PSY 311: prejudice article 4: stereotype threat

area of prejudice least studied

= no correspondence attention to experience of being the target of prejudices and stereotypes =started to change in the sense that there having emerged greater interest in effects of/reactions to societal devaluation **stigmatization

study 1 methods

=28 men and women selected from intro psych pool at UMich =all required to have 1 semester of calculus and received B or better =required to have scored above 85th percentile on SAT or ACT =on 11 point scale of strongly agree-strongly disagree, P's had to strongly agree (1-3) that they are good at math and that it's important to them to be good at math---measures whether person is self-schematic in domain 2x2 design (male and female) (easier and difficult math test) dependent variables= performance on test and time spent working

study 2 methods

=30 women, 24 men selected from intro psych pool at UMich using same criteria as study 1 =2x2 mixed model design with one b/w participants factor (sex) and one w/in participants factor (test characterization) =primary dependent variables were performance on math test and time spent on test difficult test in study 1 was divided in 2 halves. P's had 15 mins to complete each half. 1st 6 questions from earlier test comprised first test & next 5 questions comprised second test half participants told that first test was one on which there were gender differences and second test had no gender differences other half told opposite, that first test was one for which no gender differences and second test was one of which there were gender differences *random assignment to these order conditions*

study 3 methods

=36 women and 31 men selected from intro psych pool at state U of NY @ buffalo =with adapting experiment to different Participant pop., used somewhat easier test and selected participants scoring between 400-650 on math of SAT and completed no more than 1 year of calculus ---changes maintain basic experimental situation of test being quite difficult for P's but still within upper ranges of their ability =2x2 factorial design (sex x test characterization) =primary dependent measure was Ps' performance on test and measures of evaluation apprehension, state anxiety, and self-efficacy =paper exam =only 1 test and 20 mins to complete =similar to test in 1 and 2, except that b/c of changes in selection procedures/ participant pop., test is somewhat easier --- based on graduate management test math portion =directions basically same as 1 and 2, except intro of questions designed to examine possible mediators collected after test characterization manipulation/prior to actual test =the no gender difference condition is the same as study 2 =in control, subjects were given no info on gender differences = after instructions, questionairre of 4 questions measuring evaluation apprehension, 5 questions measuring self-efficacy, and state-trait anxiety index =sample question involved to allow P's to know test was difficult, but to ensure it did not affect mediation measures, instructions read aloud to emphasize instructions and standardize exposure to sample question =single experimenter was female

stereotype threat

=begins with fact that most devaluing group stereotypes are widely known throughout society ie: sample of P's with varying views on prejudice toward African-Americans, Devine found all P's knew stereotypes about this group **this lies the threat --everyone, including those stereotyped, know allll the stereotypes about them --in situtations where stereotype applies, they face implication that anything they do or any feature they have that fits the stereotype makes it more plausible that they will be evaluated based on the stereotype

study 2

=causes of difference in test performance position= women experience stereotype threat (possibility of being stereotyped) when taking math tests and the threat is likely to undermine performance on difficult tests =HOWEVER, alternative interpretation of study 1: maybe women equaled men on easier math test not because stereotype had less effect, but because only advanced material sensitive to real ability differences between men/women study 2 -- tests effects of stereotype directly by giving all participants a difficult exam (similar to one in study 1)-- but varied whether gender stereotype was relevant to their performance =manipulated relevance of stereotype by varying how test was represented =in relevance condition, P's told that test had shown gender differences in past----characterization that explicitly evoked the stereotype about women's math ability =in condition where stereotype is irrelevant, P's told test had NEVER shown gender differences in past ---representing the test in a way to make stereotype irrelevant to interpreting women's performance on particular math test

study 2 goals

=direct test of theory that it is a stereotype-guided interpretation of performance difficulty that causes women's underperformance on difficult math test in these experiments =if women underperformed on difficult test in Study 1 because of stereotype threat(possibility that one's performance could be judged stereotypically)--then making stereotype irrelevant to interpreting their performance should eliminate this underperformance =if underperformance is due to an ability difference between men and women that is detectable only with difficult math items, women should underperform regardless of relevance of stereotype

stigmatization

=experience of bearing a "spoiled identity" -- some characteristic that, in eyes of society, causes one to be broadly devalued

present research on being target of prejudice

=extending focus by examining experience of being in a situation where one faces judgement based on societal stereotypes about one's group---experience=stereotype threat

study 3 goals

=if reducing gender relevance of the stereotype still improves women's performance under these conditions, then we can have greater confidence that stereotype is playing significant role in women's math performance =explore mediation of effect of stereotype threat on women's math performance (presumably, predicament caused by stereotype threat adds to normal self-evaluative risk of performance the further risk for women of confirming or being judged by negative stereotype about their math ability...condition differences in self-efficacy rather than differences in stereotype threat could have mediated effects of previous studies) --as preliminary test of these interpretations, present study measure participants evaluation apprehension, state anxiety, and self-efficacy after they received instructions that manipulated stereotype threat and before they took the difficult math test (if any of these variables mediate effects of stereotype threat, they should vary with the instructions with the test... that is: independent variable manipulation---and with performance on this test, thus accounting for effects of instructions on performance goal is to replicate same results of study 2, while eliminating confounding factors of less highly selected members, coming from a different university, with wider range of questions (to eliminate floor effect) and with control group that does not explicitly mention gender difference relevance or irrelevance---manipulation of independent variable by measuring participants evaluation apprehension, state anxiety, and self efficacy after instructions that manipulated stereotype threat and before difficult math test used as a preliminary test of interpretations to prove whether these variables mediated the effects of stereotype threat

study 3

=study 2 did provide evidence that stereotype threat can depress women's performance on a difficult math test and that eliminating the threat can eliminate their depressed performance HOWEVER, =3 reasons experiment did not make point convincingly as it might have: 1. floor effect on second test of study 2 raises possibility that effect of stereotype threat might be limited to small number of questions. 2. highly selected sample used in studies 1 and 2, raises possibility that stereotype threat effects may have limited generalizability. 3. study 2 explicitly stated there were gender differences, leaving open possibility that stereotype threat effects will only emerge when gender differences are alleged. *****THEREFORE, in this study sought to replicate effect of study 2 but with less highly selected sample from another university, on test with wider range of problems, and with control group in which no explicit mention of gender differences is made

study 1 results

=the pattern observed in the literature can be replicated with highly selected and identified group of participants =women underperformed in comparison to men on difficult test but did just as well on easy test

study 1 goal

=to test whether the pattern --women underperform on difficult tests but perform just as well on easier test-- holds true with highly selected participants

study 2 results

=two halves of test did not prove to be equally difficult -- mean for both and women on second test was not different from 0---creating floor effect for that test (so used performance of first test as dependent variable)--test characterization then treated as b/w participants fact such that people told first test yielded differences made up one level of factor and people told it did not yield gender difference made up other level =when P's explicitly told that test yielded gender differences, women GREATLY underperformed in relation to men =when test was purported not to yield gender differences, women performed same level as equally qualified men =presenting test as one on which gender differences do not occur, we made stereotype of women's math inability irrelevant to interpreting their performance on test (this particular test) *STRONG EVIDENCE that women's underperformance on these difficult math tests result from stereotype threat, rather than from sex linked ability differences that are detectable only on advanced math material

central position on stereotype threat

=when a stereotype about one's group indicts an important ability, one's performance in situations where that ability can be judged comes under an extra pressure -- that of possibly being judged by or self-fulfilling the stereotype -- and this extra pressure may interfere with performance

example of stereotype threat

=when women perform math, unlike men, they risk being judged by negative stereotype that women have weaker math ability --hypothesize that apprehension it causes may disrupt women's math performance (the negative stereotype causes apprehension and disruption of performance of the people stereotyped against)

study 3 results

=women underperformed relative to men in control condition =women performed equally with men in no-gender-difference condition =men outperformed women only in control condition =women's math performance improved when stereotype threat was decreased =THUS, replicated results of study 2 with less selected population, with a test with more items, and comparing the no-gender-differences condition to a control group where no explicit mention of differences was made with this and study 2: compelling evidence that reducing stereotype threat can increase women's math performance =nonsignificant tendency for men to perform slightly worse in no-gender-differences condition than in control conditions--- possibly suggesting that characterizing the test as not producing gender differences has a negative effect on men's performance ^^ allows us to begin exploration of the mediation of effect of test characterization on women's math performance =3 possible mediators: evaluation apprehension, anxiety, and self-efficacy -----to verify these measure, they conducted factor analysis, with varimax rotation, on questionnaire that assessed them --4 factors emerged with Eigen values greater than 1---these correspond closely with 3 mediational variables intended to measure, so made scales for these variables from items that loaded most strongly on factors to which they correspond =examined effect of stereotype threat manipulation on each potential mediator, and then examined whether these mediators accounted for effect of stereotype threat manipulation on women's test performance. finally examined each of these mediators individually to see if they could account for effect of stereotype threat manipulation on women's performance =no evidence of relationship between stereotype threat manipulation and evaluation apprehension---however, manipulation did have marginally significant effect on anxiety ---representing test as gender-fair tended to lower women's anxiety, but had no discernable effect on their evaluation apprehension or their self-efficacy =examined whether any of potential mediators predicted test performance and whether these mediators accounted for effect of stereotype threat manipulation on test performance =analysis revealed that anxiety was significantly related to performance, evaluation apprehension was significantly related to performance, and self-efficacy was not significantly related to performance ---anxiety and evaluation apprehension were related to test performance, but self-efficacy was not

study 1

replication of pattern in literature--women underperform in comparison to men on difficult tests, but perform equally with men on easy tests-- in sample of highly qualified equally prepared men and women (men/women selected to have strong math background) -varied difficult on test given -difficult test taken from GRE exam in mathematics--involving calculus, abstract algebra, and real variable theory -easier test taken from quantitative section of GRE general exam--advanced algebra, trigonometry, and geometry -administered on computer allowing measurement of amount of time participants spent on the test--assessing extent to which differences in performance might be related to differences in participants effort

operational definition of stereotype threat in relation to women and math study

widely known stereotype that women have less ability in mathematics and related domains than men THUS, in situations where math skills are exposed to judgments-- formal test, classroom participation, or simply computing waiter's tip--women bear the extra burden of having a stereotype that alleges a sex-based inability