Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Affirmation and majority students: Can affirmation impair math performance?
(USC Thesis Other)
Affirmation and majority students: Can affirmation impair math performance?
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
Running Head: AFFIRMATION AND MAJORITY STUDENTS i
AFFIRMATION AND MAJORITY STUDENTS: CAN AFFIRMATION IMPAIR MATH
PERFORMANCE?
BY
Eduardo Jones
___________________________________________________
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PSYCHOLOGY)
December 2017
Copyright 2017 Eduardo Jones
AFFIRMATION AND MAJORITY STUDENTS ii
Acknowledgements
It has been an incredible journey to get to this point in my academic career. One that
would not have been possible without the contribution of so many individuals who played
invaluable roles in my development and allowed me to pursue my academic interests
wholeheartedly. First, I must thank Dr. Ursula Whiteside, my undergraduate mentor, who saw in
me a potential for scholarly contribution and paved the road to graduate school for me. You
believed in me and pushed open the doors of academia to me.
To Dr. Steven Lòpez, who initiated one of my most life changing days when you called
and invited me to be a graduate student at USC. I will never forget that day and the knowing that
my life had forever changed. Thank you for your support, for seeing my promise, and chairing
my master’s thesis and serving on my dissertation committee. You changed my life.
To Dr. Stan Huey, my faculty mentor and dissertation chair, what a journey this has been
for us. During all the trials and tribulations, the successes and victories, you have been a true
compass guiding me in the night, and anchoring me during the turbulent winds and choppy
waters of graduate school. You invested in me. You taught me how to critically read. You taught
me how to write. You imparted skills that will last a lifetime and for which I will forever be
grateful.
I am also deeply appreciative to Drs. Daphna Oyserman, Sarah Townsend, and Cleopatra
Abdou. Thank you for your time and support, and for the contribution of your knowledge and
expertise in the development of my dissertation topic, design, and execution. I am grateful for
each of you serving on my committee and helping shepherd me through the dissertation process.
AFFIRMATION AND MAJORITY STUDENTS iii
I would like to thank my graduate school cohort: Jean Kim, Jackie Tilley, Ali Cram, and
Jung Jang. Our journey at USC has been a wonderful one and I am happy to have shared it with
all of you. To my lab-mates: Taona Chithambo, Caitlin Smith, Marie Gillespie, Gabby Lewine,
Sylvanna Vargas, Miriam Rubenson, Nina Jhaveri, and Crystal Wang. We have enjoyed so
many incredible conversations and moments over the years. Each of you played a special role in
my development and I hope you all continue to support the Huey Lab ‘ohana. Marie Gillespie,
thank you for being the best office-mate in the world. For your humor, kindness, and support. To
Andy Larsen and Ryan Essex, my USC roommates, teammates, and friends. You made being at
USC fun. We grew up so much over the years and I will always remember and cherish our time
together.
Most of all, this dissertation is dedicated to my family. To my mother, for whom my
early interests in psychology were born. To Tia Diana, who has been a mother, teacher, friend,
and a spiritual companion in this life. To OJ, who helped get me here every step of the way.
From helping me write my USC graduate school application, to getting me my first car, to
stepping into the family role of provider so I could pursue my love of psychology. For being the
best older brother and friend I could ask for, this dissertation is dedicated to you. Lastly, to
Sanjeevi Sturges, for whom I would have never made it through graduate school without. For all
your support, which came in so many ways. For your sympathetic ear, of which you gave to me
every day. For your patience, of which you always seemed to find more even during the most
difficult times. For your love, of which you gave every ounce during the seven years I spent at
USC. I will never forget the many sacrifices you made so I could pursue my dream. This
dissertation is dedicated to you.
AFFIRMATION AND MAJORITY STUDENTS iv
Table of Contents
Acknowledgements ........................................................................................................................ ii
Table of Contents ........................................................................................................................... iv
List of Tables ..................................................................................................................................v
List of Figures ................................................................................................................................ vi
List of Appendices ........................................................................................................................ vii
Abstract ............................................................................................................................................1
Background and Significance ..........................................................................................................2
Current Studies...............................................................................................................................15
Study 1 ..........................................................................................................................................16
Method ...............................................................................................................................17
Results ................................................................................................................................20
Discussion ..........................................................................................................................26
Study 2 ...........................................................................................................................................29
Method ..............................................................................................................................31
Results ................................................................................................................................35
Discussion ..........................................................................................................................40
General Discussion ........................................................................................................................41
References ......................................................................................................................................47
AFFIRMATION AND MAJORITY STUDENTS v
List of Tables
Table 1: Study 1 participant demographic characteristics ...................................................56
Table 2: Study 1 frequency of values ranked most important ............................................57
Table 3: Study 1 awareness probe results ...........................................................................58
Table 4: Study 1 math performance and psychological outcomes ......................................59
Table 5: Study 2 participant demographic characteristics ..................................................60
Table 6: Study 2 participant math identity and values ........................................................61
Table 7: Study 2 frequency of values ranked most important ............................................62
Table 8: Study 2 awareness probe results ...........................................................................63
Table 9: Study 2 math performance and psychological outcomes .......................................64
AFFIRMATION AND MAJORITY STUDENTS vi
List of Figures
Figure 1: Study 1 math test performance ..............................................................................65
Figure 2: Study 2 math test performance .............................................................................66
AFFIRMATION AND MAJORITY STUDENTS vii
List of Appendices
Appendix A: Belonging-affirmation writing exercise .................................................................67
Appendix B: Neutral writing control exercise ............................................................................69
Appendix C: Study 1 math test ...................................................................................................71
Appendix D: Self-Integrity Scale ...............................................................................................79
Appendix E: Social Fit Questionnaire ........................................................................................80
Appendix F: Awareness probe ...................................................................................................81
Appendix G: Belonging-affirmation coding manual ..................................................................84
Appendix H: Motives for Attending College Questionnaire ......................................................85
Appendix I: Personal Mastery and Constraint Scale .................................................................86
Appendix J: Study 2 math test ...................................................................................................87
AFFIRMATION AND MAJORITY STUDENTS 1
Abstract
Self-affirmation is a promising brief intervention capable of reducing the academic achievement
gap between majority and stigmatized groups. Affirmations improve academic performance
among stigmatized groups (underrepresented minorities, women in STEM fields) by expanding
one’s sense of self, buffering social belonging, and reducing social identity threat (Cohen &
Sherman, 2014). Despite these encouraging findings, some studies suggest that affirmations may
decrease the academic performance of White students (Brady et al., 2016; Miyake et al., 2010;
Shnabel, Purdie-Vaughns, Cook, Garcia, & Cohen, 2013). We conducted experimental studies to
evaluate whether an affirmation focused on the theme of social belonging (i.e., belonging-
affirmation) decreases the math performance of White males. We hypothesized that the
belonging-affirmation would diminish math performance for White male participants, but
enhance performance for female participants. Two studies were conducted to evaluate these
hypotheses: 1) a lab-based study involving White male and mixed-ethnicity female
undergraduate college students, and 2) an online study involving young adult White male and
female participants. Results failed to support study hypotheses, with no significant differences
found between conditions on math performance among male or female participants. There were
also no significant differences between conditions on psychological outcomes (e.g., self-
integrity, social belonging). These findings are consistent with recent large-scale field replication
failures of self-affirmation interventions, suggesting that the phenomena may be more nuanced
and fragile than suggested by early research findings.
AFFIRMATION AND MAJORITY STUDENTS 2
Background and Significance
Over the past two decades, self-affirmation has received a sizable degree of empirical
attention as a brief, cost-effective, intervention capable of reducing the academic achievement
gap between majority and stigmatized groups. Affirmations encourage individuals to consider
core personal values, and reflect on the importance of these values in their lives. Values
affirmation interventions have been shown to improve the academic performance of African
American and Latino middle-school students (Cohen, Garcia, Apfel, & Masters, 2006; Sherman
et al., 2013), female undergraduate physics students (Mikaye et al., 2010), first-generation (i.e.,
those without a parent with a 4-year degree) undergraduate biology students (Harackiewicz et
al., 2014), and Latino undergraduate students (Brady et al., 2016).
How Do Affirmations Work?
According to self-affirmation theory, values affirmation inventions are effective because
they reduce the performance limiting effects of stereotype and identity threat, the fear that one
will be judged negatively in light of their ethnic background, gender, or other social identities
(Sherman & Cohen, 2006). Lab-based experimental studies have demonstrated that affirmations
completed directly before a threatening academic task can improve the academic performance of
threatened students (Martens, Johns, Greenberg, & Schimel, 2006; see Sherman & Cohen, 2006
for review). Additionally, randomized field trials of affirmation interventions have shown that
only a few brief affirmation writing exercises throughout the academic year can improve the
overall GPA of minority students over the course of two years (Cohen et al., 2006; Cohen,
Garcia, Purdie-Vaughns, Apfel, & Brzustoski, 2009). In these real-world settings, values
affirmation interventions have been shown to significantly reduce the academic achievement gap
between marginalized and majority students (see Cohen & Sherman, 2014, for a review).
AFFIRMATION AND MAJORITY STUDENTS 3
How can a brief 10 to 15-minute writing exercise can have such a profound and long-
standing effect on academic achievement? In academic settings, where students are subject to
routine evaluation, underrepresented minority students and women in math and science must
contend with the additional worry that their performance may confirm negative stereotypes
about the intellectual ability of their group (Spencer, Steele, & Quinn, 1999; Steele, 1997; Steele
& Aronson, 1995). This largely unconscious worry or threat can impair students’ working
memory and executive functioning, and significantly dampen performance on cognitively
demanding tasks (Beilock, Rydell, & McConnell, 2007; Schmader, Johns, & Forbes, 2008).
Over time, the effects of this chronic threat compound and lead vulnerable students to question
their sense of belonging in their academic institutions and to deidentify with academics (Cook,
Purdie-Vaughns, Garcia, & Cohen, 2011; Steele, 1997). Interrupting the deleterious effects of
stereotype threat can thus bolster students’ performance immediately and in the long-term.
Hence, affirmations do not work by boosting performance per se. Instead, they mitigate the
impact of social identity threat, and allow the performance of marginalized students to return to
where it would be in the absence of such threat (Cohen & Sherman, 2014). When students
vulnerable to stereotype threat complete affirmations before an academic stressor, they can
reflect on important transcendental values unrelated to the domain of the threat (e.g.,
relationships, hobbies). This helps bring about a more expansive view of the self and broadens
their sense of self-construal (Critcher & Dunning, 2015; Sherman et al., 2013). Consequently,
academic stressors are interpreted differently and no longer threaten students’ sense of academic
belonging (Brady et al., 2016; Cook et al., 2011; Layous et al., 2017). These underlying
processes are believed to negate the effect of social identity threat, and help explain how
AFFIRMATION AND MAJORITY STUDENTS 4
affirmation interventions can have cascading long-term benefits that alter the academic trajectory
of stigmatized students years later (Goyer et al., 2017; Tibbetts et al., 2016).
Evidence of Performance Hindering Effects of Affirmation?
Despite encouraging results showing that affirmation can reduce the academic
achievement gap between marginalized and majority students, an emerging small body of
literature suggests that affirmations may decrease academic performance for some students.
Because affirmations are thought to only affect threatened students, they should have no effect
on the academic performance of nonthreatened White students (Cohen et al., 2006; 2009).
Findings of reduced academic performance for nonstereotyped students run contrary to self-
affirmation theory, and have appeared unexpectedly and sporadically across studies. It is
possible that such findings are spurious and indicative of the multiple-comparison problem (Dee,
2015). With no established theoretical support for these findings, they are typically minimized,
and little research has directly investigated the potentially negative effects of self-affirmation
(Vohs, Park, & Schmeichel, 2013). However, as research findings showing that affirmation can
reduce performance continues to mount (Brady et al., 2016; Dee, 2015; Kizilcec, Saltarelli,
Reich, & Cohen, 2017), it appears that the effect warrants increased attention.
At least five studies have reported results suggesting that affirmations may have a
negative effect on the academic performance of majority students (Brady et al., 2016; Cook et
al., 2011; Mikaye et al., 2010; Shnabel, Purdie-Vaughns, Cook, Garcia, & Cohen, 2013; Woolf,
McManus, Gill, & Dacee, 2011). First, a randomized trial by Woolf and colleagues (2009)
examined the effects of a values-affirmation intervention on the academic performance of ethnic
minority and White undergraduate medical students in the United Kingdom. The ethnic minority
AFFIRMATION AND MAJORITY STUDENTS 5
students were a heterogeneous group consisting primarily of South Asian (i.e., Asian Indian,
Pakistani, and Bangladeshi) students. Because South Asian students have been found to
underperform academically when compared with White students (Woolf, Potts & McManus
2011), the researchers hypothesized that their performance may be impaired by stereotype threat,
and reasoned that they might benefit from a values affirmation intervention in the same manner
as negatively stereotyped African American middle-school youth (Cohen et al., 2006; 2009).
Woolf et al. (2009) found that affirmation was effective at reducing the achievement gap
between ethnic minority and White students on an end-of-year written medical assessment, but
in an unexpected way – it had an iatrogenic effect on the performance of White students.
Whereas ethnic minority students in the affirmation condition performed no better on the written
assessment than those in the control condition, White students in the affirmation condition
performed significantly worse than those in the control condition. This pattern of results was
limited to the written assessment (primary outcome measure), and was not replicated on a post-
intervention clinical assessment (secondary outcome measure) in which both affirmed ethnic
minority and White students outperformed their unaffirmed peers. The authors speculated that
the benefits of affirmation on the clinical assessment may have been related to the nature of the
exam, which involved face-to-face interaction with examiners. Since affirmations can lead to
feelings of love and connectedness towards others (Crocker, Niiya, & Mischkowski, 2008), they
may have facilitated more positive in-person interactions between students and examiners, and
thus improved performance. Presumably perplexed by the iatrogenic results of affirmation on the
written exam, the researchers failed to offer any explanation.
A second study evaluated the effects of affirmation on the academic performance of
undergraduate physics students. Miyake and colleagues (2010) randomly assigned
AFFIRMATION AND MAJORITY STUDENTS 6
predominately White students to a values affirmation intervention or neutral writing condition.
Because implicit stereotypes can suppress women’s academic performance (Spencer et al.,
1999), the researchers hypothesized that affirmation would mitigate threat among female
participants in a male-dominated science major, and reduce the gender achievement gap. As
predicted, the intervention lifted the performance of affirmed female students and significantly
reduced the gender achievement gap in physics exam scores and end-of-semester grades.
However, when examining the impact of the intervention on male students, an unexpected
pattern emerged – the affirmation had a negative effect. Affirmed males had significantly lower
exam scores and marginally lower course grades relative to their unaffirmed peers. The
researchers speculated that for male participants, the affirmation intervention could have led
them to divert effort from academics and invest it in other valued domains.
In a recent study, Brady and colleagues (2016) conducted a lab-based randomized trial of
a values affirmation intervention among Latino and White undergraduate students at an elite
university. Participants completed a single values affirmation in the lab in their first or second
year of college. This one intervention was found to improve the GPA of Latino students over the
course of two years and reduce the achievement gap between Latino and White students by 90%.
Affirmed Latino students also benefited psychologically. At 2-year follow-up, they reported a
greater sense of self-integrity, self-esteem, hope, and academic belonging compared with those
in the control group. The intervention had the opposite effect on White students’ academic
performance – affirmed White students had lower GPAs at 2-year follow-up compared with their
unaffirmed peers. There were no significant differences in psychological outcomes among White
participants at follow-up, leaving questions regarding why the intervention resulted in decreased
academic achievement for these students.
AFFIRMATION AND MAJORITY STUDENTS 7
To elucidate the mechanisms through which affirmation affects academic performance,
Cook and colleagues (2011) conducted a study examining the effect of intervention timing on
academic belonging (i.e., how much students felt they fit in at school). White and Black middle-
school youth were randomized to either receive a values affirmation the third day of 7th grade or
four weeks later. The authors hypothesized that African American students with a history of low
academic performance would benefit more from the earlier intervention than students who
received the later intervention. They reasoned that the beginning of the school year would be a
more sensitive period, and that fortifying their sense of belonging earlier would improve
academic outcomes. Results showed that for low-performing Black students, those who received
the earlier intervention reported a greater sense of belonging, and showed less of a decline in
their academic performance over the course of two years, compared with those who received the
later intervention. With respect to feelings of belonging and intervention timing, the opposite
pattern emerged for low-performing White students – those who received the earlier intervention
reported a lower sense of belonging than those receiving the later intervention. This pattern of
results did not hold for high-performing White students, who appeared to marginally benefit
from the earlier intervention. Given that the earlier intervention was more effective at buffering
feelings of belonging among African American youth, this indirectly suggests that the earlier
affirmation may have had a negative effect on low-performing White students. The researchers
speculated that the affirmation may have reminded low-performing White students of their poor
performance in the previous year, led them to devalue academics, and lowered their sense of
belonging as a result. Overall, these mixed findings suggest that for some White students, the
cognitive processes that affirmation triggers may differ and paradoxically decrease belonging
rather than buffer it. However, because all participants received an affirmation, it is premature to
AFFIRMATION AND MAJORITY STUDENTS 8
conclude that affirmation had a direct iatrogenic effect on academic belonging among White
students.
To understand the role of writing about social belonging themes on affirmation effects,
Shnabel and colleagues (2013) conducted two studies directly investigating these effects. In the
first study, the researchers conducted a content analysis of the original middle-school youth
affirmation essays collected by Cohen and colleagues (2006). They found that, regardless of
race, youth in the affirmation condition were much more likely to write about the theme of social
belonging (i.e., how their personal values connected them with others) than those in the control
condition. Mediation analysis showed that Black students in the affirmation condition were more
likely to write about belonging and this, in turn, was associated with an increase in GPA. White
students in the affirmation condition were also more likely to write about social belonging;
however, an opposite pattern developed for academic performance. In both the affirmation and
control condition, writing about belonging was associated with a decrease in White students’
GPA. Perhaps once these students’ feelings of social connectedness were affirmed, they may
have felt free to subtly disengage from academics (Miakye et al., 2010; Vohs et al., 2013).
In a second study, Shnabel and colleagues (2013) directly tested the hypothesis that
writing about social belonging accounts for the benefits of affirmation. They randomly assigned
predominately White undergraduates to one of four conditions: (1) standard affirmation, (2)
neutral writing control, (3) individuating-affirmation, and (4) belonging-affirmation. The first
two conditions replicated the intervention and control conditions used in prior values affirmation
research (Cohen, Aronson, & Steele, 2000). The third condition, the individuating-affirmation,
differed in that it instructed participants to write a brief essay on how their top ranked value
made them “feel independent and self-sufficient.” Finally, the fourth condition, the one of
AFFIRMATION AND MAJORITY STUDENTS 9
primary interest, instructed participants to write about how their top ranked value made them
“feel closer and more connected with people.” Following the affirmation, all participants
completed a challenging math test designed to induce social identity threat (i.e., students were
told the test was diagnostic of math ability).
As the researchers hypothesized, results showed that the prevalence of social belonging
themes differed by condition. Only 15% of essays included belonging themes in the neutral
writing control condition, whereas nearly all (97%) belonging-affirmation essays included
belonging themes (versus 60% for individuating-affirmation, and 63% for standard affirmation).
When comparing the math performance of males and females in the neutral writing control
condition, there was a significant gender gap, with males outperforming females. In the
individuating-affirmation and the standard-affirmation conditions, female performance improved
and the gender gap was eliminated. However, in the belonging-affirmation condition, the gender
gap was reversed, with female participants performing significantly better than male participants.
Consistent with results of their first study, there was a trend for conditions that elicited more
writing about belonging to be associated with poorer performance among White males. Males
completing the belonging-affirmation condition performed marginally worse than males in the
other three conditions, suggesting the adapted affirmation had an iatrogenic effect. In two
separate studies with different samples (middle-school students vs. undergraduate college
students), Shnabel et al. (2013) found evidence that writing about social belonging resulted in
decreased academic performance among presumably nonthreatened White students. This
suggests that one of the active ingredients of values affirmation interventions, affirming social
belonging, may be detrimental to majority students in some contexts.
AFFIRMATION AND MAJORITY STUDENTS 10
Affirmation Benefiting Majority Students?
The studies reviewed above provide examples of when affirmation reduced academic
outcomes of majority students. Yet, there also exists evidence that White students may benefit
academically from affirmation. In a study with an ethnically mixed sample of low-income
middle-school youth, Bowen and colleagues (2013) evaluated the effects of an affirmation
adaptation designed to enhance intervention outcomes. The researchers reasoned that teachers
that were made aware of their students’ values would develop more positive expectancies of
them and that this, in turn, would bolster intervention effectiveness. Results showed that,
compared with students in the neutral writing control, those who completed an affirmation did
not experience the decline in course grades that is typical of negatively stereotyped youth.
Moreover, those students who completed an affirmation and had their teachers read their essay
had higher end-of-year grades than those who completed the affirmation only. A somewhat
surprising result was that the affirmation intervention appeared equally efficacious for African
American and non-African American youth (e.g., Asian, European American, multiracial).
Although this may have been due to limited statistical power to detect differences between
African Americans and other ethnic groups, which were relatively small in size, the researchers
speculated that the high poverty rates of school youth may have subjected them to other class-
based stereotypes. Under such conditions, these youths may have been negatively impacted by
social identity threat, enabling affirmed youth to benefit from the intervention regardless of
ethnicity
To evaluate whether affirmation effects extend to college generation status, a proxy for
socioeconomic status, Harackiewicz and colleagues (2014) evaluated the effects of an
affirmation intervention with predominately White first-generation (i.e., those without a parent
AFFIRMATION AND MAJORITY STUDENTS 11
with a 4-year degree) and continuing-generation (i.e., those with at least one parent with a 4-year
degree) students in an introductory biology class. The researchers hypothesized that first-
generation students face unique psychological challenges (e.g., uncertainty about academic and
social belonging) in adapting to the individualistic culture of elite American universities
(Stephens, Fryberg, Markus, Johnson, & Covarrubias, 2012a). Due to these additional obstacles,
they reasoned that first-generation students would benefit from an affirmation intervention
regardless of ethnicity. Indeed, the values affirmation intervention significantly improved the
academic outcomes of first-generation students, and narrowed the social-class achievement gap
by 50%. Compared to control, affirmed first-generation students were more likely to earn Bs
than Cs, and were more likely to enroll in a higher-level biology class the following semester.
The benefits of the affirmation intervention were long-lasting, and affirmed first-generation
students had significantly higher GPAs at 3-year follow-up compared with their unaffirmed
peers (Tibbett et al., 2016). Interestingly, intervention effects were isolated to first-generation
students and did not extend to a combined group of first- and continuing-generation
underrepresented minorities. The findings suggest that it was the intervention’s impact on factors
relevant to first-generation students (e.g., cultural mismatch, belonging uncertainty), not
ethnicity, which accounted for the positive outcomes. Thus, it appears that affirmations can
benefit first-generation college students regardless of ethnicity.
In a recent experimental study, Layous and colleagues (2017) conducted a values
affirmation intervention among a predominately White sample of undergraduate students in an
introductory psychology class. Compared with those in the control group, affirmed students saw
an increase in GPA over three semesters. Surprisingly, the benefits of the affirmation appeared
to extend to males only, and were particularly pronounced among males who had low pre-
AFFIRMATION AND MAJORITY STUDENTS 12
intervention GPAs and reported a low sense of social belonging. Vulnerable males in the control
condition saw a decline in GPA over three semesters, whereas those in the affirmation condition
saw an increase in GPA over the same time period. These findings bolster support that
affirmation may be beneficial for majority students at times, particularly if they are
psychologically threatened.
Taken together, the studies reviewed previously demonstrate that affirmation may have
null, beneficial, or iatrogenic effects on White students’ academic performance. Typically, White
students are not suspected to be at-risk of social identity threat in academic settings, and early
affirmation research showed that they did not benefit from affirmation interventions (Cohen et
al., 2006; 2009; Sherman et al., 2013). More recent research suggests that White students who
are low-income, first-generation, or psychologically threatened may benefit from affirmation
(Bowen, Wegmann, &Webber, 2013; Harackiewicz et al., 2014; Layous et al., 2017). Note,
however, that the available evidence suggests that the most privileged subgroup, White male
continuing-generation students, may be most likely to experience negative effects from
affirmation.
Why Might Affirmation Hinder Performance of Majority Students?
Why might some majority students be negatively affected by affirmation? To understand
the iatrogenic effects of affirmation on academic performance, we draw on cultural mismatch
theory and stereotype lift theory. Cultural mismatch theory posits that there is a cultural
mismatch between the interdependent norms of first-generation students and the independent
norms of American universities (Stephens et al., 2012a). These students tend to endorse more
interdependent and fewer independent motives for attending college relative to continuing-
generation students (Harackiewicz et al., 2014; Stephens et al., 2012a). Because of this cultural
AFFIRMATION AND MAJORITY STUDENTS 13
mismatch, first-generation students face additional psychological challenges in adapting to a
college culture that places increased emphasis on the individual and independence above other
values. Such a mismatch can create additional stress for first-generation students, undermine
their academic performance, and ultimately contribute to the social-class achievement gap
(Stephens, Townsend, Markus, & Phillips, 2012b; Stephens, Markus, & Phillips, 2014a).
Interventions that framed colleges’ communication with first-generation students using
interdependent norms (Stephens et al., 2012a), or that conveyed that students with different
social-class backgrounds were valued (Stephens, Hamedani, & Destin, 2014b), reduced
biological markers of stress and improved the academic performance of these students (Stephens
et al., 2012b; Stephens, Townsend, Hamedani, Destin, & 2015).
Continuing-generation students do not confront the same challenges as their first-
generation peers in adapting to college life, as their individualistic values are generally thought
to be more congruent with university culture (Stephens et al., 2014a). That said, they may still be
susceptible to threat derived from a cultural mismatch between their independent values and
affirmations focused on interdependent themes. Recall that Shnabel et al. (2013) found that
White male participants saw a performance decrease in the belonging-affirmation condition, but
not in the individuating-affirmation or control condition. Only when male participants were
explicitly directed to write about how their core values connected them with others, did they
experience a dip in performance. Perhaps, affirmations focused on social connectedness conflict
with these students’ need to feel independent and distinct and result in an alternate cultural
mismatch (i.e., independent values and an interdependent task). In the absence of psychological
threat, a cultural mismatch between a belonging-affirmation and the independent values of
majority students could attenuate performance. With this line of reasoning, we would expect
AFFIRMATION AND MAJORITY STUDENTS 14
nonthreatened majority students to perform worse on psychological outcomes (e.g., self-
integrity, social belonging) following a belonging-affirmation intervention. This has yet to be
investigated, but there is some evidence to suggest that interventions focused on fortifying
students’ belonging can decrease feelings of belonging among White students (Walton & Cohen,
2007).
Stereotype lift theory provides a competing hypothesis for why affirmation may attenuate
nonthreatened students’ performance. Stereotype lift is a performance boost that individuals gain
from making downward social comparisons to negatively stereotyped groups (Walton & Cohen,
2003). When nonthreatened group members (e.g., White males in science and math) know their
performance is being compared to a denigrated group (e.g., underrepresented minorities,
women), downward social comparisons are thought to increase their self-efficacy and
motivation. This lift in confidence allows these individuals to persevere in the face of adversity
and enhances their performance (Walton & Cohen, 2003). Because affirmations can increase
feelings of connectedness towards others (Crocker et al., 2008), they may reduce the likelihood
of majority students making negative “stereotype-inspired social comparisons” (Walton &
Cohen, 2003). A reduction in stereotype lift may explain why majority students in the control
condition in affirmation studies sometimes outperform those receiving the affirmation
intervention (Brady et at., 2016; Woolf et al., 2009).
This theory appears to be roughly in line with the reasoning advanced by Miyake and
colleagues (2010) in explaining why male physics students in the affirmation condition
performed worse than those in the control condition. For majority students, affirmations may
divert attention away from their privileged social identities (i.e., students who perform better
academically compared with marginalized groups). Rather than allowing these students to stay
AFFIRMATION AND MAJORITY STUDENTS 15
academically engaged in the face of adversity – as is the case with stigmatized students –
affirmations may lead these majority students to reinvest their time in other valued life domains.
If this reasoning holds, we would expect affirmation to lead to decreased academic performance
among majority students, but have limited, or perhaps even beneficial, effects on psychological
functioning (e.g., self-integrity, social belonging).
To summarize, cultural mismatch theory suggests that a mismatch between values and
context can reduce academic performance. Affirmations focused on interdependent values (e.g.,
belonging-affirmation) may subtly threaten independently-oriented majority students, and
decrease their academic performance as a result. By comparison, stereotype lift suggests that
affirmations may reduce majority students’ academic performance not by subtly communicating
threat (e.g., cultural mismatch), but by removing a psychological advantage these students
receive from comparing themselves to negatively stereotyped groups. These theories provide a
preliminary framework for why affirmations may decrease the performance of majority students.
However, with little research directly investigating the performance hindering effects of
affirmation, there is limited empirical evidence to support the relevance of these theories to
affirmation interventions. The first step in understanding the potential negative consequences of
affirmation on majority students is to directly examine the effects of affirmation with this
population specifically.
Current Studies
We conducted two experimental studies to examine whether an affirmation focused on
social connectedness (i.e., belonging-affirmation) negatively affects the math performance and
psychological functioning of White male students. In the first study, White male and mixed-
ethnicity female undergraduate students completed a lab-based belonging-affirmation
AFFIRMATION AND MAJORITY STUDENTS 16
intervention, and we evaluated its effects on math performance, self-integrity, and social
belonging. The aim of this study was to replicate previous findings (Shnabel et al., 2013)
demonstrating that a belonging-affirmation improves the math performance of female
undergraduates while hindering the performance of White male undergraduates. In a second
study, we again sought to evaluate the effects of a belonging-affirmation on majority students.
We recruited an online sample of White male and female young adults through Amazon
Mechanical Turk (MTurk). This online study examined the effects of a belonging-affirmation on
math performance, self-integrity, and personal mastery, and tested for potential moderators of
affirmation effects (i.e., math identity, cultural values).
Study 1
Because Study 1 was intended to replicate prior research findings, study design and
procedures were similar to Shnabel et al. (2103), with a few exceptions. First, because our
primary population of interest was White male students, males of other ethnic backgrounds were
excluded from participation. Asian female students were also excluded from participation due to
the possibility of positive stereotypes regarding Asians and math ability (e.g., stereotype boost,
stereotype susceptibility) obfuscating the effects of the affirmation intervention (Armenta, 2010;
Gibson, Losee, & Vitiello, 2014; Shih, Pittinsky, & Ambady, 1999). This effect was not
expected to occur for women of other ethnic backgrounds, as we reasoned that women from
underrepresented backgrounds (e.g., African American, Latina) who are negatively stereotyped
in math based on two identities (gender and ethnicity) are more likely to benefit from
affirmation. Second, because we were specifically interested in the effects of the belonging-
affirmation, participants were randomized to complete only the belonging-affirmation or neutral
writing control (i.e., we did not include the standard affirmation or independent-affirmation
AFFIRMATION AND MAJORITY STUDENTS 17
conditions). Third, we included two post-intervention measures to assess the effect of the
affirmation manipulation on psychological factors (i.e, self-integrity, social fit). We
hypothesized that among female participants, affirmation would lead to enhanced math
performance, ratings of self-integrity, and feelings of social belonging compared with control.
By comparison, among White male participants, we hypothesized that the belonging-affirmation
would lead to diminished math performance, ratings of self-integrity, and feelings of social
belonging compared with control.
Method
Participants
One hundred twenty-two undergraduate college students were recruited using the USC
subject pool and were compensated with course credit for their participation. Two participants
did not complete the affirmation exercise correctly (e.g., the second page of the exercise was left
blank), and thus were excluded from subsequent analyses. The final sample consisted of 40
White male and 82 mixed ethnicity female participants. Ages ranged from 18 to 24 years old (M
= 19.92, SD = 1.43). Among female participants, the majority identified as White (69.5%) with
the remaining participants split among those identifying as Black (11%), Latino (9.8%), and
biracial (9.8%). Most participants (82.3%) were continuing-generation students (see Table 1).
Procedures
Participants were run individually or in groups of up to seven participants at a time, but
completed all tasks at separate desks and did not interact throughout the study. At the beginning
of the experiment, participants were greeted by the experimenter (who was blind to condition)
and provided informed consent. They were instructed that they would be participating in two
ostensibly separate studies: one examining students’ values, and a second focused on how
AFFIRMATION AND MAJORITY STUDENTS 18
students solve math problems. Once seated, participants were provided a large sealed envelope
with their names written on it, containing the affirmation writing exercise materials. All
participants were presented a list of 11 values and asked to rank them in order of personal
importance (i.e., being good at art, physical attractiveness, creativity, independence, membership
in a social group, music, politics, relationships with friends or family, religious values, sense of
humor, and sports ability). Next, participants completed one of two randomly assigned hand-
written exercises: 1) belonging-affirmation, or 2) neutral writing control (Shnabel et al., 2013).
Participants in the belonging-affirmation condition were instructed to think about a time when
their top ranked value made them feel closer with others, and write a brief essay about why it
made them feel closer and more connected (see Appendix A). Participants in the neutral writing
control were instructed to think about a time when an unimportant value (i.e., 9
th
ranked value)
was important to someone else, and write a brief essay about why it would be important to
someone else (see Appendix B). They were given 10 minutes to complete the writing exercise
and were instructed to reflect on the value they wrote about if they finished early. Following
completion of the affirmation writing exercise, participants were thanked and told they would be
beginning the “second” study, which was completed in the same room with the same
participants.
In the ostensible second study, participants were given a very challenging math test and
told that it was diagnostic of math ability. The test was labeled “USC Math Assessment Test”
and included 30-multiple choice questions drawn from difficult GRE math problems (adapted
from Shnabel et al., 2013; see Appendix C). Participants were instructed that they had 20
minutes to complete the test and could not use a calculator. Further, they were instructed that in
order to get the best assessment of their current math ability, it was very important for them to
AFFIRMATION AND MAJORITY STUDENTS 19
give their best effort. Following completion of the math test, participants completed several post-
intervention measures, including a brief demographic questionnaire, a measure of self-integrity,
a measure of social fit, and an awareness probe. The awareness probe was intended to assess
what factors participants’ believed contributed to their math performance, and whether they
suspected the values writing exercise was intended to influence their math performance.
Following completion of these measures, participants were debriefed and thanked for their
participation.
Measures
Demographic and personal characteristics. Participant age, gender, ethnicity, transfer
status, grade level, college generation status (first-generation vs. continuing-generation), and
SAT/ACT scores were assessed through self-report at the end of the study (i.e., after completion
of the other post-intervention measures).
Self-integrity. The 8-item Self-Integrity Scale (SIS) was used to assess the extent to
which participants feel a general sense of efficacy and moral adequacy (adapted from Sherman
et al., 2009; see Appendix D). Items include statements such as “I have the ability and skills to
deal with whatever comes my way” and “I feel that I’m basically a moral person.” Participants
rated items on a scale of 1 (Strongly disagree) to 7 (Strongly agree), with higher scores
indicating a greater sense of self-integrity. The SIS has been used as a psychological outcome
measure in prior affirmation research (Sherman et al., 2009), and demonstrated acceptable
reliability with the present sample (α = .79).
Social fit. The six-item Social Fit Questionnaire was used to assess social fit at college
(adapted from Walton & Cohen 2007; see Appendix E). Measure items include statements such
as “I feel like I belong as a student at USC” and “I feel like I fit in with the academic community
AFFIRMATION AND MAJORITY STUDENTS 20
at USC.” Participants rated their agreement with these statements on a scale of 1 (Strongly
disagree) to 7 (Strongly agree). The mean of these six items served as our measure of social
belonging, with higher scores indicating a greater sense of social fit (α = .80).
Awareness probe. The awareness probe was included to elicit factors that participants
believe affected their math performance (adapted from Sherman et al., 2009; see Appendix F).
The measure first asks in an open-ended fashion for participants to list any factors that might
have contributed to their performance. Then, participants are asked to estimate the degree to
which 12 specific factors (i.e., math ability, effort on the test, personal background, beliefs, and
attitudes, familiarity with math problems, first writing exercise, reasoning ability, self-esteem,
mood, gender of the research assistant, time of day, and today’s weather) contributed to their
math performance. Participants rated each factor on a scale of 1 (Did not contribute at all) to 9
(Contributed a great deal). Open-ended responses were reviewed by a research assistant to
determine whether participants suspected that the affirmation writing exercise was intended to
influence participants’ performance on the math test.
Results
Participant Values and Manipulation Check
During the values ranking task, across conditions, most participants (69.4%) selected
“relationships with friends and family” as their top value. Next, “religious values” (13%) and
“independence” (12%) were selected most frequently. All other values were infrequently ranked
(n < 3; see Table 2). To assess whether our manipulation was effective, we coded participants’
affirmation essays to determine whether the two conditions successfully produced different
frequencies of social belonging themes. One trained undergraduate coder evaluated whether each
AFFIRMATION AND MAJORITY STUDENTS 21
essay included a social belonging theme based on the coding instructions used by Shnabel and
colleagues (2013) (see Appendix G).
As expected, the affirmation manipulation was effective at producing essays with
belonging themes in the belonging-affirmation condition but not the neutral writing control
condition. All essays in the belonging-affirmation condition included belonging themes whereas
none of the essays in the control condition included belonging themes. These results were similar
to Shnabel et al. (2013), who found that 97% of the essays in the belonging-affirmation
condition included social belonging themes, while only 15% of essays in the control condition
included belonging themes. Excerpts from essays in the belonging-affirmation are as follows:
“Maintaining good relationships with friends and family is extremely important to me
because it provides a feeling of support and togetherness. It allows you to feel like people
are looking out for you and care for you, especially in times when you need it.”
“Independence is a value that makes me feel closer and more connected with people
because without it, I would not be able to do things the way I want. For instance, I
consider myself to be independent because I live away from home and that independence
has allowed me to meet many new people here at school who I have grown so close with
in these short years. I think I feel so connected with these people because they are kind of
like a second family since I don’t see mine as often now.”
Excerpts from essays in the neutral writing control condition, which focused on why an
unimportant value (i.e., 9
th
ranked value), might be important to someone else included:
“Sports ability would be important to someone else who is reliant on their ability for
income purposes (professionals), continuation of a scholarship (college athletes), or even
AFFIRMATION AND MAJORITY STUDENTS 22
someone hoping to one day be a professional or make it into a better school than what
their grades reflect (high school students).”
“Being good at art would likely be important to someone who plans to be an artist or
perhaps has a major in art school. American society doesn’t put much importance on
being good at art, but in other cultures, art may have a great significance, and
consequently, a person may say that being good at art is particularly important to them.”
Awareness Probe Results
On an open-ended question where participants were instructed to report any factors that
contributed to their math performance, none of the participants reported that they believed the
first study (values affirmation) was intended to influence their performance on the math test. Of
eleven listed factors, “familiarity with math problems,” “math ability,” “reasoning ability,” and
“effort,” were the four highest rated by participants as factors affecting their performance (see
Table 3). The values affirmation exercise was rated as the least influential factor (11
th
) by male
participants, and the second least influential (10
th
) factor by female participants. Overall,
participants did not suspect that the values affirmation was intended to influence their
performance on the math test.
Primary Analyses
We first tested our primary hypotheses that (a) women’s math performance would be
enhanced in the affirmation compared to control, and (b) men’s math performance would be
poorer in the affirmation condition compared to control. A two-way ANOVA did not reveal any
significant main effects for affirmation condition, F (1, 121) < 0.01, p = .99, or gender, F (1,
121) = 3.07, p = .82, on number of problems correctly answered. Contrary to study hypotheses,
AFFIRMATION AND MAJORITY STUDENTS 23
there was no significant condition by gender interaction, F (1, 121) < 0.01, p = .98. The number
of problems correctly answered by affirmed females (M = 3.59, SE = 0.38), was similar to that of
unaffirmed females (M = 3.61, SE = 0.37) (see Table 4). Likewise, the math performance of
affirmed males (M = 4.43, SE = 0.55) and unaffirmed males (M = 4.42, SE = 0.54) was nearly
identical. Overall, the affirmation manipulation had no effect on participants’ math performance
(see Figure 1).
Next, we evaluated our secondary hypotheses that (a) women’s ratings of self-integrity
and social fit would be enhanced in the affirmation condition compared to control, and (b) men’s
ratings of self-integrity and social fit would be poorer in the affirmation condition compared to
control. With respect to ratings of self-integrity, there were no main effects of condition, F (1,
121) = 0.44, p = .51, or of gender, F (1, 121) = 0.24, p = .63. The condition by gender interaction
was also not significant, F (1, 121) = 0.65, p = .42. Perceptions of self-integrity were similar
among females in the affirmation condition (M = 45.39, SE = 0.81) and females in the control
condition (M = 46.84, SE = 0.70). Among male participants, perceptions of self-integrity were
also similar between those who were affirmed (M = 46.67, SE = 1.12) and those who were
unaffirmed (M = 46.53, SE = 1.33) (see Table 4).
Concerning ratings of social fit, there were also no main effects of condition, F (1, 121) =
1.35, p = .25, or of gender, F (1, 121) = 2.11, p = .15. The condition by gender interaction was
also not significant, F (1, 121) = 0.02, p = .89. Affirmed females’ ratings of social fit (M =
33.20, SE = 0.74) were not significantly different from those of unaffirmed females (M = 34.29,
SE = 0.88), and affirmed males’ ratings of social fit (M = 31.52, SE = 1.26) were not
significantly different from those of unaffirmed males (M = 32.89, SE = 1.48) (see Table 4).
Thus, participants’ perceptions of self-integrity or social fit appeared unaffected by affirmation.
AFFIRMATION AND MAJORITY STUDENTS 24
Supplementary Analyses
Our procedures differed from Shnabel and colleagues (2013) in several ways that could
have contributed to our outcomes (i.e., no significant differences in math test performance
between conditions for male or female participants). Shnabel and colleagues (2013) reasoned
that the presence of other participants, particularly males, may help increase female participants’
sense of evaluative pressure during the math test and cue stereotype threat. Consequently, they
had participants complete the study procedures in mixed-gender groups of three to nine
participants. In line with the procedures by Shnabel et al. (2013), we attempted to have
participants complete the study in mixed-gender groups. However, logistical constraints
interfered with our ability to follow these procedures precisely, and some participants completed
the study individually and without the presence of a male participant. Within our sample, most
participants (72.1%) completed the study in a group of two or more (2-7 participants), and
approximately half the participants (48.4%) completed the study in a mixed-gender cohort (i.e.,
at least one male and female participant present). Most participants had a female experimenter
(80.3%). Some prior research on stereotype threat and women’s math performance used a White
male experimenter to help cue stereotype threat (Martens et al., 2006; Schmader, 2002).
Research on the effects of affirmation interventions have used both male and female
experimenters (Brady et al., 2016; Shnabel et al., 2013).
To assess whether participant characteristics or differences in study design and
procedures (e.g., experimenter gender, group size) contributed to study outcomes, we conducted
a series of exploratory analyses. We evaluated whether the following impacted math test
performance or psychological outcomes (i.e., self-integrity, social fit): 1) minority status
(underrepresented vs. majority); 2) college generation status (first-generation vs. continuing-
AFFIRMATION AND MAJORITY STUDENTS 25
generation); 3) grade level (underclassman vs. upperclassman); 4) transfer student (yes vs. no);
5) experimenter gender (male vs. female); 6) group status (individual vs. group); and 7) gender
composition (mixed vs. homogenous). We conducted these analyses using Welch’s t-tests for the
pooled sample and by gender. This resulted in 63 between-group comparisons (3 samples
[pooled sample, men only, women only], 7 factors, and 3 outcome measures). Of these
comparisons, only one reached significance – upperclassmen reported a greater sense of self-
integrity compared with underclassmen, Welch’s t(99.91) = 4.42, p < .05. The remaining
comparisons produced nonsignificant results (ps > .05). To explore the possible effects of group
size and gender ratio on outcomes, we computed correlations between these two factors and our
three outcomes measures (i.e., math test performance, self-integrity, and social fit) for the pooled
sample and by gender. We did not observe any significant associations between participant
cohort characteristics and our outcome measures (ps > .11) Given the relative lack of significant
findings among these exploratory analyses, it is less likely that any one of these factors unduly
influenced our results.
Lastly, there was a substantial gender gap in our sample size, with the number of female
participants (n = 82) being more than twice the number of male participants (n = 40). This
resulted in large unequal cell sample sizes, which can significantly reduce the power of ANOVA
(Wilcox, 2012). To account for this difference, we separated groups by gender and conducted
between-group comparisons (affirmation vs. control) using nonparametric tests (e.g., Welch’s t-
test, Mann-Whitney U test). Using these robust methods, we were able to compare outcomes by
gender (i.e., males separately, females separately), without violating ANOVA statistical
assumptions of normality and homoscedasticity (Wilcox, 2012). These analyses were consistent
AFFIRMATION AND MAJORITY STUDENTS 26
with the results of our two-way ANOVA tests, and did not reveal any significant differences
between conditions for any of our three outcome measures (ps > .24).
Discussion
Results of Study 1 failed to support our hypotheses. The affirmation manipulation did not
significantly affect math performance or psychological outcomes (i.e., self-integrity, social fit) of
male or female participants. This was the case when analyzing the data together or separately by
gender. Our results are inconsistent with those of Shnabel and colleagues (2013), who found that
men outperformed women in the control condition, but saw this pattern reversed in the
belonging-affirmation condition, where women outperformed men. We attempted to replicate
procedures of Shnabel et al. (2013) and used a comparable sample (e.g., undergraduate college
students recruited from psychology classes), the same study instructions (e.g., participants were
told they were participating in two separate studies), the same math test, and the same
affirmation materials. However, as described previously, our procedures differed in that some
participants completed the experimental procedures individually, and some did not complete
them in a mixed-gender group (i.e., only women were present). Prior research suggests that
affirmations can be effective when completed individually or in group settings (Covarrubias,
Herrmann, & Fryberg, 2016; Tibbetts et al., 2016). We conducted exploratory analyses to see
whether these factors (e.g., group size, gender composition) influenced our outcomes, and did
not find evidence to support this. The results of these exploratory analyses are consistent with a
recent meta-analysis of stereotype threat and women’s math performance which did not identify
an effect of gender composition (e.g., women only groups, majority women groups) on
stereotype threat effects (Picho, Rodriguez, & Finnie, 2013).
AFFIRMATION AND MAJORITY STUDENTS 27
A second possible explanation for the null results regarding math performance is that the
math test was too onerous for most participants. Although the test was multiple choice, nearly
half the sample (49%) answered fewer than four problems correctly. The mean number of
problems participants answered correctly was relatively low (M = 3.87, SD = 2.44; 30 problems
total), and problem accuracy (i.e., problems correct divided by problems attempted) was also low
(M = 0.31, SD = 0.19). This may have had the unintended consequence of limiting variability in
our outcome measure and masking the effects of affirmation. However, the average math test
scores of our female (M = 3.60, SD = 2.75) and male (M = 4.43, SD = 2.46) participants across
conditions were similar to the average test scores of female (M = 4.09, SD = 2.75) and male (M =
4.18, SD = 2.57) participants across the belonging-affirmation and the control condition reported
by Shnabel et al. (2013), suggesting that our sample was comparable in math ability and that the
math test used was a viable outcome measure. That stated, it is worth noting that this math test
has also been used in studies of affirmation effects as part of a two-test procedure. In this design,
the test is initially administered to cue stereotype threat in participants. Once cued, participants
are asked to complete a less challenging math test during which the performance buffering
effects of affirmation are observed (Sherman et al., 2009). Perhaps using this two-test procedure
or a moderately challenging math test may have allowed affirmation effects to emerge.
With respect to sample size, we recruited fewer White male participants (n = 40) than we
had initially proposed (n = 60). Of those participants, only 33 were continuing-generation
students, the population whose performance we reasoned would be impaired by the belonging-
affirmation. One could speculate that the small sample size could have rendered statistical power
too low to observe a performance diminishing effect of affirmation among males. There are
several limitations to this reasoning. First, our sample size was similar to Shnabel et al. (2013),
AFFIRMATION AND MAJORITY STUDENTS 28
who had 117 participants (62 males). Although we had fewer male participants, we had more
male and female participants per condition because we only compared the belonging-affirmation
with the neutral writing control (Shnabel et al., 2013 compared four conditions). Second,
because mean scores of math performance between affirmed and unaffirmed males were nearly
identical (Mdiff = 0.01), a larger sample would likely not have yielded significant results if mean
estimates are accurate. Third, because the relatively larger sample of female participants (n = 82)
also showed very small differences in mean scores of math performance between conditions
(Mdiff = 0.02), it is unlikely that sample size was the primary contributor to the null results.
Factors related to the study design (e.g., participants, outcome measures) and the phenomena
being examined (e.g., affirmation effects, stereotype threat) are the more likely culprits
impacting study outcomes.
A final limitation worth noting is that we did not assess for potential moderators of
affirmation effects (e.g., math identity, cultural values). Research on affirmation suggests that
participants’ identification with the domain of the perceived threat (e.g., academics, math) can
moderate intervention effects. Specifically, those who are high in academic or math
identification are thought to be more vulnerable to social identity threat in academic settings, and
may be more likely to benefit from affirmation (Schmader, 2002; Sherman et al., 2013). We
attempted to collect SAT/ACT scores as proxies for math identification, but only half the sample
(54%) reported these scores. For those who provided scores, there were no significant
differences by gender, and there were no significant differences between conditions on outcome
measures. With respect to cultural values, we speculated that the belonging-affirmation would be
beneficial for females but harmful for White males, possibly because it conflicted with their
more independent values. We did not, however, assess cultural values directly. This may be an
AFFIRMATION AND MAJORITY STUDENTS 29
important next step in understanding who may benefit from and who may be hindered by
affirmation.
Study 2
The goal of Study 2 was identical to Study 1 – to evaluate whether an affirmation
focused on social belonging (i.e., belonging-affirmation) could result in enhanced math
performance among White females, but diminished math performance among White males.
However, due to the null findings in Study 1, we made several changes to the procedures in
Study 2. These changes and their rational are discussed next.
First, to increase the sample of White males, we recruited a convenience sample through
Amazon Mechanical Turk (MTurk). In recent years, MTurk has gained popularity as an effective
crowdsourcing platform for recruiting participants for behavioral science experiments (see
Chandler & Shapiro, 2016, for a review). When compared with college student samples, MTurk
participants generated data with identical or superior data reliability (Behrend et al., 2011).
Moreover, experimental studies comparing effect size estimates of MTurk samples with college
student samples have generally found equivalent effects (Berinsky, Huber, & Lenz, 2012).
Although affirmation interventions are typically administered in academic settings and
affirmation essays are hand-written, research suggests that affirmations may also be effective
when completed online and in non-academic settings (Covarrubias et al., 2016; Kizilcec et al.,
2017; Townsend & Sood, 2012). Thus, MTurk appears to be a viable platform on which to
extend affirmation research and test study hypotheses.
Second, to facilitate a more refined evaluation of the effects of the belonging-affirmation
on majority students, we limited study participation to those identifying as European American,
AFFIRMATION AND MAJORITY STUDENTS 30
and who reported having at least one parent who received a 4-year degree (i.e., first-generation
students were excluded). We reasoned that including only White participants and those who
came from households with at least one parent who received a 4-year degree would allow us to
remove ethnicity as a confound from the female sample, and college generation status as a
confound from the male sample. This allowed for a clear assessment of affirmation effects on
White students.
Third, to evaluate potential moderators of affirmation effects, we assessed participants’
math identification and cultural values in a screening survey prior to the affirmation
manipulation. With regard to math identification, some research suggests that the effects of
stereotype threat may depend on the degree to which one is identified with the domain of threat.
Generally, those who are more identified are more at risk of stereotype threat, and thus have
more to gain from affirmation (Nguyen & Ryan, 2008; Sherman et al., 2009; Spencer et al.,
1999). Based on this literature, we hypothesized that those who were highly math-identified
would benefit more from affirmation than those who were low in math identification. We also
assessed participants’ independent and interdependent motives for attending college. Previous
research suggests that those with more interdependent values (i.e., Latinos) may be more likely
to benefit from affirmations focused on interdependence (Covarrubias et al., 2016). We
speculated that belonging-affirmations may hinder the performance of those more
independently-oriented students, while enhancing the performance of those who are more
interdependent in orientation.
Fourth, our instructions regarding the affirmation were different from Study 1 in that we
chose not to use deception (i.e., participants were not instructed that they were participating in
two separate studies). Instead, we remained vague about the purpose of the study and instructed
AFFIRMATION AND MAJORITY STUDENTS 31
participants that it was about “understanding values and problem-solving.” Recent research
suggests affirmations can be effective when participants are aware of their intent if they are
freely chosen (Silverman, Logel, & Cohen, 2013; Walton et al., 2014), so it seemed unnecessary
to directly deceive participants. Fifth, because we expected our online sample to be less skilled at
math than the sample in Study 1 (MTurk workers vs. college undergraduate students) and
wanted to reduce the likelihood of a floor effect, we used a different math test as our primary
outcome measure. The math test was comprised of 15 moderately challenging GRE problems, as
opposed to the most difficult GRE problems. Sixth, assuming it would not be a relevant outcome
to online participants due to their diverse locations, we chose not to evaluate participants’ sense
of social fit following the affirmation manipulation. Instead, we evaluated participants’ sense of
control (i.e., personal mastery), as we hypothesized this construct would capture feelings similar
to self-integrity and would be affected by the affirmation manipulation.
Method
Participants
Participants were recruited to complete the study based on meeting eligibility criteria
assessed through a screening survey advertised on MTurk as a “Survey of Academic
Background, Values, and Motivation.” A large number of MTurk workers took the survey (N =
1,265), and were compensated 25 cents for their participation. To meet eligibility criteria,
participants had to reside in the United States, identify as European American, have at least one
parent who completed a 4-year degree, and be between the ages of 18 to 35 years old. Of those
MTurk workers who completed the screening survey, 292 met eligibility criteria and were
contacted by email and invited to participate in the online affirmation study. Participants could
AFFIRMATION AND MAJORITY STUDENTS 32
complete the study at their convenience by following a link in the email, but were required to
complete it within three weeks of receiving the invitation. All participants were compensated $3
for completing the study.
The sample completing the study consisted of 198 participants. One participant did not
complete the affirmation essay and was thus eliminated from subsequent analyses. The final
sample consisted of 103 White male and 94 White female participants. Ages ranged from 18 to
35 years old (M = 28.25, SD = 4.36; see Table 5). Participants tended to be well educated, with
most having completed a 4-year degree or higher (61.9%). Most participants were also employed
full-time (61.4%).
Procedures
Procedures were similar to Study 1, but were completed online through MTurk and
Qualtrics. Participant demographic information and putative moderators of affirmation effects
(e.g., math identity, cultural values) were collected through the screening survey as to not
interfere with the affirmation manipulation. Participants were informed that the study was to
understand values and problem-solving. They were told they would be writing about commonly
held values and complete a problem-solving task. Participants who consented completed a value
ranking task which was tailored slightly to the online sample. This is consistent with
recommendations that values affirmations include values relevant to the persons completing the
exercise (Cohen, Aronson, & Steele, 2000). Participants ranked 12 (rather than 11) commonly
held values (e.g., career, the environment, creativity, independence, learning and gaining
knowledge, athletic ability, financial success, music, politics or government, relationships with
friends and family, spiritual or religious values, and sense of humor). They were then
AFFIRMATION AND MAJORITY STUDENTS 33
randomized to complete the belonging-affirmation or neutral writing control, which were
identical to those used in Study 1. Participants in the belonging-affirmation condition wrote
about how their top ranked value made them feel more connected with people. Participants in the
neutral writing control condition wrote about why their 9
th
ranked value may be important to
someone else. Following completion of the affirmation exercise, participants were instructed that
they had 15 minutes to complete a math test comprised of 15 problems. As in Study 1,
participants were instructed that it was very important for them to try their best on the math test
to get the best assessment of their math ability. After the math test, participants completed a
measure of self-integrity, a measure of personal mastery, and an awareness probe.
Measures
Demographic and personal characteristics. Participant age, gender, ethnicity,
education, employment status, and college generation status were assessed through the
prescreening survey.
Math identification. To assess the degree to which participants were invested in their
math performance, we used the following two items: (1) “How skilled are you at math?”, and
(2) “How important is doing well on standardized math tests to you?” (adapted from Sherman et
al., 2009). Participants rated the first item on a 7-point scale of 1 (Not very skilled) to 7 (Very
skilled), and the second item on a 7-point scale of 1 (Not important at all) to 7 (Very important).
The mean of these two items served as our measure of math identification, with higher scores
indicating that participants were more math-identified (α=.75; see Table 6).
Motives for attending college. To assess cultural values, we used the Motives for
Attending College Questionnaire (Stephen et al., 2012; see Appendix H). This 12-item measure
is comprised of two subscales: a) a six-item scale assessing independent motives, and b) a six-
AFFIRMATION AND MAJORITY STUDENTS 34
item scale assessing interdependent motives. Participants were instructed “There are many
reasons why people attend college. Please read the following list of possible motivating factors
below. Then, indicate the number that best represents the importance of each motive to you.”
They were then provided with the following prompt before each item: “I was motivated to attend
college because I want to….” Independent items included statements such as “become an
independent thinker” and “expand my knowledge of the world.” Interdependent items included
statements such as “be a role model for people in my community” and “help my family after I’m
done with college.” The mean of the six independent items served as our measure of
independent values (α = .78), and the mean of the six interdependent items served as our
measure of interdependent values (α = .88).
Self-integrity. We again used the six-item SIS to assess perceptions of self-integrity
(Sherman et al., 2009). Scale reliability with the present sample was good (α = .85).
Personal mastery. To assess sense of control, we used six items from the Personal
Mastery and Constraint Scale (adapted from Lachman & Weaver, 1998; see Appendix I). Items
included statements such as “I can do just about anything that I set my mind to” and “When I
really want to do something, I usually find a way to succeed at it.” Participants rated items on a
scale of 1 (Strongly disagree) to 7 (Strongly agree), with higher scores indicating a greater sense
of personal mastery. The mean of these six items served as our measure of personal mastery (α =
.81)
Awareness probe. We again used an awareness probe but adapted it for our online
sample. Participants were asked to estimate the degree to which 10 specific factors (i.e., math
ability, effort on the test, personal background, beliefs, attitudes, familiarity with math problems,
first writing exercise, reasoning ability, self-esteem, mood, their gender, and the amount paid to
AFFIRMATION AND MAJORITY STUDENTS 35
complete the task) contributed to their math performance. Participants rated each factor on a
scale of 1 (Strongly disagree) to 7 (Strongly agree) indicating how much they believed it
contributed to their math test performance. Lastly, participants were asked in an open-ended
fashion what they believed the study was about.
Results
Participant Values and Manipulation Check
Similar to Study 1, “relationships with friends and family” (52%) and “spiritual or
religious values” (10.7%) were the most frequent top ranked values. Next, were “learning and
gaining knowledge” (8.6%) and “independence” (8.1%). The remaining values were selected
infrequently (<5%) as participants’ most cherished value (see Table 7).
Using the same method as Study 1, we coded participants’ essays to determine whether
the study conditions produced different frequencies of social belonging themes. Affirmation and
control essays were coded by the principal investigator. Essays in the affirmation condition
included belonging themes 90% of the time, whereas essays in the control condition included
belonging themes only 3.5% of the time. The percentage of belonging-affirmations that included
themes of social belonging was lower than in Study 1 (90% vs. 100%). A closer examination of
essays revealed that a small number of participants in the belonging-affirmation condition did
not believe their top ranked value (e.g., financial success) helped them to feel more connected to
others. Overall, participants in the affirmation condition were still much more likely to include a
social belonging theme in their affirmation essays than those in the control condition, χ
2
(2, N =
197) = 149.20, p < .001.
AFFIRMATION AND MAJORITY STUDENTS 36
Awareness Probe Results
Consistent with Study 1, participants rated their “reasoning ability,” “effort,” “math
ability,” and “familiarity with math problems” as the four factors most relevant to their math
performance (see Table 8). For both men and women, the values affirmation writing exercise
was ranked 9
th
out of 10 listed factors. In the open-ended response section where participants
were queried regarding what they believed the study was about, only a small percentage (6.6%)
reported that they suspected the values affirmation was intended to affect their performance on
the math test. Removing these participants from subsequent analyses did not change our results.
Primary Analyses
To adapt our procedures to the online sample and decrease the likelihood of a floor
effect, we used a less challenging math test than the one used in Study 1 (see Appendix J). Study
2 participants answered nearly half the test problems correctly (M = 7.86, SD = 3.29; 15
problems total) suggesting that the math test was not too onerous for participants, and it did not
result in a floor effect.
We used two-way ANOVA to test our primary hypotheses that (a) women’s math
performance would be enhanced in the affirmation condition compared to control, and (b) men’s
math performance would be diminished in the affirmation condition compared to control.
Results did not reveal a significant main effect for affirmation condition, F (1, 196) = 0.22, p =
.64, on number of math problems correctly answered. There was a significant main effect of
gender, F (1, 196) = 17.17, p < .001, ηp
2
= 0.08, such that men (M = 8.79, SE = 0.31), on
average, answered more problems correctly than women (M = 6.86, SE = 0.31; see Table 9).
Contrary to predictions, there was no significant condition by gender interaction, F (1, 197) =
AFFIRMATION AND MAJORITY STUDENTS 37
0.86, p = .36. The number of problems correctly answered by affirmed females (M = 6.58, SE =
0.44) was similar to that for unaffirmed females (M = 7.22, SE = 0.50). Affirmed males’ (M =
8.89, SE = 0.47) math performance was nearly identical to that of unaffirmed males (M = 8.68,
SE = 0.42). In sum, math test performance was not affected by the affirmation manipulation
directly (Figure 2).
Next, using two-way ANOVA tests, we evaluated our secondary hypotheses that (a)
women’s ratings of self-integrity and personal mastery would be enhanced in the affirmation
condition compared to control, and (b) men’s ratings of self-integrity and personal mastery
would be poorer in the affirmation condition compared to control. With respect to ratings of self-
integrity, there were no main effects of condition, F (1, 196) = 0.24, p = .63, or of gender, F (1,
121) = 0.97, p = .33. The condition by gender interaction was also not significant, F (1, 196) =
0.33, p = .57. Perceptions of self-integrity were similar for females in the affirmation condition
(M = 45.09, SE = 0.84) and those in the control condition (M = 45.02, SE = 0.95). Among male
participants, there were no significant differences on ratings of self-integrity between those who
were affirmed (M = 45.46, SE = 0.90) and those who were unaffirmed (M = 46.39, SE = 0.81;
see Table 9). The affirmation manipulation did not affect participants’ sense of self-adequacy.
When examining participants’ ratings of personal mastery, we also failed to observe any
main effects of condition, F (1, 196) = 0.77, p = .38, or of gender, F (1, 196) = .96, p = .33. The
condition by gender interaction was also not significant, F (1, 196) = 0.34, p = .56. Affirmed
females’ ratings of personal mastery (M = 30.83, SE = 0.79) were similar to those of unaffirmed
females (M = 29.63, SE = 0.89), and affirmed males’ ratings of personal mastery (M = 31.15, SE
= 0.83) were similar to those of unaffirmed males (M = 30.91, SE = 0.75; see Table 9). Thus, the
affirmation manipulation did not influence participants’ feelings of personal mastery.
AFFIRMATION AND MAJORITY STUDENTS 38
Supplementary Analyses
We conducted moderator analyses to evaluate whether math identification interacted with
the affirmation condition to affect math performance. Participants’ level of math identification
was based on the mean scores of our two-item measure of math identity, and ranged from 1 to 7
(M = 4.38, SD = 1.48). As expected, math identity was positively correlated with math test
performance (r = .35, p < .001), with participants who were more math identified answering
more math problems correctly. When comparing math identification of male and female
participants, males reported a higher level of math identification (M = 4.71, SD = 1.42),
compared with females (M = 4.02, SD = 1.47; t(197) = 3.66, p < .001).
To evaluate whether math identification moderated the effects of affirmation, we
regressed math test performance (problems correct) on the mean-centered math identification
measure, affirmation condition (contrast coded control condition = -1, belonging-affirmation =
1), and the interaction. There was a significant effect of math identity, ß(193) = 0.34, p < .001,
but no significant effect of condition, ß(193) < 0.01, p = .99, or the interaction effect of condition
by math identity, ß(193) = 0.04, p = .57, on math test performance. Since we hypothesized that
the affirmation manipulation would have a different effect on male and female participants, we
also examined whether there was a three-way interaction between affirmation condition, gender,
and math identity on math performance. We regressed math test performance on math identity,
affirmation condition, gender, and all possible two-way and three-way interactions. There was
no significant three-way interaction, ß(189) < 0.01, p = .97, and the only significant effects were
those previously reported: the significant main effect of gender, ß(189) = 0.20, p < .01, and of
math identity, ß(189) = 0.29, p < .001. In sum, participants’ math identification did not interact
with the affirmation manipulation and gender to affect math performance.
AFFIRMATION AND MAJORITY STUDENTS 39
Next, we examined whether independent and interdependent values moderated the effects
of the belonging-affirmation on math performance. We hypothesized that participants who
reported more interdependent motives would benefit more from the belonging-affirmation,
whereas participants who reported more independent motives would be impaired. There were no
significant differences between males’ ratings of independent motives (M = 31.50, SD = 6.55)
and females’ ratings of independent motives (M = 32.57, SD = 6.00). There were also no
significant differences between males’ ratings of interdependent motives (M = 24.00, SD = 7.69)
and females’ ratings (M = 24.56, SD = 7.56). Interdependent motives were negatively correlated
with math performance (r = -.16, p < .05), such that those who reported more interdependent
motives tended to score poorer on the math test. There was no significant association between
participants’ ratings of independent motives and math performance.
To assess whether independent motives moderated the effects of affirmation on math test
performance, we regressed math test performance (problems correct) on the mean-centered
independent motives subscale, affirmation condition (contrast coded control condition = -1,
belonging-affirmation = 1), and the interaction. There was no significant effect of independent
motives, ß(193) = 0.07, p = .19, affirmation condition, ß(193) = 0.31, p = .43, or the interaction
of independent motives by condition, ß(193) = 0.39, p = .33. Participants’ independent motives
did not moderate affirmation effects on math performance.
To evaluate interdependent motives as a moderator of affirmation effects, we regressed
math test performance (problems correct) on the mean-centered interdependent motives
subscale, affirmation condition (contrast coded control condition = -1, belonging-affirmation =
1), and the interaction. Regression analyses revealed a significant effect for interdependent
motives ß(193) = -0.18, p = .01, but no significant effect of affirmation condition, ß(193) = -
AFFIRMATION AND MAJORITY STUDENTS 40
0.08, p = .26, or the interaction effect of interdependent motives and affirmation condition,
ß(193) = 0.11, p = .13. Overall, participants’ value-orientation did not moderate the effects of
affirmation on math performance.
Discussion
Study 2 did not find evidence to support our hypotheses. There was a significant gender
gap on the math test, with males outperforming females, but no effect of the affirmation
manipulation on math performance or psychological outcomes (i.e., self-integrity, personal
mastery). There was also no evidence to suggest that math identification or cultural values
moderated affirmation effects. It is worth noting, however, that participants’ interdependent
values were negatively associated with math performance, with those who reported more
interdependent motives performing worse. This finding is consistent with previous literature
suggesting that those students with more interdependent norms (e.g., first-generation students)
are at risk of poorer academic outcomes (Stephens et al., 2012a). The gender gap in average test
scores between female and male participants suggests that female participants may have been
affected by stereotype threat, and the belonging-affirmation may have failed to attenuate the
threat. However, because males reported being more math identified than female participants,
and there was a significant association between math identity and math performance, differences
in performance could also be attributed to differences in math skill between male and female
participants in our sample.
Methodologically, the present study improved on Study 1 in several ways: a) we included
a larger sample of White male participants (103 vs. 40); b) we limited participation to European
Americans who had at least one parent with a 4-year degree (i.e., were not current or former
AFFIRMATION AND MAJORITY STUDENTS 41
first-generation students); c) we evaluated potential moderators of affirmation effects; and d) we
decreased the variability in study procedures (e.g., group size, experimenter gender) by having
them administered online in a standardized format. Despite these improvements, Study 2 results
were largely consistent with Study 1, and the affirmation manipulation did not appear to affect
outcomes. In the next section, we review potential reasons for the null results within the broader
context of the research literature.
General Discussion
The present research aimed to assess the impact of an affirmation focused on the theme
of social connectedness (i.e., belonging-affirmation) on math performance. We conducted two
experimental studies to evaluate whether a belonging-affirmation enhanced the math
performance of females, while diminishing the performance of White males. We also examined
whether the affirmation affected psychological functioning (i.e., self-integrity, social fit, and
personal mastery), and hypothesized that the affirmation would buffer women psychologically
but hinder men. Results failed to support study hypotheses, as we did not observe any effects of
affirmation on math performance or psychological outcomes for either gender. Next, we draw on
recent stereotype threat and affirmation research to interpret the implications of our findings.
Stereotype Threat and Women’s Math Performance
Stereotype threat has received considerable attention as a contributor to the gender
achievement gap in STEM fields, particularly in mathematics (Stoet & Geary, 2012).
Affirmations are viewed as a cutting-edge psychological intervention for reducing the gender
gap in male-dominated fields (Yeager & Walton, 2011), with randomized field trials supporting
their efficacy (Miyake et al., 2010; Walton, Logel, Peach, Spencer, & Zanna, 2014). Our results,
AFFIRMATION AND MAJORITY STUDENTS 42
however, failed to replicate prior research demonstrating that affirmations enhanced women’s
math performance (Martens et al., 2006; Shnabel et al., 2013). One possible explanation for why
we did not observe a performance boosting effect of affirmation among women was that their
performance was not significantly limited by stereotype threat. Only Study 2 found evidence of a
gender gap in math achievement, but this may have been explained by gender differences in
math skill. Study 1 used a more difficult math test and had participants complete procedures in a
setting that mirrored a real-world academic environment. Under these circumstances, one would
expect that gender differences in performance would be more likely to emerge (Spencer et al.,
1999). Since significant gender differences in math performance were not observed, we are left
with questions regarding whether female participants were affected by stereotype threat, and
whether it impacted their math performance.
These questions appear relevant to the research community at large, as researchers are
divided on whether gender differences in math performance exist, where gender differences
exist, and whether stereotype threat is a significant contributor to these presumed gender
differences (Finnigan & Corker, 2016; Flore & Wicherts, 2015; Nguyen & Ryan, 2008). In a
critical review and meta-analysis of the effects of stereotype threat on women’s math
performance, Stoet and Geary (2012) concluded that stereotype threat is at best a modest
contributor to gender differences. Their review found that only 30% of unconfounded
experimental studies replicated the original stereotype threat finding (Spencer et al., 1999),
demonstrating that women performed poorer on a math test when they were explicitly told it had
shown gender differences compared with when they were not told this. The researchers argue
that enthusiasm for stereotype threat as an explanation of gender differences in math
achievement has precluded a more critical examination of the literature, and assert that
AFFIRMATION AND MAJORITY STUDENTS 43
additional research is needed to better understand other factors that may contribute to any
observed gender differences. Values affirmations are believed to be effective in improving
academic performance because they mitigate stereotype threat (Sherman & Cohen, 2014).
However, if stereotype threat does not consistently impair women’s math performance, results of
affirmation interventions in this domain are likely to be mixed, with null findings being routine
rather than exceptions.
Short- vs. Long-Term Benefits of Affirmation
Another possible explanation for our null results is that we only evaluated the immediate
consequences of affirmation and did not assess whether it had long-term effects on performance.
It is worth noting that early lab-based affirmation research highlighted its immediate impact in
reducing bias, psychological discomfort, and stereotype threat (Martens et al., 2006; McQueen &
Klein, 2006; Sherman & Cohen, 2006). As the affirmation literature evolved, research began to
emphasize its long-term benefits in bolstering academic outcomes (Brady et al., 2016; Cohen et
al., 2009; Sherman et al., 2013). Cohen and Sherman (2014) assert that values affirmations exert
their influence through psychological processes that have recursive effects, interact with the
environment and other psychological processes, and compound over time. Thus, they contend
that the benefits of affirmation interventions, particularly within educational settings, are likely
to become more salient over time and observed long-term. Two recent randomized affirmation
trials with undergraduate students support this reasoning. Both found evidence that affirmation
lifted the academic achievement of vulnerable students over multiple years, but neither study
identified any immediate academic benefits of affirmation (Brady et al., 2016; Layous et al.,
2017). With increasing evidence suggesting that affirmation effects are likely to unfold over time
AFFIRMATION AND MAJORITY STUDENTS 44
through recursive processes, future research should consider a focus on the distal, rather than
proximal, effects of affirmation interventions.
Robustness of Affirmation Interventions
A third explanation for why we may have failed to identify any benefits of affirmation is
that effects may not be as robust as initially thought. The first set of randomized field trials
showed that affirmation reduced the racial achievement gap of middle-school youth by 30%
(Cohen et al., 2006; 2009; Sherman et al., 2013), and initial replication studies with college
students found that affirmation interventions reduced achievement gaps by 50% or more
(Harackiewicz et al., 2014; Miyake et al., 2010). Despite these promising findings, recent
research suggests affirmation effects are more fragile than indicated by these earlier studies.
Several recent field-experimental studies of values affirmation interventions failed to
replicate Cohen and colleagues (2006; 2009) earlier findings demonstrating that affirmation
resulted in significant academic gains for minority youth (Dee, 2015; Haselman, Rozek, Grigg,
& Borman, 2015; Protzko and Aronson, 2016). These replication studies included a larger
number of students and schools than the initial studies by Cohen et al. (2006; 2009). In one
particularly large-scale replication trial, Dee (2015) conducted a values affirmation intervention
with an ethnically mixed sample of 2,500 middle-school youth. This well-powered study found
little support that affirmation enhanced minority students’ academic outcomes, and even found
evidence suggesting it decreased academic achievement for some students (e.g., 8
th
-grade girls).
Such findings call into question the robustness of these interventions, and suggest that
affirmation effects are likely more nuanced and modest than indicated by earlier research.
AFFIRMATION AND MAJORITY STUDENTS 45
Affirmation and Majority Students
If the benefits of affirmation interventions to marginalized students are more elusive and
uncertain than initially suspected, what is the impact of these interventions on nonthreatened
majority students? The primary intent of the present research was to identify a performance
limiting effect of affirmation on White males, a group we reasoned was not at-risk. We reviewed
theories (e.g., cultural mismatch, stereotype lift) that shed light on why affirmations may
decrease performance for some students. However, because our studies did not produce any
significant results related to math performance or psychological outcomes, we are unable to
provide support for these theories.
Given that findings of performance hindering effects of affirmation have been less
consistent than the performance enhancing effects of affirmation, should they be ignored? We
argue that they should not. Although these results have been sporadic, they have not disappeared.
In a large-scale study of MOOC (massive open online course) participants, Kizilcec and
colleagues (2017) found that a values affirmation intervention resulted in higher rates of course
completion among participants in less-developed countries, but lower rates of course completion
among participants in more-developed countries. These recent findings provide further support
that affirmation can result in poorer academic outcomes for privileged students, possibly because
they feel they have less at stake, and are reminded by the values affirmation of other areas in
which they could invest their time (Miyake et al., 2010). Perhaps for these nonthreatened
students, affirmations may trigger an alternate cycle of “recursive processes” that can lead to
academic disengagement over time (Vohs et al., 2013). If affirmations untether students’
academic performance from their sense of self adequacy and belonging, affirmed privileged
students may feel free to reduce their focus on academics and instead redirect this attention
AFFIRMATION AND MAJORITY STUDENTS 46
toward other valued life domains (e.g., relationships, physical health, personal hobbies). Future
research will need to bear out these speculations, and should focus on identifying the contexts in
which affirmation is most likely to be beneficial, while being least likely to confer significant
adverse effects to nonthreatened students. This will likely require researchers to reconsider and
rethink the mechanisms of values affirmations, and to explicitly investigate the pathways and
underlying processes through which affirmation may depress academic outcomes. For now, we
contend that such research is warranted, as sufficient evidence has accumulated to indicate that
values affirmations can impair academic performance.
AFFIRMATION AND MAJORITY STUDENTS 47
References
Armenta, B. E. (2010). Stereotype boost and stereotype threat effects: The moderating role of
ethnic identification. Cultural Diversity and Ethnic Minority Psychology, 16(1), 94-98.
doi: 10.1037/a0017564
Behrend, T. S., Sharek, D. J., Meade, A. W., & Wiebe, E. N. (2011). The viability of
crowdsourcing for survey research. Behavior Research Methods, 43(3), 800. doi:
10.3758/s13428-011-0081-0
Beilock, S. L., Rydell, R. J., & McConnell, A. R. (2007). Stereotype threat and working
memory: mechanisms, alleviation, and spillover. Journal of Experimental Psychology:
General, 136(2), 256. doi: 10.1037/0096-3445.136.2.256
Berinsky, A. J., Huber, G. A., & Lenz, G. S. (2012). Evaluating online labor markets for
experimental research: Amazon. com's Mechanical Turk. Political Analysis, 20(3), 351-
368.
Bowen, N. K., Wegmann, K. M., & Webber, K. C. (2013). Enhancing a brief writing
intervention to combat stereotype threat among middle-school students. Journal of
Educational Psychology, 105(2), 427. doi: 10.1037/a0031177
Brady, S. T., Reeves, S. L., Garcia, J., Purdie-Vaughns, V., Cook, J. E., Taborsky-Barba, S., ...
& Cohen, G. L. (2016). The psychology of the affirmed learner: Spontaneous self
affirmation in the face of stress. Journal of Educational Psychology, 108(3), 353.
doi: 10.1037/edu0000091
AFFIRMATION AND MAJORITY STUDENTS 48
Chandler, J., & Shapiro, D. (2016). Conducting clinical research using crowdsourced
convenience samples. Clinical Psychology, 12(1), 53. doi: 10.1146/annurev-clinpsy-
021815-093623
Cohen, G. L., Aronson, J., & Steele, C. M. (2000). When beliefs yield to evidence: Reducing
biased evaluation by affirming the self. Personality and Social Psychology Bulletin,
26(9), 1151-1164. doi: 10.1177/01461672002611011
Cohen, G. L., Garcia, J., Apfel, N., & Master, A. (2006). Reducing the racial achievement gap:
A social-psychological intervention. Science, 313, 1307–1310.
doi:10.1126/science.1128317
Cohen, G. L., Garcia, J., Purdie-Vaughns, V., Apfel, N., & Brzustoski, P. (2009). Recursive
processes in self-affirmation: Intervening to close the minority achievement gap. Science,
324, 400–403. doi: 10.1126/science.1170769
Cohen, G. L., & Sherman, D. K. (2014). The psychology of change: Self-affirmation and social
psychological intervention. Annual Review of Psychology, 65, 333-371. doi:
10.1146/annurev-psych-010213-115137
Cook, J. E., Purdie-Vaughns, V., Garcia, J., & Cohen, G. L. (2011). Chronic threat and
contingent belonging: Protective benefits of values affirmation on identity development.
Journal of Personality and Social Psychology, 102, 479–496. doi:10.1037/a0026312
Covarrubias, R., Herrmann, S. D., & Fryberg, S. A. (2016). Affirming the interdependent self:
Implications for Latino student performance. Basic and Applied Social Psychology,
38(1), 47-57.
Critcher, C. R., & Dunning, D. (2015). Self-affirmations provide a broader perspective on self-
threat. Personality and Social Psychology Bulletin, 41(1), 3-18.
AFFIRMATION AND MAJORITY STUDENTS 49
doi: 10.1177/0146167214554956
Crocker, J., Niiya, Y., & Mischkowski, D. (2008). Why does writing about important values
reduce defensiveness? Self-affirmation and the role of positive other-directed
feelings. Psychological Science, 19(7), 740-747. doi: 10.1111/j.1467-9280.2008.02150.x
Dee, T. S. (2015). Social identity and achievement gaps: Evidence from an affirmation
intervention. Journal of Research on Educational Effectiveness, 8(2), 149-168.
doi: 10.1080/19345747.2014.906009
Finnigan, K. M., & Corker, K. S. (2016). Do performance avoidance goals moderate the effect of
different types of stereotype threat on women’s math performance? Journal of Research
in Personality, 63, 36-43. doi: 10.1016/j.jrp.2016.05.009
Flore, P. C., & Wicherts, J. M. (2015). Does stereotype threat influence performance of girls in
stereotyped domains? A meta-analysis. Journal of School Psychology, 53(1), 25-44.
doi:10.1016/j.jsp.2014.10.002
Gibson, C. E., Losee, J., & Vitiello, C. (2014). A replication attempt of stereotype susceptibility
(Shih, Pittinsky, & Ambady, 1999): Identity salience and shifts in quantitative performance.
Social Psychology, 45(3), 194-198. doi: 10.1027/1864-9335/a000184
Goyer, J. P., Garcia, J., Purdie-Vaughns, V., Binning, K. R., Cook, J. E., Reeves, S. L., ... &
Cohen, G. L. (2017). Self-affirmation facilitates minority middle schoolers' progress
along college trajectories. Proceedings of the National Academy of Sciences, 114(29),
7594-7599. doi: 10.1073/pnas.1617923114
Hanselman, P., Rozek, C. S., Grigg, J., & Borman, G. D. (2017). New evidence on self-
affirmation effects and theorized sources of heterogeneity from large-scale
AFFIRMATION AND MAJORITY STUDENTS 50
replications. Journal of Educational Psychology, 109(3), 405-424. doi:
10.1037/edu0000141
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(2), 375-389. doi:
10.1037/a0034679
Kizilcec, R. F., Saltarelli, A. J., Reich, J., & Cohen, G. L. (2017). Closing global achievement
gaps in MOOCs. Science, 355(6322), 251-252. doi: 10.1126/science.aag2063
Lachman, M. E., & Weaver, S. L. (1998). The sense of control as a moderator of social class
differences in health and well-being. Journal of Personality and Social
Psychology, 74(3), 763. doi: 10.1037/0022-3514.74.3.763
Layous, K., Davis, E. M., Garcia, J., Purdie-Vaughns, V., Cook, J. E., & Cohen, G. L. (2017).
Feeling left out, but affirmed: Protecting against the negative effects of low belonging in
college. Journal of Experimental Social Psychology, 69, 227-231.
http://dx.doi.org/10.1016/j.jesp.2016.09.008
Martens, A., Johns, M., Greenberg, J., & Schimel, J. (2006). Combating stereotype threat: The
effect of self-affirmation on women’s intellectual performance. Journal of Experimental
Social Psychology, 42(2), 236-243. doi:10.1016/j.jesp.2005.04.010
McQueen, A., & Klein, W. M. (2006). Experimental manipulations of self-affirmation: A
systematic review. Self and Identity, 5(4), 289-354. doi: 10.1080/15298860600805325
Miyake, A., Kost-Smith, L. E., Finkelstein, N. D., Pollock, S. J., Cohen, G. L., & Ito, T. A.
(2010). Reducing the gender achievement gap in college science: A classroom study of
values affirmation. Science, 330, 1234–1237. doi:10.1126/science.1195996
AFFIRMATION AND MAJORITY STUDENTS 51
Nguyen, H. D., & Ryan, A. M. (2008). Does stereotype threat affect test performance of
minorities and women? A meta-analysis of experimental evidence. Journal of Applied
Psychology, 93(6), 1314-1334. doi: 10.1037/a0012702
Picho, K., Rodriguez, A., & Finnie, L. (2013). Exploring the moderating role of context on the
mathematics performance of females under stereotype threat: A meta-analysis. The
Journal of Social Psychology, 153(3), 299-333. doi: 10.1080/00224545.2012.737380.
Protzko, J., & Aronson, J. (2016). Context moderates affirmation effects on the ethnic
achievement gap. Social Psychological and Personality Science, 7(6), 500-507.
doi: 10.1177/1948550616646426
Schmader, T. (2002). Gender identification moderates stereotype threat effects on women's math
performance. Journal of Experimental Social Psychology, 38(2), 194-201.
doi: 10.1006/jesp.2001.1500
Schmader, T., Johns, M., & Forbes, C. (2008). An integrated process model of stereotype threat
effects on performance. Psychological Review, 115(2), 336. doi: 10.1037/0033-
295X.115.2.336.
Sherman, D. K., & Cohen, G. L. (2006). The psychology of self-defense: Self-affirmation
theory. Advances in Experimental Social Psychology, vol 38. (pp. 183-242) Elsevier
Academic Press, San Diego, CA. doi: 10.1016/S0065-2601(06)38004-5
Sherman, D. K., Hartson, K. A., Binning, K., Purdie-Vaughns, V., Garcia, J., Taborsky-Barba,
S., . . . Cohen, G. (2013). Deflecting the trajectory and changing the narrative: How self-
affirmation affects academic performance and motivation under identity threat. Journal
of Personality and Social Psychology, 104, 591–618. doi:10.1037/a0031495
AFFIRMATION AND MAJORITY STUDENTS 52
Sherman, D. K., Cohen, G. L., Nelson, L. D., Nussbaum, A. D., Bunyan, D. P., & Garcia, J.
(2009). Affirmed yet unaware: Exploring the role of awareness in the process of self-
affirmation. Journal of Personality and Social Psychology, 97(5), 745-764.
doi: 10.1037/a0015451
Shih, M., Pittinsky, T. L., & Ambady, N. (1999). Stereotype susceptibility: Identity salience and
shifts in quantitative performance. Psychological Science, 10(1), 80-83.
doi: 10.1111/1467-9280.00111
Shnabel, N., Purdie-Vaughns, V., Cook, J. E., Garcia, J., & Cohen, G. L. (2013). Demystifying
values-affirmation interventions: Writing about social belonging is a key to buffering
against identity threat. Personality and Social Psychology Bulletin, 39, 663– 676.
doi:10.1177/ 0146167213480816
Silverman, A., Logel, C., & Cohen, G. L. (2013). Self-affirmation as a deliberate coping
strategy: The moderating role of choice. Journal of Experimental Social Psychology,
49(1), 93-98. doi: http://dx.doi.org/10.1016/j.jesp.2012.08.005
Spencer, S. J., Steele, C. M., & Quinn, D. M. (1999). Stereotype threat and women's math
performance. Journal of Experimental Social Psychology, 35(1), 4-28. Retrieved from
http://search.proquest.com.libproxy2.usc.edu/docview/619412944?accountid=14749
Steele, C. M. (1997). A threat in the air: How stereotypes shape intellectual identity and
performance. American psychologist, 52(6), 613. doi: 10.1037/0003-066X.52.6.613
Steele, C. M., & Aronson, J. (1995). Stereotype threat and the intellectual test performance of
African Americans. Journal of personality and social psychology, 69(5), 797.
doi: 10.1037/0022-3514.69.5.797
AFFIRMATION AND MAJORITY STUDENTS 53
Stephens, N. M., Fryberg, S. A., Markus, H. R., Johnson, C. S., & Covarrubias, R. (2012a).
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. doi:10.1037/a0027143
Stephens, N. M., Hamedani, M. G., & Destin, M. (2014a). Closing the social-class achievement
gap: A difference-education intervention improves first-generation students’ academic
performance and all students’ college transition. Psychological Science, 25(4), 943-953.
doi: 10.1177/0956797613518349
Stephens, N. M., Markus, H. R., & Phillips, L. T. (2014b). Social class culture cycles: How three
gateway contexts shape selves and fuel inequality. Annual Review of Psychology, 65,
611-634. doi: 10.1146/annurev-psych-010213-115143
Stephens, N. M., Townsend, S. M., Markus, H., & Phillips, L. (2012b). A cultural mismatch:
Independent cultural norms produce greater increases in cortisol and more negative
emotions among first-generation college students. Journal of Experimental Social
Psychology, 48, 1389–1393. doi:10.1016/j.jesp.2012.07.008
Stephens, N. M., Townsend, S. S., Hamedani, M. G., Destin, M., & Manzo, V. (2015). A
difference-education intervention equips first-generation college students to thrive in the
face of stressful college situations. Psychological Science, 26(10), 1556-1566.
doi: 10.1177/0956797615593501
Stoet, G., & Geary, D. (2012). Can stereotype threat explain the gender gap in mathematics
performance and achievement? Review of General Psychology., 16(1), 93-102.
doi: 10.1037/a0026617
AFFIRMATION AND MAJORITY STUDENTS 54
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(5), 635. doi: 10.1037/pspa0000049
Townsend, C., & Sood, S. (2012). Self-affirmation through the choice of highly aesthetic
products. Journal of Consumer Research, 39(2), 415-428. doi: 10.1086/663775
Vohs, K. D., Park, J. K., & Schmeichel, B. J. (2013). Self-affirmation can enable goal
disengagement. Journal of Personality and Social Psychology, 104(1), 14.
doi: 10.1037/a0030478
Walton, G. M., & Cohen, G. L. (2003). Stereotype lift. Journal of Experimental Social
Psychology, 39(5), 456-467. doi: 10.1016/S0022-1031(03)00019-2
Walton, G. M., & Cohen, G. L. (2007). A question of belonging: Race, social fit, and
achievement. Journal of Personality and Social Psychology, 92, 82–96.
doi:10.1037/0022-3514.92.1.82
Walton, G. M., Logel, C., Peach, J. M., Spencer, S. J., & Zanna, M. P. (2015). Two brief
interventions to mitigate a “chilly climate” transform women’s experience, relationships,
and achievement in engineering. Journal of Educational Psychology, 107(2), 468.
doi:10.1037/a0037461
Wilcox, R. R. (2012). Introduction to robust estimation and hypothesis testing. Amsterdam:
Elsevier.
Woolf, K., Potts, H. W., & McManus, I. C. (2011). Ethnicity and academic performance in UK
trained doctors and medical students: systematic review and meta-analysis. BMJ, 342,
d901. doi:10.1136/bmj.d901
AFFIRMATION AND MAJORITY STUDENTS 55
Woolf, K., McManus, I. C., Gill, D., & Dacre, J. (2009). The effect of a brief social intervention
on the examination results of UK medical students: a cluster randomised controlled trial.
BMC Medical Education, 9(1), 35. doi: 10.1186/1472-6920-9-35
Yeager, D. S., & Walton, G. M. (2011). Social-psychological interventions in education: They’re
not magic. Review of educational Research, 81(2), 267-301.
doi: 10.3102/0034654311405999
AFFIRMATION AND MAJORITY STUDENTS 56
Table 1.
Study 1 Participant Demographic Characteristics
Male
(n = 40)
Female
(n = 82)
M SD M SD
Age, years 20.25 1.55 19.75 1.35
n % n %
Ethnicity - - - -
Caucasian 40 100 57 69.5
Black 0 0 9 11.0
Latino 0 0 8 9.8
Biracial 0 0 8 9.8
Transfer Student - - - -
Yes 14 35.0 18 22.0
No 26 65.0 64 78.0
Generation status - - - -
First-generation 7 17.5 15 18.3
Continuing-generation 33 82.5 67 81.7
Grade - - - -
Freshman 10 25.0 22 26.8
Sophomore 8 20.0 32 39.0
Junior 11 27.5 11 13.4
Senior 11 27.5 17 20.7
AFFIRMATION AND MAJORITY STUDENTS 57
Table 2.
Study 1 Frequency of Values Ranked Most Important
Male
(n = 40)
Female
(n = 82)
Value n % n %
Relationships with friends and family 30 75.0 55 67.1
Independence 5 12.5 10 12.2
Spiritual or religious values 2 5.0 14 17.1
Creativity 1 2.5 2 2.4
Physical Attractiveness 1 2.5 0 0
Music 1 2.5 0 0
Sense of humor 0 0 1 1.2
Sports Ability 0 0 0 0
Politics or government 0 0 0 0
Being good at art 0 0 0 0
Membership in a social group 0 0 0 0
AFFIRMATION AND MAJORITY STUDENTS 58
Table 3.
Study 1 Awareness Probe Results
Factor
Male
(n = 40)
M (SD)
Female
(n = 82)
M (SD)
Familiarity with math
problems
6.28 (2.15) 5.85 (2.35)
Math ability 5.85 (2.39) 5.63 (2.54)
Reasoning ability 5.40 (2.21) 5.24 (2.06)
Effort* 4.88 (2.24) 5.82 (1.79)
Mood 4.60 (2.42) 5.05 (2.26)
Self-esteem* 3.85 (2.06) 3.94 (2.09)
Time of day 3.68 (2.20) 3.40 (2.02)
Personal background, beliefs
and attitudes
3.15 (2.13) 3.12 (2.22)
Weather 2.78 (2.14) 2.38 (1.73)
Experimenter gender 2.61 (2.00) 1.38 (0.93)
First writing exercise (values
affirmation)
2.28 (1.78) 2.43 (1.86)
*p<.05
AFFIRMATION AND MAJORITY STUDENTS 59
Table 4.
Study 1Math Performance and Psychological Outcomes
Male Female
Measure
Affirmation
(n = 21)
M (SD)
Control
(n = 19)
M (SD)
Affirmation
(n = 44)
M (SD)
Control
(n = 38)
M (SD)
Math performance
(problems correct)
4.43 (2.54) 4.42 (2.36) 3.59 (2.55) 3.61 (2.97)
Self-Integrity 46.67 (5.13) 46.53 (5.82) 45.39 (5.39) 46.84 (4.33)
Social Fit 31.52 (5.77) 32.89 (6.45) 33.20 (4.90) 34.29 (5.44)
AFFIRMATION AND MAJORITY STUDENTS 60
Table 5.
Study 2 Participant Demographic Characteristics
Male
(n = 103)
Female
(n = 94) Characteristic
M SD M SD
Age, years 29.44 4.52 27.48 3.91
n % n %
Education - - - -
High school graduate 3 2.9 5 5.3
Some college but no degree 24 23.3 19 20.2
Associate degree 11 10.7 13 13.8
Bachelor’s degree 46 44.7 35 37.2
Master’s degree 12 11.7 19 20.2
Doctoral degree 4 3.9 2 2.1
Professional degree (JD, MD) 3 2.9 1 1.1
Employment status - - - -
Employed full-time 72 69.9 49 52.1
Employed part-time 8 7.8 21 22.3
Unemployed looking for work 5 4.9 3 3.2
Unemployed not looking for work 1 1.0 9 9.6
Undergraduate student 13 12.6 10 10.6
Graduate student 4 3.9 0 0
Retired 0 0 0 0
Disabled 0 0 2 2.1
Current Student 17 16.3 10 10.8
AFFIRMATION AND MAJORITY STUDENTS 61
Table 6.
Study 2 Participant Math Identity and Values
Measure
Male
(n = 103)
M (SD)
Female
(n = 94)
M (SD)
Math identity* 4.71 (1.42) 4.02 (1.47)
Independent motives 31.50 (6.55) 32.57 (6.00)
Interdependent motives 24.00 (7.69) 24.56 (7.56)
*p<.05
AFFIRMATION AND MAJORITY STUDENTS 62
Table 7.
Study 2 Frequency of Values Ranked Most Important
Male
(n = 103)
Female
(n = 94)
Value n % n %
Relationships with friends and family 46 44.6 56 59.6
Spiritual or religious values 10 9.7 11 11.7
Learning and gaining knowledge 10 9.7 7 7.4
Career 9 8.7 1 1.1
Independence 7 6.8 9 9.6
Financial success 7 6.8 2 2.1
The environment 6 5.8 3 3.2
Creativity 4 3.9 3 3.2
Music 2 1.9 1 1.1
Politics or government 2 1.9 0 0
Sense of humor 0 0 1 1.1
Athletic ability 0 0 0 0
AFFIRMATION AND MAJORITY STUDENTS 63
Table 8.
Study 2 Awareness Probe Results
Factor
Male
(n = 103)
M (SD)
Female
(n = 94)
M (SD)
Reasoning ability 5.97 (0.95) 5.79 (0.87)
Effort 5.87 (1.01) 5.77 (0.98)
Math ability 5.59 (1.39) 5.57 (1.44)
Familiarity with math
problems
5.58 (1.35) 5.49 (1.33)
Mood 4.69 (1.56) 4.27 (1.54)
Self-esteem* 4.30 (1.81) 3.43 (1.61)
Amount paid* 3.92 (2.06) 2.99 (1.82)
Personal background, beliefs
and attitudes
3.83 (1.90) 3.37 (1.90)
First writing exercise (values
affirmation)*
3.48 (2.07) 2.85 (1.77)
Gender* 2.60 (2.01) 2.01 (1.47)
*p < .05
AFFIRMATION AND MAJORITY STUDENTS 64
Table 9.
Study 2 Math Performance and Psychological Outcomes
Male
(n = 103)
Female
(n = 94)
Measure
Affirmation
(n = 46)
M (SD)
Control
(n =57)
M (SD)
Affirmation
(n = 53)
M (SD)
Control
(n = 41)
M (SD)
Math test performance
(problems correct)
8.89 (3.41) 8.68 (3.26) 6.58 (3.02) 7.22 (2.92)
Self-Integrity 45.46 (6.22) 46.39 (5.62) 45.09 (6.46) 45.02 (6.15)
Personal Mastery 31.15 (4.88) 30.91 (6.24) 30.83 (5.87) 29.63 (5.46)
AFFIRMATION AND MAJORITY STUDENTS 65
Figure 1. Study 1 participant mean number of math problems correct by condition and gender.
There were no significant differences in number of problems correctly answered by affirmation
condition or gender. Standard errors are represented in the figure by the error bars attached to
each column.
0
1
2
3
4
5
6
Men Women
Study 1 Math Test Problems Correct
Control Belonging-Affirmation
AFFIRMATION AND MAJORITY STUDENTS 66
Figure 2. Study 2 participant mean number of math problems correct by condition and gender.
Males answered more problems correctly than females. There were no significant differences in
number of problems correctly answered by affirmation condition for male or female participants.
Standard errors are represented in the figure by the error bars attached to each column.
0
1
2
3
4
5
6
7
8
9
10
Men Women
Study 2 Math Test Problems Correct
Control Belonging-Affirmation
AFFIRMATION AND MAJORITY STUDENTS 67
Appendix A: Belonging-Affirmation Writing Exercise
(adapted from Shnabel et al., 2013)
WHAT ARE YOUR PERSONAL VALUES?
Below is a list of characteristics and values, some of them may be important to you; some may
be unimportant to you. Please rank them from 1 to 11 according to how important they are to
you (“1” being the most important item, “11” being the one that is least important to you). Use
each number only once.
Qqqqqq Being Good at Art
Qqqqqq Physical Attractiveness
Qqqqqq Creativity
Qqqqqq Independence
Qqqqqq Membership in a Social Group (such as your community, racial group, or school club)
Qqqqqq Music
Qqqqqq Politics
Qqqqqq Relationships with Friends or Family
Qqqqqq Religious Values
Qqqqqq Sense of Humor
Qqqqqq Sports Ability
AFFIRMATION AND MAJORITY STUDENTS 68
Directions:
1. Look at the value you picked as most important to you (the value you ranked #1 on the
previous page).
2. Think about times when this value made you feel closer and more connected with people.
3. Describe why this value is important in making you feel closer and more connected with
people.
Focus on your thoughts and feelings and don’t worry about spelling, grammar, or how
well written it is.
Again, look at the value you picked as most important. List the top two reasons why this value
makes you feel closer and more connected with people.
1.
2.
Circle how much you agree or disagree with each of these statements.
1. This value has influenced my relationships with people.
Strongly
Disagree
Disagree Somewhat
Disagree
Somewhat
Agree
Agree Strongly
Agree
2. Other people in my life are an important part of what makes this value special.
Strongly
Disagree
Disagree Somewhat
Disagree
Somewhat
Agree
Agree Strongly
Agree
AFFIRMATION AND MAJORITY STUDENTS 69
Appendix B: Neutral Writing Control Exercise
(adapted from Shnabel et al., 2013)
WHAT ARE YOUR PERSONAL VALUES?
Below is a list of characteristics and values, some of them may be important to you; some may
be unimportant to you. Please rank them from 1 to 11 according to how important they are to
you (“1” being the most important item, “11” being the one that is least important to you). Use
each number only once.
Qqqqqq Being Good at Art
Qqqqqq Physical Attractiveness
Qqqqqq Creativity
Qqqqqq Independence
Qqqqqq Membership in a Social Group (such as your community, racial group, or school club)
Qqqqqq Music
Qqqqqq Politics
Qqqqqq Relationships with Friends or Family
Qqqqqq Religious Values
Qqqqqq Sense of Humor
Qqqqqq Sports Ability
AFFIRMATION AND MAJORITY STUDENTS 70
Directions:
1. Look at the value you ranked as #9 on the previous page.
2. Think about times when this value would be important to someone else (like another
student at your school or a person you’ve heard about).
3. Describe why this value would be important to someone else.
Focus on your thoughts and feelings and don’t worry about spelling, grammar, or how
well written it is.
Again, look at your #9 value. List the top two reasons why someone else would pick this as their
most important value.
1.
2.
Circle how much you agree or disagree with each of these statements.
1. This value has influenced some people.
Strongly
Disagree
Disagree Somewhat
Disagree
Somewhat
Agree
Agree Strongly
Agree
2. This value is important to some people.
Strongly
Disagree
Disagree Somewhat
Disagree
Somewhat
Agree
Agree Strongly
Agree
AFFIRMATION AND MAJORITY STUDENTS 71
Appendix C: Study 1 Math Test
University of Southern California Math Assessment Test
You will be asked to complete a very difficult math test.
You will have 20 minutes to complete 30 multiple choice questions. Calculators are not allowed
for this test. Please write your ID number and name on the Scantron form. Fill in your answers in
with a pencil or pen. If you guess on any question, please mark “G” next to the appropriate
test item on the Scantron form.
Please do not write on this test. Scratch paper is available on the table.
You may find that this test is very challenging and difficult. This is the case in order to
provide the best measure of current levels of math ability. In order to get the best
assessment of current ability, it is very important that you try your best on this test. At the
end of the study, you will have the opportunity to get feedback on your answers.
AFFIRMATION AND MAJORITY STUDENTS 72
AFFIRMATION AND MAJORITY STUDENTS 73
AFFIRMATION AND MAJORITY STUDENTS 74
AFFIRMATION AND MAJORITY STUDENTS 75
AFFIRMATION AND MAJORITY STUDENTS 76
AFFIRMATION AND MAJORITY STUDENTS 77
AFFIRMATION AND MAJORITY STUDENTS 78
AFFIRMATION AND MAJORITY STUDENTS 79
Appendix D: Self-Integrity Scale
(adapted from Sherman et al., 2009)
Please indicate your agreement with the statements below by writing the appropriate number
next to the statement using the following scale.
Strongly Disagree Disagree Agree Strongly Agree
1 2 3 4 5 6 7
______ 1. I have the ability and skills to deal with whatever comes my way.
______ 2. I feel that I’m basically a moral person.
______ 3. On the whole, I am a capable person.
______ 4. I am a good person.
______ 5. When I think about the future, I’m confident that I can meet the challenges that I will
face.
______ 6. I try to do the right thing.
______ 7. Even though there is always room for self-improvement, I feel a sense of
completeness about who I fundamentally am.
______ 8. I am comfortable with who I am.
AFFIRMATION AND MAJORITY STUDENTS 80
Appendix E: Social Fit Questionnaire
(adapted from Walton and Cohen, 2007)
Please indicate your agreement with the statements below by writing the appropriate number
next to the statement that best reflects how you feel
Strongly disagree Disagree Agree Strongly Agree
1 2 3 4 5 6 7
______ 1. I feel like I belong as a student at USC.
______ 2. I feel like I fit in with the academic community at USC.
______ 3. If my parents were to visit me or when they have visited me at USC, I would feel or
have felt comfortable introducing them to my friends.
______ 4. I feel a part of the USC community.
______ 5. I expect that the social experience at USC will be difficult for me.
______ 6. In the future, I could see myself having a lot of friends at USC.
AFFIRMATION AND MAJORITY STUDENTS 81
Appendix F: Awareness Probe
(adapted from Sherman et al., 2009)
In the space below, please list whatever factors you think may have contributed to your
performance on the math test you completed
AFFIRMATION AND MAJORITY STUDENTS 82
Please estimate the extent to which each of the following factors contributed to your
performance on the math test
1. Your math ability
1 2 3 4 5 6 7 8 9
Didn’t contribute at all Contributed a great deal
2. Your effort on the test
1 2 3 4 5 6 7 8 9
Didn’t contribute at all Contributed a great deal
3. Your personal background, beliefs, and attitudes
1 2 3 4 5 6 7 8 9
Didn’t contribute at all Contributed a great deal
4. The first writing exercise you completed (about values)
1 2 3 4 5 6 7 8 9
Didn’t contribute at all Contributed a great deal
5. Your familiarity with math problems
1 2 3 4 5 6 7 8 9
Didn’t contribute at all Contributed a great deal
AFFIRMATION AND MAJORITY STUDENTS 83
6. Your reasoning ability
1 2 3 4 5 6 7 8 9
Didn’t contribute at all Contributed a great deal
7. Your self-esteem
1 2 3 4 5 6 7 8 9
Didn’t contribute at all Contributed a great deal
8. Your mood
1 2 3 4 5 6 7 8 9
Didn’t contribute at all Contributed a great deal
9. The gender of the research assistant who gave you test
1 2 3 4 5 6 7 8 9
Didn’t contribute at all Contributed a great deal
10. The time of the day
1 2 3 4 5 6 7 8 9
Didn’t contribute at all Contributed a great deal
11. Today's weather
1 2 3 4 5 6 7 8 9
Didn’t contribute at all Contributed a great deal
AFFIRMATION AND MAJORITY STUDENTS 84
Appendix G: Belonging-Affirmation Coding Manual
(adapted from Shnabel et al., 2013)
If a participant explicitly mentions one of the following in their essay, then they included a
belonging theme and their essay should be coded as a “1.” If they did not, please code their essay
as a “0.”
a. Stating that one (i.e., the participant) values an activity because it is done with others.
b. Stating that one (i.e., the participant) feels part of a group of people because of a certain
value or while engaging in a certain activity.
An essay does not include a belonging theme if it only references how others might connect
because of a value, or fails to state how the value connects the participant personally with others.
If an essay only includes one of these themes, and does not state how the participant values the
activity because it is done with others or makes them feel part of a group, code the essay as a
“0.”
a. How “others” might value an activity because it is done with others
b. How “others” feel part of a group of people because of a certain value or while engaging
in a certain activity
AFFIRMATION AND MAJORITY STUDENTS 85
Appendix H: Motives for Attending College Questionnaire
(adapted from Stephens et al., 2014b)
There are many reasons why people attend college. Please read the following list of reasons why
people attend college. Then, select the choice that best represents the importance of each motive
to you.
I was motivated to attend college because I want to…
______ 1. become an independent thinker.
______ 2. explore new interests.
______ 3. expand my knowledge of the world.
______ 4. explore my potential in many domains.
______ 5. learn more about my interests.
______ 6. expand my understanding of the world
______ 7. give back to my community.
______ 8. provide a better life for my own children.
______ 9. show that people with my background can do well.
______ 10. bring honor to my family.
______ 11. be a role model for people in my community.
______ 12. help my family out after I'm done with college.
Note: Items 1-6 comprised the independent motives subscale and items 7-12 comprised the
interdependent motives subscale.
AFFIRMATION AND MAJORITY STUDENTS 86
Appendix I: Personal Mastery and Constraint Scale
(adapted from Lachman & Weaver, 1998)
Please indicate your agreement with each of the following statements by selecting the option on
the scale that best reflects how you feel.
Strongly disagree Disagree Agree Strongly Agree
1 2 3 4 5 6 7
______ 1. I can do just about anything I set my mind to.
______ 2. I sometimes feel I am being pushed around in life.
______ 3. When I really want to do something, I usually find a way to succeed at it.
______ 4. I often feel helpless in dealing with problems of my life.
______ 5. What happens to me in the future depends mostly on me.
______ 6. There is little I can do to change the important things in my life.
Note: Items 2, 4, and 6 are reverse coded.
AFFIRMATION AND MAJORITY STUDENTS 87
Appendix J: Study 2 Math Test
This task is comprised of 15 math problems. You will have 15 minutes to complete the problems.
After 15 minutes has passed, the screen will automatically advance to the next section. There is a
timer at the bottom of the page. You can use scratch paper and a pencil or pen to complete the
task but may not use a calculator.
You may find that this test is challenging and difficult. This is the case in order to provide the
best measure of your current level of math ability. In order to get the best assessment of your
math ability, it is very important that you try your best on this test.
(1) The average of three numbers is 55. The second is 1 more than twice the first, and the third is
4 less than three times the first. Find the largest number.
a. 165
b. 57
c. 88
d. 80
e. None of the above
(2) How much tea worth 93 cents per pound must be mixed with tea worth 75 cents per pound to
produce 10 pounds worth 85 cents per pound?
a. 2 2/9
b. 3 1/2
c. 4 4/9
d. 5 5/9
e. 12/17
AFFIRMATION AND MAJORITY STUDENTS 88
(3) Jennifer has enough money to buy 45 bricks. If the bricks each cost 10 cents less, Jennifer
could buy 5 more bricks and have no money left. How much money does Jennifer have to
spend on bricks?
a. $100
b. $50
c. $45
d. $40
e. $30
(4) During one season, a tennis team won 21 matches and lost 30% of their matches. What was
the number of matches that the team lost?
a. 70
b. 7
c. 5
d. 30
e. 9
(5) Macey is three times as old as Mike. In 8 years, she will be twice as old as Mike. How old
was Macey 3 years ago?
a. 21
b. 30
c. 8
d. 24
e. 5
AFFIRMATION AND MAJORITY STUDENTS 89
(6) A small college reduced its faculty by approximately 13 percent to 195 professors. What was
the original number of faculty members?
a. 220
b. 224
c. 230
d. 236
e. 242
(7) Three factories of Conglomerate Corporation are capable of manufacturing hubcaps. Two of
the factories can each produce 100,000 hubcaps in 15 days. The third factory can produce
hubcaps 30% faster. How many days would it take to produce a million hubcaps with all
three factories working simultaneously?
a. 38
b. 42
c. 46
d. 50
e. 54
(8) An incredible punch is composed of buttermilk, orange juice, and brandy. How many pints of
orange juice are required to make 60 pints of punch containing twice as much buttermilk as
orange juice and three times as much orange juice as brandy?
a. 16
b. 18
c. 20
d. 22
AFFIRMATION AND MAJORITY STUDENTS 90
e. 24
(9) A bookseller sells his books at a 20% markup in price. If he sells a book for $12.00, how
much did he pay for it?
a. $14.40
b. $14.00
c. $11.80
d. $10.00
e. $9.60
(10) Two quarts containing 2/3 water and 1/3 formula are mixed with three quarts containing 3/8
water and 5/8 formula. Approximately what percent of the combined five-quart mixture is
water?
a. 40%
b. 45%
c. 50%
d. 55%
e. 60%
(11) The average of four numbers is x. If the average of the first three numbers is e, what is the
value of the fourth number?
a. 3x-4e
b. 3e-4x
c. 4x-e
d. (x+e)/4
e. 4x-3e
AFFIRMATION AND MAJORITY STUDENTS 91
(12) A coat is offered at a 20% discount. What is the original price of the coat if the sale price is
$18.40?
a. $14.72
b. $22.08
c. $23.00
d. $24.60
e. $36.80
(13) The smallest of three consecutive even integers is 40 less than three times the largest. Find
the largest of these integers.
a. 14
b. 17
c. 18
d. 19
e. 20
(14) A salesperson receives a salary of $150 per week and earns a commission of 15% on sales
she makes. How many dollars’ worth of sales does she need to make in order to bring her
total weekly income to $600?
a. $3,000
b. $3,150
c. $4,000
d. $4,150
e. $5,000
AFFIRMATION AND MAJORITY STUDENTS 92
(15) A man invests $1500, part at 8% and the remainder at 5%. If his interest from both
investments for the year totaled $102.00, what is the amount he invested at 8%?
a. $500
b. $600
c. $800
d. $900
e. $1,200
Abstract (if available)
Abstract
Self-affirmation is a promising brief intervention capable of reducing the academic achievement gap between majority and stigmatized groups. Affirmations improve academic performance among stigmatized groups (underrepresented minorities, women in STEM fields) by expanding one’s sense of self, buffering social belonging, and reducing social identity threat (Cohen & Sherman, 2014). Despite these encouraging findings, some studies suggest that affirmations may decrease the academic performance of White students (Brady et al., 2016
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Towards validating therapists’ in-session behaviors of cultural competence
PDF
Cultural influences on mental health stigma in Asian and European American college students
PDF
Identity, perceived discrimination, and attenuated positive psychotic symptoms among college students
PDF
The role of social status and gender in the composition of bully-victim dyads in early adolescence
PDF
An academic and professional preparatory curriculum design and supplemental academic advisement tool: self-regulation, ethics, and communication for engineering graduate students
PDF
The effects of math anxiety and low self-efficacy on students’ attitudes and interest in STEM
PDF
The Bridge Program and underrepresented Latino students: an evaluation study
PDF
Trunk control during dynamic balance: effects of cognitive dual-task interference and a history of recurrent low back pain
Asset Metadata
Creator
Jones, Eduardo Odell
(author)
Core Title
Affirmation and majority students: Can affirmation impair math performance?
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Psychology
Publication Date
09/21/2017
Defense Date
08/14/2017
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
OAI-PMH Harvest,self-affirmation,stereotype threat,values affirmation
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Huey, Stanley Jr. (
committee chair
), Abdou, Cleopatra (
committee member
), Lòpez, Steven (
committee member
), Oyserman, Daphna (
committee member
), Townsend, Sarah (
committee member
)
Creator Email
eduardo.o.jones@kp.org,eduardoj@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c40-430328
Unique identifier
UC11264388
Identifier
etd-JonesEduar-5741.pdf (filename),usctheses-c40-430328 (legacy record id)
Legacy Identifier
etd-JonesEduar-5741-0.pdf
Dmrecord
430328
Document Type
Dissertation
Rights
Jones, Eduardo Odell
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
self-affirmation
stereotype threat
values affirmation