Stereotype threat theory (STT; Steele & Aronson, 1995) emerged from the field of the psychology of social regulation of performance. STT offers a target’s perspective on the effect of negative stereotypes on performance. According to Steele and Aronson (1995), stigmatized targets underperform on challenging tests partly because “they face the threat of confirming or being judged by a negative societal stereotype –a suspicion– about their group’s intellectual ability and competence” (p. 797). This underperformance can place stigmatized targets at a disadvantage in the education system. In their seminal studies, Steele and Aronson (1995) varied the stereotype vulnerability of African American students by varying test presentation. The same test was described as a diagnostic test of intellectual ability vs. a laboratory tool (i.e., nondiagnostic). When African American students perform a diagnostic test of intellectual ability, unlike White students, they risk being judged by the negative stereotype that African Americans have a weaker intellectual ability but not when they perform a nondiagnostic test. Indeed, African American students underperformed in comparison to White students only when the test was described as diagnostic of intellectual ability. Stereotype threat can be experienced by any group member in a domain in which group ability is expected to be poor (e.g., women in math; Spencer, Steele, & Quinn, 1999). However the social context can have profound implications for stereotyped targets. Marx and Goff (2005) showed that when an African American experimenter gave a diagnostic test to African American students, their performance did not suffer. It means that merely varying who administers a test (the source) can alter the beliefs targets have about how they may be stereotyped (Marx & Roman, 2002; Marx, Stapel, & Muller, 2005). Indeed, a competent African American experimenter (i.e., a role model) provides stereotype disconfirming information. In the current study, we also varied who activates the negative stereotype, but in terms of the source’s implication in the experimental hypothesis. The stereotype was either activated by an outgroup experimenter, or by a bogus participant (i.e., a confederate). The outgroup experimenter (high implication) will provide stereotype confirming expectations but not the confederate. We postulate that only the experimenter will threaten participants and therefore lead to performance decrement. Conversely, STT postulates no effect of the source’s implication.

Participants were 126 female introductory psychology students. Participants first performed a math test in small groups, which performance would be used as an individual baseline. Participants then performed a second math test under stereotype threat conditions. In the high implication condition (HIC) a male experimenter added “this test had been shown to produce gender differences” (Spencer et al., 1999). In the low implication condition (LIC) a bogus female participant wondered whether or not “this test had been shown to produce gender differences”. In the control condition no mention of “gender” was made.

Results revealed that in comparison to individual baseline, participants of the HIC performed less well in comparison to both LIC and control conditions. Performance in these last two conditions did not differ significantly (the hypothesis of the source’s implication fits the data). These results are encouraging. We will discuss them in terms of their applications in classroom settings. Indeed, who conveys the negative stereotype, teacher vs. other pupils, should lead to different interventions to reduce or even eliminate stereotype threat.