. Each participant was randomly assigned to
one of the experimental conditions in the Assessment Method (assessment for selection vs.
assessment for learning) x Target’s SES (low vs. high) x Function of Assessment (selection
vs. education) between-participants design. To avoid increasing the complexity of the
experimental design, and as target’s sex was not a variable of interest, this factor was not
included in the present design; we only used boys as targets. Data from 10 participants were
Running head: SELECTION AND SOCIAL CLASS
!
28
excluded because they expressed suspicion, were unable to assess the test or were not
students. Data from 117 participants were excluded because they failed the manipulation
checks regarding the Target’s SES (N = 37), the Function of Assessment (N = 74) or both (N
= 6) (see supplemental material for analysis with the full sample). The number of participants
per condition ranged from 40 to 52. We believe this high number of failures can be explained
by the lack of involvement from students in the collective sessions in class, a recruitment
method that we did not use in the two previous experiments. Even though they were explicitly
asked to carefully read the instructions, failure on the manipulations checks indicate that they
did not read the main instructions or that they refused to comply with them. More
importantly, including participants who were unable to accurately report the function of
assessment might prevent us from properly testing the main hypothesis of this study: the
underlying role of this structural factor in the creation of a SES gap. We thus considered that
the validity of these data was questionable and that including them would increase noise
(Oppenheimer, Meyvis, & Davidenko, 2009). The final sample of 374 students consisted of
212 female, 153 male, 9 unspecified (Mage = 22.39, SDage = 2.65).
Material and procedure. The procedure was similar to the one followed in
Experiment 2. After reading about the assessment method they have to use, and the profile of
the target, participants were required to assess the dictation test. At the top of the page
containing the dictation test, we presented a reminder of the assessment method and the
specific instructions about the function of the assessment. Participants in the selection
condition read that the mistakes they would find in the test would eventually help them to
decide whether the student should move to next grade or not. Participants in the education
condition read that the mistakes they would find in the test would help them to propose
learning strategies that allow the student to improve. After they underlined the mistakes in the
Running head: SELECTION AND SOCIAL CLASS
!
29
test, participants answered the manipulation checks and the socio-demographic questions8.
Finally, participants were thanked and debriefed.
Results
Perceived SES. Perception of the students’ socio-economic background was analyzed
in a regression with Assessment Method (assessment for selection coded -0.5, assessment for
learning coded 0.5), Target’s SES (low-SES coded -0.5, high-SES coded 0.5), Function of
Assessment (selection coded -0.5, education coded 0.5) and all interactions as predictors. The
analysis was conducted on the sample of non-suspicious participants (N = 491, but 3 missing
values). The Target’s SES influenced the perception of socio-economic background in the
expected direction (MlowSES = 3.88, SDlowSES = .92, [3.77, 3.99]; MhighSES = 5.97, SDhighSES =
.79, [5.86, 6.08]), b = 2.09, [1.93, 2.24], t(480) = 26.49, p < .001, η2p = .59, [.54, .64]. No
other effects reached significance (ts < 1.60, ps > .11). Among the participants excluded, 37
were taken out from the final sample because they did not correctly report the target’s socio-
economic background.
Number of mistakes. The number of mistakes found in the test was analyzed in a
regression with Assessment Method, Target’s SES, Function of Assessment and all
interactions as predictors9. The analysis revealed a significant interaction between the
Function of Assessment and the Target’s SES, b = 1.34, [0.12, 2.56], t(365) = 2.16, p = .03,
η2p = .01, [.00, .04] , Cohen’s d = .20. As shown in Figure 4, when participants thought the
assessment was aimed at selecting the students, they found more mistakes in the test of a low-
SES students (M = 12.71, SD = 2.75, [12.07, 13.35]) than in the test of a high-SES student (M
= 11.59, SD = 2.94, [10.98, 12.20]), b = -1.12, [-2.00, -0.24], t(365) = -2.50, p = .01, η2p = .02,
[.00, .05]. This social class gap was not significant when the assessment was presented with
an educational purpose, (MlowSES = 11.44, SDlowSES = 3.35, [10.84, 12.05]; Mhigh SES = 11.66,
SDhigh SES = 2.84, [11.08, 12.14]), b = 0.22, [-0.62, 1.06], t(365) = 0.51, p = .61, η2p = .00, [.00,
Running head: SELECTION AND SOCIAL CLASS
!
30
.02]. The three-way interaction between Assessment Method, Function of Assessment and
Target’s SES did not reach significance, b = 1.72, [-0.72, 4.15], t(365) = 1.38, p = .17, η2p =
.01, [.00, .03]. These results were not impacted or further moderated by the participant’s level
of competence or social class (see supplemental material).
Discussion
This study sought to test the hypothesis that the selective (rather that educational) role
of assessment is the mechanism that leads evaluators to create a performance gap that
corresponds to existing status asymmetries in the absence of actual differences in
performance. Results supported our hypothesis: The Function of Assessment x Target’s SES
interaction revealed that when assessment was presented as a way to select the best students,
evaluators found on average 1.12 more mistakes in a dictation supposedly produced by a low-
SES student than in that attributed to a high-SES student. This artificial performance gap was
reduced when assessment was presented as a way to help students improve. The results did
not show a significant moderating effect of the assessment method (i.e., assessment for
selection vs. learning). These findings suggest that it is not so much normative grading and
formative comments per se that lead evaluators to respectively create or not a social class
performance gap, but rather the function attributed to the assessment tools.
However, a high number of participants had to be excluded in particular for not
properly reporting the function of assessment. This raises concerns about the design and
suggests a possible conflict between instructions about the assessment method and the
function of assessment. For example, it might have been difficult for participants to
understand or comply with the instruction of both looking for mistakes to help improve
learning and decide whether the student should move to the next grade. This is actually not
surprising if we consider that the function of selection is positively associated with support for
grading whereas negatively associated with support for formative comments (Autin et al.,
Running head: SELECTION AND SOCIAL CLASS
!
31
2015). Contradicting the usual associations might have led to unpredictable consequences on
the participants’ behavior.
Therefore, we designed a fourth study to test the hypothesis that it is the function of
assessment that triggers or not the creation of a SES performance gap. To avoid confusion
between the assessment tools and their functions, we kept the assessment tool constant.
Because of the link between normative grading and a selective function and between
formative comments and educational functions, we decided not to use these assessment
methods. Rather, we relied on a less common procedure based on the highlighting of sections
of the student’s work. Participants had to highlight in two different colors the positive and
negative aspects of an essay (see Croft & Schmader, 2012, for a similar design). The fourth
study thus aims to replicate the previously observed findings on a different measure of
evaluation.
Moreover, this assessment tool provides information about both positive and negative
feedback, which could shed light on how the SES performance gap is created. Indeed, in the
previous studies, we observed the creation of a difference between the low- and high-SES
student but could not definitely determine whether this difference resulted from negative
behavior against the low-SES student or advantage given to the high-SES student. Indeed,
inequalities were traditionally framed as the product of discrimination, bias against low status
groups but they actually also result from favoritism, bias for high status groups (e.g., Adams
et al., 2008). Some even argue that in societies where hostility toward low status groups and
intergroup conflicts are not acceptable, favoritism is more prevalent (Brewer, 1999;
DiTomaso, 2015; Greenwald & Pettigrew, 2014). Disentangling the processes at play behind
the creation of the social class achievement gap is not the central question addressed in the
present paper, yet we could expect that when evaluating an essay with a selective purpose, if
participants discriminate against low-SES students, they will provide more negative feedback
Running head: SELECTION AND SOCIAL CLASS
!
32
to this student compared to a high-SES student. If participants favor high-SES students, they
might provide more positive feedback to this student compared to a low-SES student. When
assessment is used with an educational purpose, these differences should be attenuated.
Finally, the fourth study investigated a potential conflation between the educational vs.
selection function of assessment and a growth vs. fixed mindset. A sizable literature has
shown that individuals can adopt a growth mindset that refers to the belief that one’s qualities
are malleable and expandable through learning, or a fixed mindset that corresponds to the
belief that qualities are unchangeable (Dweck, 2012). Rattan, Good and Dweck (2012)
showed that instructors with a fixed theory of intelligence, compared to a malleable theory,
attribute low ability to low-performing students and give them less engaging feedback.
Because the induction of the educational function focuses evaluators on improvement and
learning, it might be associated with a growth mindset. Conversely, the function of selection
focuses evaluators on the student’s stance relative to the requirement and might relate to a
fixed mindset. We included a measure of the evaluators’ perception of the malleability of
students’ intelligence to test whether the function of assessment affects their mindset.
Experiment 4
Method
Participants. A total of 335 students in a French university participated in the study,
in exchange for course credit (N = 227) or were recruited in a university library (N = 108). We
aimed for at least 50 participants per cell; as we anticipated attrition, we oversampled.
Twenty-eight participants were excluded for not being able to report the function of
assessment (N = 7), the SES of the target (N = 19) or both (N = 3) (see supplemental material
for analyses on the full sample). The final 306 participants (246 women, 57 men, 3
unspecified, Mage = 19.69, SD = 3.80) were randomly assigned to one of the experimental
Running head: SELECTION AND SOCIAL CLASS
!
33
conditions in the Function of Assessment (selection vs. education) x Target’s SES (low vs.
high) between-participants design.
Material and procedure. Participants had to imagine that they were a history teacher
who has to assess an essay produced by an 8th grade male student using a new assessment
tool. They had to highlight in one color (yellow) the parts of the essay that were well written
(i.e., clear, logical, important for the structure of the text) and in another color (orange) the
parts that needed to be revised (i.e., unclear, misplaced regarding the logical organization of
the text, spelling, syntax or grammar errors). An example was provided.
Manipulation of the function of assessment. To emphasize the function of selection,
half of the participants read that their evaluation of the student’s skills counted toward his
semester GPA. Evaluating the essay would give them information to decide whether the
student should move to the next grade or not by identifying his strengths and weaknesses in
this kind of exercise. To focus the other half of the participants on the educational function,
they read that their evaluation of the student’s skills was part of a learning program.
Evaluating the essay would give them information to help the student improve his learning by
identifying strategies to make progress in this kind of exercise.
Manipulation of the student’s SES. Participants were then asked to read the file of the
student who supposedly produced the essay. The files were similar to the ones used in
Experiment 3 but adapted to the French context. The target’s SES was manipulated by
changing the student’s name, parental occupation and extracurricular activities.
Implicit theories of intelligence. Eight items were adapted from Souchal and Toczek
(2010) to measure participants’ conception of students’ intelligence. Four items referred to an
entity theory (e.g., “ Students have a certain level of intelligence and no matter what they do,
it can not change”) and 4 to an incremental theory (e.g., “Students’ intelligence grows with
every new experience they live”). A factor analysis revealed one factor (value = 2.69, 33.6%
Running head: SELECTION AND SOCIAL CLASS
!
34
of explained variance) including the four entity items and 2 incremental items (reversed). We
then computed a score of entity by averaging the scores on these items (Cronbach’s α = .78).
After completing the questionnaire, participants assessed the essay. They were briefly
reminded of the Function of Assessment and of the instructions regarding the use of
highlighters to provide positive and negative feedback. The essay was a picture of a 20 line-
long handwritten text inspired from actual essays. The number of characters highlighted in
each color was computed as indicators of the quantity of positive (yellow) and negative
(orange) feedback.
After the assessment of the essay10, participants reported the function of the
assessment, two types of information presented in the student file and estimated his
background on a 7-point scale (1 “highly disadvantaged” to 7 “highly advantaged”). They
provided socio-demographic information (age, sex, parental level of education and
occupations), were thanked and debriefed.
Results.
Perceived SES. Participants’ perception of the target’s SES was analyzed on the full
sample in a regression with the Function of Assessment (selection coded -0.5, education
coded 0.5), Target’s SES (low-SES coded -0.5, high-SES coded 0.5), and the interaction term
as predictors11. Results showed a main effect of Target’s SES, b = 1.75, [1.56, 1.94], t(328) =
18.34, p < .001, η2p = .51, [.43, .57]. The low-SES target was perceived as coming from a
more disadvantaged background (M = 3.98, SD = 0.86, [3.84, 4.11]) than the high-SES target
(M = 5.72, SD = .88, [5.59, 5.86]). The Function main effect and interaction did not reach
significance, b = -0.09, t(328) = -0.91, p = .36 and b = -0.09, t(328) = 1.13, p = .26.
Ratio of negative feedback. We computed the number of characters highlighted in
each color as indicators of negative and positive feedback. We calculated a ratio of negative
feedback relative to the total amount of positive and negative feedback such that higher scores
Running head: SELECTION AND SOCIAL CLASS
!
35
indicate more negativity in the evaluation. This ratio was created to have a negative
evaluation indicator that is comparable to evaluation in our previous studies (i.e. finding
mistakes in a test). This ratio was analyzed in a regression with the Function of Assessment
(selection coded -0.5, education coded 0.5), Target’s SES (low-SES coded -0.5, high-SES
coded 0.5), and the interaction as predictors12. The results showed no main effect of the
Target’s SES, b = -0.03, [-0.05, 0.00], t(300) = -1.82, p = .07, η2p = .01, [.00, .05] or the
Function of Assessment, b = -0.004, [-0.03, 0.02], t(300) = -0.29, p = .77, η2p = .00, [.00, .02].
The expected interaction between SES and Function did not reach significance, b = 0.03, [-
0.03, 0.08], t(300) = 0.91, p = .36, η2p = .00, [.00, .03] , Cohen’s d = .10. However, this
interaction was in the expected direction with a larger difference in ratio between low and
high SES students when the assessment was meant to select (M low SES = .42, SD low SES = .12,
[.39, .44] vs. M high SES = .38, SD high SES = .13, [.35, .41]) rather than to improve learning (M
low SES = .40, SD low SES = .12, [.38, .43] vs. M high SES = .39, SD high SES = .13, [.36, .42]).
Positive and negative feedback. The number of characters highlighted was analyzed
in a 2 (Function of Assessment: selection vs. education) X 2 (Target’s SES: high vs. low) X 2
(Type of Feedback: negative vs. positive) mixed ANOVA with the last factor as a within-
participant factor13. The analysis revealed a main effect of the Type of feedback, F (1, 299) =
189.34, p < .001, η2p = .39, indicating that participants gave more positive feedback (M =
418, SD = 189) than negative feedback (M = 272, SD = 133). This effect was qualified by an
interaction with the Target’s SES, F (1, 299) = 7.05, p = .008, η2p = .023, 90 % [.00, .06].
Participants gave more positive feedback to a high SES student (M = 446, SD = 204, [413,
479]) compared to a low SES student (M = 392, SD = 171, [365, 419]), F (1, 299) = 5.98, p =
.015, η2p = .02, [.00, .05]. This difference between SES was not significant for negative
feedback, F(299) = .38, p = .54, η2p = .00, [.00, .01]. No other effect reached significance Fs
Running head: SELECTION AND SOCIAL CLASS
!
36
< 2.53, ps > .11, ηs2p < .01, including the expected interaction between the Type of feedback,
SES and the Function of Assessment, F(299) = 0.03, p = .86, η2p = .00, [.00, .01].
Entity theory of intelligence. The score of belief in an entity theory of intelligence
was analyzed in a regression with the Function of Assessment (selection coded -0.5,
education coded 0.5), Target’s SES (low-SES coded -0.5, high-SES coded 0.5), and the
interaction as predictors14. The main effect of Function that would indicate an impact of that
induction on mindset did not reach significance b = 0.02, [-0.14, 0.18], t(298) = 0.25, p = .80,
η2p = .00, [.00, .01]. The main effect of SES and the interaction were also non-significant,
respectively b = 0.05, [-0.12, 0.21], t(298) = 0.55, p = .59, η2p = .00, [.00, .02] and b = -0.23,
[-0.55, 0.10], t(298) = -1.36, p = .17, η2p = .01, [.00, .04].
Supplementary analyses investigating the impact of the participants’ own social class
were conducted (see supplemental material for the details) and showed no change or
moderation of the described results.
Discussion
This fourth study first aimed at testing the hypothesis that the function of selection,
while keeping the assessment tool constant, triggers the creation of a SES performance gap,
compared to the educational function. The analysis on the ratio of negative feedback showed
a pattern that was congruent with this hypothesis, but the effect was not significant. The
results on the number of characters highlighted showed an overall favoritism of the higher
social class student. Irrespective of the function of assessment, participants provided more
positive feedback to the high-SES student than the low-SES student. This result is in line with
the idea that nowadays the creation of inequalities relies on a favorable bias for high status
group members who are offered more positive experience (e.g., DiTomaso, 2015). However,
further replication of this effect is needed as we initially predicted that it would appear only
when the context emphasizes on selection. Moreover, a possible way to disentangle favoritism
Running head: SELECTION AND SOCIAL CLASS
!
37
toward the high-SES student from negative treatment toward the low-SES student could be to
include a control condition with no information about the target’s SES (i.e., anonymous). This
would provide information about whether it is the low-SES condition that triggers more
negative assessment than the control or the high-SES condition that triggers more positive
assessment, or whether both discrimination and favoritism are at play.
A secondary goal of this study was to examine whether the function of assessment
could impact the evaluators’ mindset, with selective purposes fostering a more entitative
theory of the student’s abilities than educational purposes. The results are not in line with this
proposition. Previous research showing changes in mindset used direct intervention by telling
participants that intelligence is fixed or can grow (Rattan, Savani, Chugh, & Dweck, 2015). It
could be that information about the function of assessment is not sufficiently powerful to
affect the mindset. Yet, some studies suggest that mindsets are also sensitive to subtle
information such as praise or generic statements about categories (i.e., talking about boys in
general instead of a boy in particular) (Cimpian & Markman, 2011; Mueller & Dweck, 1998).
At this stage, the impact of the function of assessment on the belief in an entity theory of
intelligence remains an open question, although the effect size (η2p = .00, [.00, .01]) could
suggest a possibly negligible effect.
Finally, it should be recognized that the sample of this experiment presents an
imbalance in terms of gender and recruitment location. Such an imbalance makes it difficult
to test the effects of these variables, but as the personal characteristics of the participants do
not seem to alter the observed effects (see Supplemental Materials), we believe that this
asymmetry should not be a source of concern.
Meta-analysis
To understand more precisely the size of the effect of interest, we ran an internal meta-
analysis on the four experiments (Cumming, 2013) and estimated the effect size of the
Running head: SELECTION AND SOCIAL CLASS
!
38
moderation of the SES performance gap by the orientation of assessment toward selection
(i.e., grading or selection function) or education (i.e., formative comments or educational
function). We computed the standardized mean difference corresponding to the difference of
simple effects of SES between the selection and education assessment practices
[(X
̅lowSES_selection - X
̅highSES_selection) – (X
̅lowSES_education - X
̅highSES_
education)]/2*sp (where sp is the pooled SD) (Westfall, 2015). In Experiment 1, we used the
effect size of the SES x Assessment Method interaction at a moderate time spent on the study.
In Experiment 4, we used the effect size of the SES x Function of Assessment interaction on
the ratio of negative feedback, as this measure is the functional equivalent to the number of
mistakes measured in the previous three studies. We used a weighted random-effects model
(Cumming, 2013). A weighted model lowers the contribution of studies with higher variance
around the effect size. Random effects models take into account the heterogeneity between
studies and postulates that different studies can estimate different effect sizes.
The analysis revealed a small and significant effect size, d = 0.19, p = .002, [0.07;
0.30]. The variance index between the four studies was not significant, suggesting low
heterogeneity between studies Q(df = 3) = 0.84, p = .84. This internal meta-analysis provides
evidence that evaluators artificially create a greater SES performance gap when assessment is
used to select rather than foster learning. The effect size is small but we nonetheless believe it
should be interpreted in light of the length of education and the frequency of assessment. Very
small differences in repeated evaluations can have important consequences on the overall
experiences and educational outcomes of students when they accumulate over time.
General Discussion
A growing line of research has addressed the question of the cultural and structural
determinants underlying the social class achievement gap (e.g., Croizet & Claire, 1998;
Stephens et al., 2012). This endeavor has been particularly valuable in revealing the
Running head: SELECTION AND SOCIAL CLASS
!
39
sociocultural influences that contribute to the social class inequalities. However, the majority
of these studies have focused on the psychological processes (e.g., stereotype threat, cultural
mismatch) that impact the academic performance of students. In the present research, we
argue that, in addition, a new stream of research should emerge that addresses how
evaluators’ behavior contributes to the social class achievement gap, independent of the
students’ actual performance. We proposed that the endemic use of normative grading in
education, given its strong association with the meritocratic ideal and the function of selection
of educational institutions, leads evaluators to reproduce existing social class asymmetries. In
contrast, assessment with an educational function and an egalitarian ethos should reduce the
impact of the student’s social class on evaluation. More specifically, we hypothesized that
evaluators would differentially assess the work produced by low- and high-SES students
when using assessment for selection, even in the absence of any objective differences. This
tendency should be reduced when using assessment for learning.
Experiments 1 and 2 showed consistent evidence that when evaluators used an
assessment method oriented toward selection (i.e., normative grading; cf. Autin et al., 2015),
they actively detected more mistakes for low-SES students than for high-SES students. This
effect emerged in a dictation test that objectively contained the same number of mistakes in
all conditions (with a moderation by the time spent on the task in Experiment 1). The creation
of such an artificial social class achievement gap was not observed when evaluators used an
assessment method oriented toward education (i.e., formative comments). We believe that a
strong asset of these results is the use of a behavioral measure – the number of mistakes that
participants actually found in the test – that did not allow participants to control the social
desirability of their responses (Dompnier, Darnon, & Butera, 2009).
Experiments 3 and 4 manipulated the mechanism that we assumed to explain the
results observed in the first two experiments: the function of assessment. Indeed, we expected
Running head: SELECTION AND SOCIAL CLASS
!
40
that evaluators reproduce in their assessment existing social class asymmetries when using
normative grading because this form of assessment epitomizes the function of selection of
educational institutions. The results support this hypothesis. In Experiment 3, making the
selection function of assessment salient led evaluators to find significantly more mistakes for
low-SES students than for high-SES students, regardless of the assessment method they used.
Such a differential treatment was no longer significant when the educational function of
assessment was made salient. In Experiment 4, the replication of this finding with a different
assessment tool (i.e., highlighting sections in the student’s essay) showed a consistent but not
significant pattern in the ratio of negative feedback. And accordingly, we conducted a small-
scale meta-analysis to test the overall support to our main hypothesis received from the four
studies. Overall, the results of the meta-analysis support the hypothesis that even in the
absence of objective differences in performance, social class inequalities can be perpetuated
by evaluators who re-create an achievement gap, especially when the selective purpose of
assessment is put to the fore.
Experiment 4 secondarily aimed at specifying how the social class achievement gap
was created in the selection context, by using both negative and positive feedback. This
question goes beyond the scope of the present article, as the primary goal here was to
document the creation of a SES performance gap. Yet, we postulated that inequalities are the
byproduct not only of negative treatment against low status individuals but also of the
privilege of high status individuals (e.g., Adams et al., 2008). The results showed an overall
effect of favoritism toward high-SES students. This unexpected effect calls for further
investigation but is consistent with an understanding of the mechanisms of inequalities that
emphasizes the implication of favoritism. For example, it has been observed that high-status
individuals receive advantages, and especially better evaluation from both high- and low-
status actors (DiTomaso, Post, Smith, Farris, & Cordero, 2007). Previous research
Running head: SELECTION AND SOCIAL CLASS
!
41
documented how higher social class students benefit from many aspects of the educational
institutions such as valued forms of knowledge, language or posture, compatible self-models
and boosting evaluative settings (Bourdieu & Passeron, 1977; Croizet et al., 2017; Lareau,
2011; Stephens, Markus, et al., 2014). The present results suggest that evaluator’s behavior
during assessment might also be one of the privileges that enhance the academic experience
of higher social class students.
Contribution to Ongoing Debates
The first contribution of the present research is to participate in the growing effort to
bring the study of social class to the core of social and educational psychological
investigations (S. T. Fiske & Markus, 2012). Alongside previous research directly studying
student performance (e.g., Goudeau & Croizet, 2017; Jury et al., 2015), the present article
unveils a new path through which social class inequalities are reproduced in schools via the
assessment of performance. Our research suggests that, even if the educational system could
offer a matching and non-threatening environment to all students, evaluators could still
artificially create a social class achievement gap when they assess with selective purposes.
The second contribution pertains to research in sociology of education. Observing that
evaluators create a social class achievement gap is in line with the classic sociological theory
of social reproduction, and in particular with the proposition that agents of institutions play an
important role in the reproduction of social inequalities (Bourdieu & Passeron, 1977). Our
research provides experimental evidence of the evaluators’ role in social reproduction, and
more importantly, identified an institutional factor that leads evaluators to create social
inequalities: the selection vs. educational function of assessment.
Finally, our findings are consistent with previous research in educational sciences
showing discrimination by evaluators in assessing the same product attributed to students of
different backgrounds (e.g., Sprietsma, 2013). However, the existing literature had only
Running head: SELECTION AND SOCIAL CLASS
!
42
investigated this phenomenon in settings using grading, and our comparison with alternative
forms of assessment offers new insights. Not only because this comparison shows that
discriminatory behavior is not inherent to assessment, but also—and especially—because it
shows that evaluators do not always act in a biased manner. Regarding the discriminatory
behavior previously observed in grading, our results suggest that it would be better interpreted
as the product of the selective purposes conveyed by such assessment practices, rather than
biased individual evaluators. This result is consistent with qualitative work showing that
changing the tools (e.g., replacing grades with formative comments) is not enough to change
the vision of assessment, and that ultimately teachers use all forms of assessment primarily
with quantifying and ranking purposes (McNair, Bhargava, Adams, Edgerton, & Kypros,
2003). Moreover, teacher’s use of assessment has been related to the requirement of
educational institutions stemming from their societal role of selection (Gewirtz, 2000; Hall,
Collins, Benjamin, Nind, & Sheehy, 2004; Popham, 2001). Our research concurs with an
analysis of assessment practices as contingent on the institutional function they serve.
Through the prism of selection, all forms of assessment might produce inequalities in
evaluation.
Overall, the present research underscore the need for a sociocultural approach to
understanding inequalities (Adams et al., 2008; Markus & Stephens, 2017) that highlights the
intertwining of individuals with institutions in their production (see also Kraus & Park, 2017).
Social class inequalities cannot be reduced to consequences of direct and intentional actions
of biased individuals (i.e., biased evaluators), or to byproducts of agentless institutions that
mechanically exclude lower social class students and favor higher social class students (i.e.,
biased schools). We rather propose that the educational institutions’ logic shapes evaluators’
behavior, and in turn low- and high-SES students’ experience. More specifically, the
functions of educational institutions seem to affect the way evaluators make sense of the
Running head: SELECTION AND SOCIAL CLASS
!
43
assessment situation and cause them to differentially evaluate students’ performance based on
their social class.
Limitations and Conclusion
Several limitations of the present research should be acknowledged. First, the studies
were conducted with students who were put in the position of a teacher, and not real teachers.
Replicating these findings with teachers would certainly increase their ecological validity, but
we do not expect any remarkable difference. Indeed, our theoretical approach is precisely that
institutional norms and functions shape their agents’ behaviors; thus, the observed effects
should be reproducible with actual teachers, as they have been socialized in the very context
that we have experimentally induced in this research. Furthermore, previous research using
role-playing paradigms showed that participants adjust their attitudes to the role (Covington &
Omelich, 1979; Harari & Covington, 1981; Houston & Holmes, 1975), and our results are
consistent with those obtained in research conducted with teachers (e.g., Hinnerich, Höglin, &
Johannesson, 2015; Rangvid, 2015, Sprietsma, 2013).
Second, more research is needed to understand the psychological mechanisms at play
in the discriminatory behavior of evaluators. The present research proposed a sociocultural
approach and therefore, in the last two experiments, we manipulated a structural-level
mechanism (i.e., the function of selection vs. education) believed to underlie the creation of
the social class achievement gap. However, future research may also be interested in
individual-level variables induced by both functions of assessment. We argued that
assessment for selection relates to a meritocratic ethos whereas the assessment for learning
relates to an egalitarian ethos (Autin et al., 2015). Egalitarianism and meritocratic values have
contrasted consequences in terms of stereotyping and attitudes towards groups (e.g., Wyer,
2003) and are involved in the reduction/maintenance of inequalities (e.g., Costa-Lopes,
Dovidio, Pereira, & Joste, 2013). The perception of egalitarian vs. meritocratic values is
Running head: SELECTION AND SOCIAL CLASS
!
44
therefore a possible mechanism underlying the effect of assessment for learning vs. selection
on the creation of a social class performance gap. Downstream mechanisms could also be
explored. For example, assessment as an apparatus of meritocratic selection might give
participants a greater sense of objectivity than assessment to foster learning. Feelings of
objectivity are known to increase bias in decisions (Uhlmann & Cohen, 2007). Assessment
for selection also requires a more firm, non-ambiguous response (to give a grade, to decide
about grade repetition) than assessment for learning. The desire for clear-cut answer – known
as need for closure – relates to biases in thinking (Webster & Kruglansky, 1998), and high
need for closure can be involved in discrimination in grading (Kruglansky & Freund, 1983). It
is therefore possible that the discriminatory behavior observed in the present studies is partly
due to the higher need for closure triggered by the selection context compared to the
educational context. It is also possible that different processes simultaneously occur when
evaluators focus on educational purposes. For example, participants might feel more
accountable for their evaluation when told that they have to write formative comments or
suggest learning strategies to improve, and accountability reduces bias in decisions (Lerner &
Tetlock, 1999; Uhlmann & Cohen, 2007).
Finally, in the absence of a control condition for the function of assessment, it is
impossible to conclude from the present experiments whether the effects are due to the
assessment for selection conditions or to assessment for leaning conditions. However, in the
current sociocultural and educational context it seems unlikely to have a control assessment
condition that doesn’t conjure an institutional function. Given that normative grading is by far
the most widespread assessment method in OECD countries (Knight & Yorke, 2003), and that
the social class achievement gap is also omnipresent in these countries (OECD, 2013a), we
believe that an “ecological” control condition (e.g., “please, assess this dictation test”) would
probably be interpreted as the assessment for selection conditions. Moreover, creating a
Running head: SELECTION AND SOCIAL CLASS
!
45
“neutral” control condition (without any form of assessment; e.g., “please, find all the
mistakes in this dictation to prove your skills”) would be pointless, as in such a condition the
participant would no longer be an evaluator. The difficulty to have a control condition
explains why, as the majority of research investigating the factors contributing to inequalities
in education (see Jury et al., 2017; Croizet et al., 2017), we compared a situation where these
factors are at play to a situation where they are actively countered.
To conclude, this line of research challenges the idea that educational institutions
perform a meritocratic selection based solely on an objective assessment of individuals’
qualities. To the contrary, this selection function might actually contribute to the reproduction
of social inequalities by leading evaluators to create a social class achievement gap. There is a
growing literature demonstrating how the way educational institutions operate is involved in
the reproduction of social inequalities (e.g., Croizet et al., 2017; Jury et al., 2017). Our
research adds to this literature by identifying a structural factor –the function of selection of
assessment– that shapes the role of evaluators in the creation of a social class achievement
gap, even when there are no actual differences in performance.
Running head: SELECTION AND SOCIAL CLASS
!
46
References
Adams, G., Biernat, M., Branscombe, N. R., Crandall, C. S., & Wrightsman, L. S. (2008).
Beyond prejudice: Toward a sociocultural psychology of racism and oppression. In G.
Adams, M. Biernat, N. R. Branscombe, C. S. Crandall, & L. S. Wrightsman (Eds.),
Commemorating Brown: The social psychology of racism and discrimination. (pp.
215–246). Washington, DC: American Psychological Association.
Arrow, K. J. (1973). Higher education as a filter. Journal of Public Economics, 2, 193–216.
Autin, F., Batruch, A., & Butera, F. (2015). Social justice in education: How the function of
selection in educational institutions predicts support for (non)egalitarian assessment
practices. Frontiers in Psychology, 6, 707. https://doi.org/10.3389/fpsyg.2015.00707
Batruch, A., Autin, F., & Butera, F. (2017). Re-establishing the social-class order: Restorative
reactions against high-achieving, low-SES pupils. Journal of Social Issues, 73, 42–60.
https://doi.org/10.1111/josi.12203
Batruch, A., Autin, F., Bataillard, F., & Butera, F. (in press). School selection and the social
class divide: How tracking contributes to the reproduction of inequalities. Personality
and Social Psychology Bulletin.
Bell, D. (1973). The Coming Of Post-industrial Society. London, UK: Heinemann.
Bennett, R. E. (2011). Formative assessment: a critical review. Assessment in Education:
Principles, Policy & Practice, 18, 5–25.
https://doi.org/10.1080/0969594X.2010.513678
Bisseret, N. (1974). L’idéologie des aptitudes naturelles. In N. Bisseret (Ed.), Les inégaux ou
la sélection universitaire (pp. 13–52). Paris, France: Presses universitaires de France.
Black, P., & Wiliam, D. (1998). Assessment and classroom learning. Assessment in
Education: Principles, Policy & Practice, 5, 7–74.
https://doi.org/10.1080/0969595980050102
Running head: SELECTION AND SOCIAL CLASS
!
47
Bloom, B. S. (1968). Learning for mastery. Evaluation Comment, 1, 1–5.
Bloom, B. S., Hastings, J. T., & Madaus, G. F. (1971). Handbook on formative and
summative evaluation of student learning. McGraw-Hill.
Bobocel, D. R., Son Hing, L. S., Davey, L. M., Stanley, D. J., & Zanna, M. P. (1998). Justice-
based opposition to social policies: Is it genuine? Journal of Personality and Social
Psychology, 75, 653–669.
Bourdieu, P., & Passeron, J. (1977). Reproduction in education, culture and society. London,
UK: Sage.
Bowen, W. G., Kurzweil, M. A., Tobin, E. M., & Pichler, S. C. (2005). Equity and excellence
in american higher education. Charlottesville, VA: University of Virginia Press.
Brewer, M. B. (1999). The psychology of prejudice: Ingroup love and outgroup hate? Journal
of Social Issues, 55, 429–444. https://doi.org/10.1111/0022-4537.00126
Butler, R. (1987). Task-involving and ego-involving properties of evaluation: Effects of
different feedback conditions on motivational perceptions, interest, and performance.
Journal of Educational Psychology, 79, 474–482.
Butler, R., & Nisan, M. (1986). Effects of no feedback, task-related comments, and grades on
intrinsic motivation and performance. Journal of Educational Psychology, 78, 210–
216.
Carson, J. (2007). The measure of merit: Talents, intelligence, and inequality in the French
and American republics, 1750-1940. Princeton, NJ: Princeton University Press.
Castilla, E. J., & Benard, S. (2010). The paradox of meritocracy in organizations.
Administrative Science Quarterly, 55, 543-676.
Chmielewski, A. K. (2014). An International comparison of achievement inequality in within-
and between-school tracking systems. American Journal of Education, 120, 293–324.
https://doi.org/10.1086/675529
Running head: SELECTION AND SOCIAL CLASS
!
48
Cimpian, A., & Markman, E. M. (2011). The generic/nongeneric distinction influences how
children interpret new information about social others: Generic/nongeneric distinction.
Child Development, 82, 471–492. https://doi.org/10.1111/j.1467-8624.2010.01525.x
Costa-Lopes, R., Dovidio, J. F., Pereira, C. R., & Jost, J. T. (2013). Social psychological
perspectives on the legitimation of social inequality: Past, present and future.
European Journal of Social Psychology, 43, 229–237
Coulmont, B. (2011). Sociologie des prénoms. Paris, France: La Découverte.
Covington, M. V., & Omelich, C. L. (1979). Are causal attributions causal? A path analysis of
the cognitive model of achievement motivation. Journal of Personality and Social
Psychology, 37, 1487–1504.
Cozzarelli, C., Wilkinson, A. V., & Tagler, M. J. (2001). Attitudes toward the poor and
attributions for poverty. Journal of Social Issues, 57, 207–227.
https://doi.org/10.1111/0022-4537.00209
Crahay, M. (2012). L’école peut-elle être juste et efficace? (2nd ed.). Bruxelle: De Boeck
Superieur.
Croft, A., & Schmader, T. (2012). The feedback withholding bias: Minority students do not
receive critical feedback from evaluators concerned about appearing racist. Journal of
Experimental Social Psychology, 48, 1139–1144.
https://doi.org/10.1016/j.jesp.2012.04.010
Croizet, J.-C., & Claire, T. (1998). Extending the concept of stereotype and threat to social
class: The intellectual underperformance of students from low socioeconimic
backgrounds. Personality and Social Psychology Bulletin, 24, 588–594.
Croizet, J.-C., & Dutrévis, M. (2004). Socioeconomic status and intelligence: Why test scores
do not equal merit. Journal of Poverty, 8, 91–107.
Running head: SELECTION AND SOCIAL CLASS
!
49
Croizet, J.-C., Goudeau, S., Marot, M., & Millet, M. (2017). How do educational contexts
contribute to the social class achievement gap: documenting symbolic violence from a
social psychological point of view. Current Opinion in Psychology, 18, 105–110.
https://doi.org/10.1016/j.copsyc.2017.08.025
Croizet, J.-C., & Millet, M. (2012). Social class and test performance: From stereotype threat
to symbolic violence and vice versa. In M. Inzlicht & T. Schmader (Eds.), Stereotype
Threat: Theory, Process, and Application (pp. 188201). New York, NY: Oxford
University Press.
Crosby, J. R., & Monin, B. (2007). Failure to warn: How student race affects warnings of
potential academic difficulty. Journal of Experimental Social Psychology, 43, 663–
670. https://doi.org/10.1016/j.jesp.2006.06.007
Cumming, G. (2013). Understanding the new statistics: Effect sizes, confidence intervals, and
meta-analysis. New York, NY: Routledge/Taylor & Francis Group.
Darnon, C., Dompnier, B., Delmas, F., Pulfrey, C., & Butera, F. (2009). Achievement goal
promotion at university: Social desirability and social utility of mastery and
performance goals. Journal of Personality and Social Psychology, 96, 119–134.
Darnon, C., Jury, M., & Aelenei, C. (2017). Who benefits from mastery-approach and
performance-approach goals in college? Students’ social class as a moderator of the
link between goals and grade. European Journal of Psychology of Education, 1-14.
Darnon, C., Smeding, A., & Redersdorff, S. (2017). Belief in school meritocracy as an
ideological barrier to the promotion of equality: School meritocracy and promotion of
equality. European Journal of Social Psychology. https://doi.org/10.1002/ejsp.2347
Désert, M., Préaux, M., & Jund, R. (2009). So young and already victims of stereotype threat:
Socio-economic status and performance of 6 to 9 years old children on Raven’s
Running head: SELECTION AND SOCIAL CLASS
!
50
progressive matrices. European Journal of Psychology of Education, 24, 207–218.
https://doi.org/10.1007/BF03173012
DiTomaso, N. (2015). Racism and discrimination versus advantage and favoritism: Bias for
versus bias against. Research in Organizational Behavior, 35, 57–77.
https://doi.org/10.1016/j.riob.2015.10.001
DiTomaso, N., Post, C., Smith, D. R., Farris, G. F., & Cordero, R. (2007). Effects of
structural position on allocation and evaluation decisions for scientists and engineers
in industrial R&D. Administrative Science Quarterly, 52, 175–207.
Dittmann, A. G., & Stephens, N. M. (2017). Interventions aimed at closing the social class
achievement gap: changing individuals, structures, and construals. Current Opinion in
Psychology, 18, 111–116. https://doi.org/10.1016/j.copsyc.2017.07.044
Domina, T., Penner, A., & Penner, E. (2017). Categorical inequality: Schools as sorting
machines. Annual Review of Sociology, 43, 311-330.
Dompnier, B., Darnon, C., & Butera, F. (2009). Faking the desire to learn: A clarification of
the link between mastery goals and academic achievement. Psychological Science, 20,
939–943. https://doi.org/10.1111/j.1467-9280.2009.02384.x
Dornbusch, S. M., Glasgow, K. L., & Lin, I.-C. (1996). The social structure of schooling.
Annual Review of Psychology, 47, 401–429.
https://doi.org/10.1146/annurev.psych.47.1.401
Dubet, F. (2004). L’école des chances: Qu’est-ce qu’une école juste? Paris, France: Seuil.
Durante, F., & Fiske, S. T. (2017). How social-class stereotypes maintain inequality. Current
Opinion in Psychology, 18, 43–48. https://doi.org/10.1016/j.copsyc.2017.07.033
Duru-Bellat, M. (2006). L’inflation scolaire: les désillusions de la méritocratie. Paris,
France: Le Seuil.
Running head: SELECTION AND SOCIAL CLASS
!
51
Duru-Bellat, M. (2008). Recent trends in social reproduction in France: should the political
promises of education be revisited? Journal of Education Policy, 23, 81–95.
https://doi.org/10.1080/02680930701754104
Dweck, C. S. (2012). Mindsets and human nature: Promoting change in the Middle East, the
schoolyard, the racial divide, and willpower. American Psychologist, 67, 614–622.
https://doi.org/10.1037/a0029783
Faniko, K., Lorenzi-Cioldi, F., Buschini, F., & Chatard, A. (2012). The influence of education
on attitudes toward affirmative action: The role of the policy’s strength. Journal of
Applied Social Psychology, 42, 387–413. https://doi.org/10.1111/j.1559-
1816.2011.00892.x
Fine, M., & Burns, A. (2003). Class notes: Toward a critical psychology of class and
schooling. Journal of Social Issues, 59, 841860. http://doi.org/10.1046/j.0022-
4537.2003.00093.x
Filer, A. (2002). Socio-historical and cultural contexts of assessment policy. In A. Filer (Ed.),
Assessment: Social Practice and Social Product (pp. 7–10). New York, NY:
Routledge.
Fiske, A. P., Kitayama, S., Markus, H. R., & Nisbett, R. E. (1998). The cultural matrix of
social psychology. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of
social psychology, Vols. 1 and 2 (4th ed.). (pp. 915–981). New York, NY: McGraw-
Hill.
Fiske, S. T., & Markus, H. R. (2012). Facing social class: How societal rank influences
interaction. New York, NY: Russell Sage Foundation
Forquin, J.-C. (1992). École et Culture: Le point de vue des sociologues britanniques.
Bruxelle: De Boeck Université.
Running head: SELECTION AND SOCIAL CLASS
!
52
Frey, B. B., & Schmitt, V. L. (2007). Coming to terms with classroom assessment. Journal of
Advanced Academics, 18, 402–423.
Gewirtz, S. (2000). Bringing the politics back in: A critical analysis of quality discourses in
education. British Journal of Educational Studies, 48, 352–370.
Glaser, R. (1963). Instructional technology and the measurement of learning outcomes: Some
questions. American Psychologist, 18, 519.
Goudeau, S., Autin, F., & Croizet, J. C. (2017). Etudier, mesurer et manipuler la classe sociale
en psychologie sociale: Approches economiques, symboliques et culturelles
[Studying, measuring and manipulating social class in social psychology: Economic,
symbolic and cultural approaches]. International Review of Social Psychology, 30,!1–
19. https://doi.org/10.5334/irsp.52
Goudeau, S., & Croizet, J.-C. (2017). Hidden advantages and disadvantages of social class:
How classroom settings reproduce social inequality by staging unfair comparison.
Psychological Science, 28, 162-170.
Greenwald, A. G., & Pettigrew, T. F. (2014). With malice toward none and charity for some:
Ingroup favoritism enables discrimination. American Psychologist, 69, 669–684.
https://doi.org/10.1037/a0036056
Hall, K., Collins, J., Benjamin, S., Nind, M., & Sheehy, K. (2004). SATurated models of
pupildom: Assessment and inclusion/exclusion. British Educational Research Journal,
30, 801–817. https://doi.org/10.1080/0141192042000279512
Harackiewicz, J. M., Canning, E. A., Tibbetts, Y., Giffen, C. J., Blair, S. S., Rouse, D. I., &
Hyde, J. S. (2014). Closing the social class achievement gap for first-generation
students in undergraduate biology. Journal of Educational Psychology, 106, 375–389.
https://doi.org/10.1037/a0034679
Running head: SELECTION AND SOCIAL CLASS
!
53
Harari, O., & Covington, M. V. (1981). Reactions to achievement behavior from a teacher
and student perspective: A developmental analysis. American Educational Research
Journal, 18, 15–28. https://doi.org/10.3102/00028312018001015
Harrison, L. A., Stevens, C. M., Monty, A. N., & Coakley, C. A. (2006). The consequences of
stereotype threat on the academic performance of White and non-White lower income
college students. Social Psychology of Education, 9, 341–357.
https://doi.org/10.1007/s11218-005-5456-6
Hinnerich, B. T., Höglin, E., & Johannesson, M. (2015). Discrimination against students with
foreign backgrounds: evidence from grading in Swedish public high schools. Education
Economics, 23, 660-676. http://doi.org/10.1080/09645292.2014.899562
Hewstone, M., Rubin, M., & Willis, H. (2002). Intergroup bias. Annual Review of
Psychology, 53, 575–604. https://doi.org/10.1146/annurev.psych.53.100901.135109
Jackman, M. R. (1994). The velvet glove: Paternalism and conflict in gender, class, and race
relations. Berkeley, CA: University of California Press.
Jasso, G., & Resh, N. (2002). Exploring the sense of justice about grades. European
Sociological Review, 18, 333–351. https://doi.org/10.1093/esr/18.3.333
Judd, C. M., McClelland, G. H., & Ryan, C. S. (2011). Data analysis: A model comparison
approach. New York : NY. Routledge
Jury, M., Darnon, C., Dompnier, B., & Butera, F. (2017). The social utility of performance-
approach goals in a selective educational environment. Social Psychology of
Education, 20(1), 215-235.
Jury, M., Smeding, A., & Darnon, C. (2015). First-generation students’ underperformance at
university: the impact of the function of selection. Frontiers in Psychology, 6, 710.
https://doi.org/10.3389/fpsyg.2015.00710
Running head: SELECTION AND SOCIAL CLASS
!
54
Jury, M., Smeding, A., Stephens, N. M., Nelson, J. E., Aelenei, C., & Darnon, C. (2017). The
experience of low-SES students in higher education: Psychological barriers to success
and interventions to reduce social-class inequality. Journal of Social Issues, 73, 23–
41. https://doi.org/10.1111/josi.12202
Katz, I., & Hass, R. G. (1988). Racial ambivalence and American value conflict: Correlational
and priming studies of dual cognitive structures. Journal of Personality and Social
Psychology, 55, 893–905.
Knight, P. T., & Yorke, M. (2003). Assessment, Learning And Employability. Maidenhead,
UK: Open University Press.
Knowles, E. D., & Lowery, B. S. (2012). Meritocracy, self-concerns, and Whites’ denial of
racial inequity. Self and Identity, 11, 202–222.
https://doi.org/10.1080/15298868.2010.542015
Kraus, M. W., Callaghan, B., & Ondish P. (2017). Social class as culture. In S. Kitayama &
D. Cohen (Eds.) Handbook of Cultural Psychology (2nd edition). NY: Guilford.
Kraus, M. W., & Park, J. W. (2017). The structural dynamics of social class. Current Opinion
in Psychology, 18, 55–60. https://doi.org/10.1016/j.copsyc.2017.07.029
Kruglanski, A. W., & Freund, T. (1983). The freezing and unfreezing of lay inferences :
Effects of impressional primacy, ethnic stereotyping and numerical anchoring. Journal
of Experimental Social Psychology, 19, 448—468.
Lareau, A. (2011). Unequal Childhoods: Class, Race, and Family Life. University of
California Press.
Lemann, N. (1999). The big test: The secret history of the American meritocracy. New York,
NY: Farrar, Straus and Giroux.
Lerner, J. S., & Tetlock, P. E. (1999). Accounting for the effects of accountability.
Psychological bulletin, 125, 255.
Running head: SELECTION AND SOCIAL CLASS
!
55
LeTendre, G. K., Hofer, B. K., & Shimizu, H. (2003). What is tracking? Cultural expectations
in the united states, germany, and japan. American Educational Research Journal, 40,
43–89. https://doi.org/10.3102/00028312040001043
Malouff, J. M., & Thorsteinsson, E. B. (2016). Bias in grading: A meta-analysis of
experimental research findings. Australian Journal of Education, 60, 245–256.
https://doi.org/10.1177/0004944116664618
Markus, H. R., & Hamedani, M. G. (2007). Sociocultural psychology: The dynamic
interdependence among self systems and social systems. In S. Kitayama & D. Cohen
(Eds.), Handbook of cultural psychology (pp. 3–39). New York, NY, US: Guilford
Press.
Markus, H. R., & Stephens, N. M. (2017). Editorial overview: The psychological and
behavioral consequences of inequality and social class: A theoretical integration.
Current Opinion in Psychology. https://doi.org/10.1016/j.copsyc.2017.11.001
McCoy, S. K., & Major, B. (2007). Priming meritocracy and the psychological justification of
inequality. Journal of Experimental Social Psychology, 43, 341–351.
https://doi.org/10.1016/j.jesp.2006.04.009
McNair, S., Bhargava, A., Adams, L., Edgerton, S., & Kypros, B. (2003). Teachers speak out
on assessment practices. Early Childhood Education Journal, 31, 23–31.
Moskowitz, G. B., Gollwitzer, P. M., Wasel, W., & Schaal, B. (1999). Preconscious control of
stereotype activation through chronic egalitarian goals. Journal of Personality and
Social Psychology, 77, 167–184. https://doi.org/10.1037/0022-3514.77.1.167
Moskowitz, G. B., Salomon, A. R., & Taylor, C. M. (2000). Preconsciously controlling
stereotyping: Implicitly activated egalitarian goals prevent the activation of
stereotypes. Social Cognition, 18, 151–177.
https://doi.org/10.1521/soco.2000.18.2.151
Running head: SELECTION AND SOCIAL CLASS
!
56
Mueller, C. M., & Dweck, C. S. (1998). Praise for intelligence can undermine children’s
motivation and performance. Journal of Personality and Social Psychology, 75, 33–
52.
Norton, M. I., Vandello, J. A., & Darley, J. M. (2004). Casuistry and social category bias.
Journal of Personality and Social Psychology, 87, 817-831.
OECD. (2006). Education at a Glance: OECD Indicators 2006. Paris, France: OECD
Publishing.
OECD. (2013a). PISA 2012 Results: Excellence through equity: Giving every student the
chance to succeed (Volume II). Paris, France: OECD Publishing.
OECD. (2013b). PISA 2012 Results: What Makes Schools Successful? Resources, Policies
and Practices (Volume IV). Paris, France: OECD Publishing.
Oppenheimer, D. M., Meyvis, T., & Davidenko, N. (2009). Instructional manipulation
checks: Detecting satisficing to increase statistical power. Journal of Experimental
Social Psychology, 45, 867–872. https://doi.org/10.1016/j.jesp.2009.03.009
Parsons, T. (1959). The school class as a social system: Some of its functions in American
society. Harvard Educational Review, 29, 297–318.
Perrenoud, P. (1995). La pédagogie à l’école des différences. Paris, France: ESF.
Popham, W. J. (2001). Teaching to the Test? Educational Leadership, 58, 16–21.
Rangvid, B. S. (2015). Systematic differences across evaluation schemes and educational
choice. Economics of Education Review, 48, 41–55.
https://doi.org/10.1016/j.econedurev.2015.05.003
Rattan, A., Good, C., & Dweck, C. S. (2012). “It’s ok — Not everyone can be good at math”:
Instructors with an entity theory comfort (and demotivate) students. Journal of
Experimental Social Psychology, 48, 731–737.
https://doi.org/10.1016/j.jesp.2011.12.012
Running head: SELECTION AND SOCIAL CLASS
!
57
Rattan, A., Savani, K., Chugh, D., & Dweck, C. S. (2015). Leveraging mindsets to promote
academic achievement: Policy recommendations. Perspectives on Psychological
Science, 10, 721–726.
Resh, N. (2009). Justice in grades allocation: Teachers’ perspective. Social Psychology of
Education, 12, 315–325. https://doi.org/10.1007/s11218-008-9073-z
Rosenholtz, S. J., & Simpson, C. (1984). The formation of ability conceptions:
Developmental trend or social construction? Review of Educational Research, 54, 31–
63.
Sabbagh, C., Faher-Aladeen, R., & Resh, N. (2004). Evaluation of grade distributions: a
comparison of Jewish and Druze students in Israel. Social Psychology of Education, 7,
313–337. https://doi.org/10.1023/B:SPOE.0000037547.11163.36
Sadler, D. R. (1989). Formative assessment and the design of instructional systems.
Instructional Science, 18, 119–144.
Shute, V. J. (2008). Focus on formative feedback. Review of Educational Research, 78, 153–
189. https://doi.org/10.3102/0034654307313795
Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2011). False-positive psychology:
Undisclosed flexibility in data collection and analysis allows presenting anything as
significant. Psychological Science, 22, 1359–1366.
https://doi.org/10.1177/0956797611417632
Simon, S., Ditrichs, R., & Grier, J. B. (1995). The simulated class as a method for studying
teacher decision making. Computers in Human Behavior, 11, 163–180.
https://doi.org/10.1016/0747-5632(94)00020-I
Smeding, A., Darnon, C., Souchal, C., Toczek-Capelle, M.-C., & Butera, F. (2013). Reducing
the socio-economic status achievement gap at university by promoting mastery-
Running head: SELECTION AND SOCIAL CLASS
!
58
oriented assessment. PLoS ONE, 8, e71678.
https://doi.org/10.1371/journal.pone.0071678
Son Hing, L. S., Bobocel, D. R., Zanna, M. P., Garcia, D. M., Gee, S. S., & Orazietti, K.
(2011). The merit of meritocracy. Journal of Personality and Social Psychology, 101,
433–450. https://doi.org/10.1037/a0024618
Souchal, C., & Toczek, M.-C. (2010). Buts de réussite, conceptions de l’intelligence,
différences de performances liées à l’appartenance socio-économique des élèves: de
nouvelles hypothèses explicatives? Les Sciences de l’éducation - Pour l’Ère nouvelle,
43, 13–35.
Spencer, B., & Castano, E. (2007). Social class is dead. Long live social class! Stereotype
threat among low socioeconomic status individuals. Social Justice Research, 20, 418–
432. https://doi.org/10.1007/s11211-007-0047-7
Stephens, N. M., Fryberg, S. A., Markus, H. R., Johnson, C. S., & Covarrubias, R. (2012).
Unseen disadvantage: How American universities’ focus on independence undermines
the academic performance of first-generation college students. Journal of Personality
and Social Psychology, 102, 1178–1197. https://doi.org/10.1037/a0027143
Stephens, N. M., Hamedani, M. G., & Destin, M. (2014). Closing the social-class
achievement gap: A difference-education intervention improves first-generation
students’ academic performance and all students’ college transition. Psychological
Science, 25, 943–953. https://doi.org/10.1177/0956797613518349
Stephens, N. M., Markus, H. R., & Phillips, L. T. (2014). Social class culture cycles: How
three gateway contexts shape selves and fuel inequality. Annual Review of
Psychology, 65, 611–634. https://doi.org/10.1146/annurev-psych-010213-115143
Taras, M. (2005). Assessment–summative and formative–some theoretical reflections. British
Journal of Educational Studies, 53, 466–478.
Running head: SELECTION AND SOCIAL CLASS
!
59
Taras, M. (2009). Summative assessment: the missing link for formative assessment. Journal
of Further and Higher Education, 33, 57–69.
https://doi.org/10.1080/03098770802638671
Thorndike, E. L. (1913). Educational psychology (Vol. Vol. 1-1). New York, NY: Teachers
College.
Tibbetts, Y., Harackiewicz, J. M., Canning, E. A., Boston, J. S., Priniski, S. J., & Hyde, J. S.
(2016). Affirming independence: Exploring mechanisms underlying a values
affirmation intervention for first-generation students. Journal of Personality and
Social Psychology, 110, 635–659. https://doi.org/10.1037/pspa0000049
Torrance, H., & Pryor, J. (1998). Introduction. In H. Torrance, & J. Pryor (Eds.), Investigating
Formative Assessment: Teaching, Learning and Assessment in the Classroom. (pp. 1–
7). Berkshire, UK: Open University Press.
Turner, R. H. (1961). Modes of social ascent through education: Sponsored and contest
mobility. In A. H. Halsey, J. Floud, & J. Anderson (Eds.), Education, economy and
society (pp. 121–139). New York, NY: Free Press.
Uhlmann, E. L., & Cohen, G. L. (2005). Constructed criteria: Redefining merit to justify
discrimination. Psychological Science, 16, 474-480.
Uhlmann, E. L., & Cohen, G. L. (2007). “I think it, therefore it’s true”: Effects of self-
perceived objectivity on hiring discrimination. Organizational Behavior and Human
Decision Processes, 104, 207-223.
Webster, D. M., & Kruglanski, A. W. (1997). Cognitive and social consequences of the need
for cognitive closure. European Review of Social Psychology, 8, 133-173.
Westfall, J. (2015, May 27). Follow-up: What about Uri’s 2n rule? Retrieved from
http://jakewestfall.org/blog/index.php/2015/05/27/follow-up-what-about-uris-2n-rule/
Running head: SELECTION AND SOCIAL CLASS
!
60
Wiederkehr, V., Bonnot, V., Krauth-Gruber, S., & Darnon, C. (2015). Belief in school
meritocracy as a system-justifying tool for low status students. Frontiers in
Psychology, 6. https://doi.org/10.3389/fpsyg.2015.01053
Wyer, N. A. (2003). Value conflicts in intergroup perception: A social cognitive perspective.
In G. V. Bodenhausen & A. J. Lambert (Eds.), Foundations of social cognition (pp.
263–289). Mahwah, NJ: Lawrence Erlbaum Associates.
Wyer, N. A. (2010). Salient egalitarian norms moderate activation of out-group approach and
avoidance. Group Processes & Intergroup Relations, 13, 151–165.
https://doi.org/10.1177/1368430209347326
Yeager, D. S., Purdie-Vaughns, V., Garcia, J., Apfel, N., Brzustoski, P., Master, A., ... &
Cohen, G. L. (2014). Breaking the cycle of mistrust: Wise interventions to provide
critical feedback across the racial divide. Journal of Experimental Psychology:
General, 143, 804.
Zdaniuk, A., & Bobocel, D. R. (2011). Independent self-construal and opposition to
affirmative action: The role of microjustice and macrojustice preferences. Social
Justice Research, 24, 341–364. https://doi.org/10.1007/s11211-011-0143-6
Zogmaister,C., Arcuri, L., & Castelli, L. (2008). The impact of loyalty and equality on
implicit ingroup favoritism. Group Processes & Intergroup Relations,11, 493–512.
http://doi.org/10.1177/1368430208095402
Running head: SELECTION AND SOCIAL CLASS
!
61
Footnotes
1 In the various lines of research presented in this section, social class is
operationalized in different ways. In some studies, it refers to socio-economic status (e.g.,
Croizet & Claire, 1998), whereas in other to socio-cultural status (e.g., first/continuing
generation to attend University; Stephens et al., 2012). A comparison of the various aspects of
social class is beyond the scope of the present work, and has been reviewed by others
(Goudeau, Autin, & Croizet, 2017; Kraus, Callaghan, & Ondish, 2017). We have referred to
these various lines of research without highlighting the differences in operationalization to the
extent that in all of these studies social class refers to a social background that relates to more
or less chances of success in education.
2 The booklet also included a questionnaire measuring the predicted future success of
the target and the perception of the assessment method and of academic performance. These
measures are not relevant for the hypothesis presented here, and we did not report the results,
but they are available upon request from the authors.
3 Two values are missing because participants did not fill the item.
4 Three outliers were excluded from the analysis due to uncommon deleted
studentized residual, centered leverage values and abnormal residuals.
5 One outlier was excluded from the analysis due to uncommon Cook’s distance and
deleted studentized residuals.
6 The booklet also included a questionnaire measuring the predicted future success of
the target, the rating of the test and tracking recommendations. These measures are not
relevant for the hypothesis presented here, and we did not report the results, but they are
available upon request from the authors.
7 One missing value
Running head: SELECTION AND SOCIAL CLASS
!
62
8 The booklet also included a questionnaire measuring the predicted future success of
the target and the rating of the test. These measures are not relevant for the hypothesis
presented here, and we did not report the results, but they are available upon request from the
authors.
9 One outlier was excluded due to uncommon deleted studentized residual.
10 The booklet also included an evaluation of the essay on a 10-point scale. The
results are available upon request from the authors
11 Three participants did not fill the item
12 Two outliers were removed from the analysis due to elevated cooks’ distances and
studentized deleted residuals
13 Three participants were excluded due to abnormal residuals, elevated cooks’
distances and studentized deleted residuals
14 Four outlier was excluded due to abnormal residuals, elevated cooks’ distances and
studentized deleted residuals
Running head: SELECTION AND SOCIAL CLASS
!
63
Figure 1. Experiment 1. Relationship between the time taken to complete the study and the
number of mistakes found in the dictation as a function of target’s SES and assessment
method.
Note. **p .01, ***p < .001.
1"
3"
5"
7"
9"
11"
13"
15"
-1" 0" 1"
Number'of'mistakes'
Time'(SD)'
Assessment'for'Selec8on'
low"SES"
high"SES"
b"=".06"
***"
**"
b"=".44***"
1"
3"
5"
7"
9"
11"
13"
15"
-1" 0" 1"
Number'of'mistakes'
Time'(SD)'
Assessment'for'Learning'
low"SES"
high"SES"
b"=".05"
b"=".12**"
Running head: SELECTION AND SOCIAL CLASS
!
64
!
Figure 2. Experiment 1. Relationship between the number of mistakes and the grade as a
function of target’s SES.
Note. **p .01, ***p < .001.
0
0,5
1
1,5
2
2,5
3
3,5
4
4,5
5
5,5
6
-1 0 1
Grade
Number of Mistakes (SD)
low SES high SES
b = -.19***
b = -.09**
Running head: SELECTION AND SOCIAL CLASS
!
65
Figure 3. Experiment 2. Number of mistakes found in the dictation as a function of the
target’s SES and the assessment method. Error bars represent the 95% confidence intervals.
0
2
4
6
8
10
12
14
for selection for learning
Number of mistakes
Assessment method
low SES high SES
Running head: SELECTION AND SOCIAL CLASS
!
66
Figure 4. Experiment 3. Number of mistakes found in the dictation as a function of the
target’s SES and the function of assessment. Error bars represent the 95% confidence
intervals.
0
2
4
6
8
10
12
14
Selection Education
Number of mistakes
Function of assessment
Low SES High SES
... Des recherches plus récentes ont même montré que l'utilisation de l'évaluation normative exerce une pression sur les évaluatrices qui les pousse à reproduire les inégalités sociales qui existent à l'école. Dans une série d'études, Autin et al. (2019) demandaient à leurs participantes de corriger la dictée d'une enfant qu'elles ne connaissaient pas, mais dont elles possédaient quelques informations, dont le métier des parents. Cette information était utilisée pour manipuler expérimentalement la classe sociale de l'élève, plus favorisée ou plus défavorisée. ...
... Ce double discours en termes de formation et de sélection n'est pas neutre et a des conséquences sur la manière dont pensent et agissent enseignant·es et élèves/étudiant·es. Concernant les enseignant·es, Autin et al. (2019) ont, par exemple, récemment montré que lorsque la fonction de sélection est rendue saillante, ils·elles attribuent des notes différentes, à copie équivalente, entre élèves issu·es de milieux défavorisés et ceux·celles issu·es de milieux favorisés. Lorsque la fonction de formation est rendue saillante, la disparité de notation n'apparaît plus. ...
... Dans leur première étude -une tâche de mémorisation de six dossiers scolaires fictifs -, ils·elles ont par exemple montré que les notes obtenues par des élèves issu·es de milieux défavorisés étaient moins bien rappelées par les participant·es, mais seulement lorsque celles-ci étaient meilleures que celles d'élèves issu·es de milieux favorisés ; lorsqu'elles étaient moins bonnes, le rappel était moins affecté (puisqu'il était plus en phase avec les représentations stéréotypées). Dans leur seconde étude, ils·elles montraient qu'à (bonne) dictée équivalente, l'élève issu·e d'un milieu défavorisé recevait une moins bonne note que l'élève issu·e d'un milieu favorisé, mais uniquement lorsque ceux·celles-ci évoluaient dans une filière jugée plus prestigieuse et dans laquelle les élèves issu·es de milieux défavorisés réussissent habituellement moins bien que leurs camarades issu·es de milieux favorisés (voir aussi Autin et al., 2019). Dit autrement, ces travaux soulignent le risque important pour des élèves dont la réussite n'est pas attendue (à cause de stéréotypes de moindres compétences) de voir mésestimer leurs compétences, particulièrement lorsque leur réussite remet en cause l'ordre établi. ...
... Teachers' belief in the universality of ability may also shape their beliefs about the purpose of education: selection and weed-out (differentiating the 'cream of the crop' from those who have lesser abilities) versus promotion of learning for all [33]. Experimental induction of a selection purpose among educators has led to greater socioeconomic disparities in grades [34] and academic track recommendations [35]. ...
... whereas those who thought about education as promoting learning for all did not exhibit disparate assessment [34,35,55]. Other experiments suggest that teachers' biases may also predict disparate assessment, although results have been mixed. ...
Article
Although researchers investigating psychological contributors to educational inequality have traditionally focused on students, a growing literature highlights the importance of teachers’ psychology in shaping disparities in students’ educational achievement and attainment. In this review, we discuss recent advances linking teachers’ attitudes, perceptions, and beliefs to inequality in students’ outcomes. First, we identify specific aspects of teacher psychology that contribute to educational disparities, including teachers’ biases, perceptions and expectations of students, beliefs about the nature of ability, and beliefs about group differences. Second, we synthesize mechanisms underlying the effects of teacher psychology on educational inequality, including teachers’ disparate assessment of students’ work and abilities, interpersonal interaction with students, and psychological impact on students. Implications for future research and interventions are discussed.
... Faculty and peers also contribute directly to the achievement and well-being of lower SES students. For example, when instructors present assessments, such as tests, as tools to sort students by ability, instructors tend to evaluate student work with biases that advantage higher SES students (Autin et al., 2019). Furthermore, everyday interactions with instructors, staff, and peers communicate hostile and derogatory slights and insults to lower SES and minoritized students (microaggressions; Rogers et al., 2020;Suárez-Orozco et al., 2015;Sue et al., 2007;Watkins et al., 2010). ...
... Students are better able to learn when instructors genuinely recognize and substantively incorporate the value associated with their students' diverse backgrounds and experiences (Silverman et al., under review). Furthermore, a range of specific and commonplace classroom practices introduce biases that have disproportionately negative effects on lower SES students (Autin et al., 2019). Regular opportunities for instructors to engage in professional development-to critically examine their practices and enhance their teaching-would certainly benefit students. ...
Article
Full-text available
As colleges and universities expand the socioeconomic diversity of their student populations, many policies and practices require reconceptualization to better serve all students. Recent social psychology and learning sciences research directly informs how to support the achievement and well-being of students from lower socioeconomic status backgrounds, with attention to intersecting minoritized identities. These approaches challenge assimilationist and deficit-based views of student identities in addressing factors at multiple levels of their sociocultural contexts. Building from the evidence, recommendations emphasize committing financial resources to allow for full access and participation in higher education. Also, specific faculty practices and development opportunities can enhance teaching. Finally, community emerges as a central theme; recommendations enhance student connections within and beyond the college environment.
... An alternative approach is to sideline bias, that is, to create situations in which bias is not functional for the goals teachers pursue 72 . For example, research shows that teachers show less bias against low-SES students in contexts emphasizing learning-helping all students learn and grow-as opposed to those emphasizing selection-identifying and rewarding the smartest students 73 . Third, praise interventions can help teachers provide praise in less biased ways. ...
Article
Full-text available
Can teachers’ inflated praise make children from low socioeconomic status (SES) backgrounds seem less smart? We conducted two preregistered experiments to address this question. We used hypothetical scenarios to ensure experimental control. An experiment with primary school teachers ( N = 106, ages 21–63) showed that when a child from a low-SES (vs. high-SES) background succeeded in school, teachers attributed this success more to hard work and delivered more inflated praise (e.g., “You did incredibly well!”) but less modest praise (e.g., “You did well!”). An experiment with primary school children ( N = 63, ages 10–13) showed that when children learned that another child received inflated praise (while an equally performing classmate received modest praise or no praise), they perceived this child as less smart but more hardworking. These studies provide converging evidence that teachers’ inflated praise, although well-intentioned, can make children from low-SES backgrounds seem less smart, thereby reinforcing negative stereotypes about these children’s academic abilities.
... However, selection in school is not neutral since it has been shown to sustain inequalities within education systems (see for example, Jury et al., 2015;Smeding et al., 2013;Souchal et al., 2014). For instance, Autin et al. (2019) recently showed that when the selection function is made salient, evaluators are more likely to create a disparity between students from disadvantaged and advantaged backgrounds in their assessment than when the education function is made salient. Thus, selecting students can alter teachers' evaluative practices, thereby contributing to the reproduction of social inequalities from school age (see also Autin et al., 2015;Batruch et al., 2017Batruch et al., , 2019. ...
Article
The implementation of inclusive practices in mainstream education remains particularly difficult in the French context and is influenced by various factors including the types of disability labels, and the type of assessment practices that are used. Indeed, how student disability is labelled could impact teacher attitudes by notably disfavouring students labelled with autism. Moreover, normative assessment is strongly linked with selection at schools—a function that works against teacher attitudes towards inclusive education. This article reports on a study in which we examined teacher intentions to use materials accommodated to special educational needs students, as a function of special needs labelling. Specifically, this refers to the use of labels for either a disability or special educational need, in connection to tasks associated with learning or assessment. The results of our study revealed that, for both types of labels, the intentions to use accommodated materials are lower when teachers are asked to assess student competence than when prompted to teach this competence. These findings are discussed with consideration to the incompatibility between selection in schools—which is aligned with the principle of meritocracy—and efforts to promote inclusive education practices.
... While these initiatives have often been successful (though see also competing findings, e.g., Brez et al., 2020;Canning, Priniski, & Harackiewicz, 2019;McCabe et al., 2020), related research suggests that individual deficits are not entirely to blame for educational inequalities, or at least cannot fully explain pervasive disparities (e.g., Destin et al., 2019). Rather, a considerable body of evidence demonstrates how biases embedded within students' academic contexts-including within educators' expectations, beliefs, and behaviors toward students who face systemic inequities-contribute to or artificially create disparities between different groups of students (e.g., Autin et al., 2019;Berkowitz, 2022;Canning, Muenks, et al., 2019;Okonofua & Eberhardt, 2015;Silverman et al., in press). ...
Article
Academic abstract: Personality and social psychology have historically viewed individuals' systemically marginalized identities (e.g., as people of color, as coming from a lower-income background) as barriers to their success. Such a deficit-based perspective limits psychological science by overlooking the broader experiences, value, perspectives, and strengths that individuals who face systemic marginalization often bring to their societies. The current article aims to support future research in incorporating a strength-based lens through tracing psychology's journey away from an emphasis on deficits among people who contend with systemic marginalization and toward three distinct strength-based approaches: the universal strengths, difference-as-strength, and identity-specific strengths approaches. Through distinguishing between each approach, we advance scholarship that aims to understand systemically marginalized identities with corresponding implications for addressing inequality. Strength-based approaches guide the field to recognize the imposed limitations of deficit-based ideologies and advance opportunities to engage in research that effectively understands and values systemically marginalized people. Public abstract: Inequalities, including those between people from higher- and lower-income backgrounds, are present across society. From schools to workplaces, hospitals to courtrooms, people who come from backgrounds that are marginalized by society often face more negative outcomes than people from more privileged backgrounds. While such inequalities are often blamed on a lack of hard work or other issues within marginalized people themselves, scientific research increasingly demonstrates that this is not the case. Rather, studies consistently find that people's identities as coming from groups that face marginalization in society often serve as a valuable source of unique strengths, not deficiencies, that can help them succeed. Our article reviews these studies to examine how future research in psychology may gain a broader understanding of people who contend with marginalization. In doing so, we outline opportunities for psychological research to effectively support efforts to address persistent inequalities.
... Both the studies by Batruch et al. (2019) and Autin et al. (2019;experiment 3) suggest that institutional selection tools such as tracking may induce SES biases in teacher evaluations. These findings highlight that contexts in which evaluators are encouraged to focus on selecting students may enhance biases. ...
Article
Full-text available
Sorting students on the basis of their academic performance into hierarchically ordered curriculums (i.e., between-school tracking) is common practice in various educational systems. International studies show that this form of tracking is associated with increased educational inequalities. As track placement is often based on teacher recommendations, biased track recommendations may contribute to this inequality. To shed light on the role that teachers play in the reproduction of inequalities in school, we conducted a systematic review of 27 recent articles on teachers' between-school tracking recommendations and students’ socio-economic or ethnic background. We find that teacher recommendations are biased against students from disadvantaged socio-economic backgrounds, yet evidence with respect to ethnic biases is more mixed. While student, parent, teacher, and contextual factors seem to play a role in tracking recommendations, they cannot account for the biases in tracking recommendations. We discuss promising areas for future studies and argue that research on institutional moderators may have more potential than research on psychological mediators to effectively reduce bias in educational institutions.
... Both the studies by Batruch et al. (2019) and Autin et al. (2019;experiment 3) suggest that institutional selection tools such as tracking may induce SES biases in teacher evaluations. These findings highlight that contexts in which evaluators are encouraged to focus on selecting students may enhance biases. ...
Preprint
Full-text available
Sorting students into hierarchically ordered tracks or streams on the basis of their academic performance (i.e., tracking) is ubiquitous in educational systems, and oftentimes based on teachers’ track recommendations. International surveys indicate that tracking is associated with educational inequalities. To determine if inequalities in tracking may be due to teacher recommendations being biased against students from disadvantaged socio-economic and/or ethnic backgrounds, we conducted a systematic review of 26 recent articles on tracking recommendations and students’ socio-economic or ethnic background. We find that teacher recommendations are biased against students from disadvantaged socio-economic backgrounds, yet evidence with respect to ethnic biases is more mixed. We also conducted an integrative review to examine which factors may account for social and ethnic inequalities in teacher tracking recommendations. We conclude that students’, parents’ and teachers’ attitudes and behaviours play a role in tracking recommendations but cannot fully account for the inequality in these recommendations. We discuss promising areas for future study, and argue that research may want to focus on finding institutional moderators in order to combat biases in educational institutions.
Presentation
Ce diaporama est issu d'une conférence dispensée au Conservatoire Royal de Liège et traitant de la problématique de la diversité et de sa gestion dans les systèmes d'éducation. Deux focales spécifiques y sont abordées : celle de l'enseignement spécialisée en Belgique francophone et celle de l'éducation aux arts.
Article
Full-text available
This research shows stereotype activation is controlled by chronic egalitarian goals. In the first 2 studies it was found that the stereotype of women is equally available to individuals with and without chronic goals, and the discriminant validity of the concept of egalitarian goals was established. In the next 2 experiments, differences in stereotype activation as a function of this individual difference were found. In Study 3, participants read attributes following stereotypical primes. Facilitated response times to stereotypical attributes were found for nonchronics but not for chronics. This lack of facilitation occurred at stimulus onset asynchronies (SOAs) where effortful correction processes could not operate, demonstrating preconscious control of stereotype activation due to chronic goals. In Study 4, inhibition of the stereotype was found at an SOA where effortful processes of stereotype suppression could not operate. The data reveal that goals are activated and used preconsciously to prevent stereotype activation, demonstrating both the controllability of stereotype activation and the implicit role of goals in cognitive control.
Article
Full-text available
Selection practices in education, such as tracking, may represent a structural obstacle that contributes to the social class achievement gap. We hypothesized that school’s function of selection leads evaluators to reproduce social inequalities in tracking decisions, even when performance is equal. In two studies, participants (students playing the role of teachers, N = 99, or preservice and in-service teachers, N = 70) decided which school track was suitable for a pupil whose socioeconomic status (SES) was manipulated. Although pupils’ achievement was identical, participants considered a lower track more suitable for lower SES than higher SES pupils, and the higher track more suitable for higher SES than lower SES pupils. A third study (N = 160) revealed that when the selection function of school was salient, rather than its educational function, the gap in tracking between social classes was larger. The selection function of tracking appears to encourage evaluators to artificially create social class inequalities.
Article
In the educational system, social inequalities are reproduced and legitimated. As a consequence, the implementation of interventions that reduce the SES achievement gap in education may face important ideological barriers. The purpose of the present paper is to test the belief in school meritocracy as one of these barriers. In three studies, participants' willingness to see an equalizing pedagogical intervention implemented in one's own university (Study 1) or in one's own children's school (Studies 2 and 3) and their intention to engage in this implementation were measured. Results indicated that a higher belief in school meritocracy (but not meritocracy endorsement, study 1) predicted lower engagement in the implementation of an equalizing method. This is not observed for inequality-maintaining interventions (Studies 2 and 3), sustaining the role of descriptive school meritocracy in the perpetuation of social inequalities within education.
Article
Although approach forms of achievement goals (mastery and performance goals) have been shown to predict academic achievement in college, recent research underscores that these associations are rather weak and not consistently observed. The present study tests students’ social class (in the present research, generational status) as a moderator of the relationships between both mastery-approach goals and performance-approach goals and final grade. One hundred students (45 first-generation students and 55 continuing-generation students, Mage = 18.9, SD = 1.52) answered an achievement goal scale related to one of their classes at the beginning of the year. Their final grade for this class was obtained three months later. As expected, performance-approach goals positively predicted final grade only for upper-class students, while mastery-approach goals tend to do so for lower-class students, supporting the idea that different kinds of motivation could predict students’ achievement depending on their social class.
Article
This article examines how the educational system participates in the reproduction of social inequality. After exposing the basics of the Social Reproduction Theory developed in sociology by Bourdieu and Passeron in 1977, we examine the research in social psychology that documents the reality of 'symbolic violence' that is the symbolic power that operates in the classroom and undermines the performance of students from underprivileged backgrounds. Three lines of research are examined: self-esteem, self-threat and research on the non-neutrality of educational settings.
Article
Despite their egalitarian ethos, schools are social sorting machines, creating categories that serve as the foundation of later life inequalities. In this review, we apply the theory of categorical inequality to education, focusing particularly on contemporary American schools. We discuss the range of categories that schools create, adopt, and reinforce, as well as the mechanisms through which these categories contribute to production of inequalities within schools and beyond. We argue that this categorical inequality frame helps to resolve a fundamental tension in the sociology of education and inequality, shedding light on how schools can-at once-be egalitarian institutions and agents of inequality. By applying the notion of categorical inequality to schools, we provide a set of conceptual tools that can help researchers understand, measure, and evaluate the ways in which schools structure social inequality.
Article
Understanding the sources of the social class achievement gap in education is an important step toward ensuring that education serves its purpose as an engine of social mobility. The goal of the current article is to provide a brief overview of the sources of the social class achievement gap as well as interventions aimed at closing this gap. We outline three major sources of the social class achievement gap-individual skills, structural conditions, and people's processes of meaning-making, or construals-and the interventions that target them. While all of these interventions can effect change, we propose that interventions will be most effective when tailored to fit the specific needs of students and the context in which they are delivered.