Title: Stereotype threat as a source of measurement bias in cognitive ability tests
Location: New York University Kimball Hall (246 Greene St.) 607W.
Description: Speaker: Jelte M. Wicherts – University of Amsterdam
Abstract: Perhaps the most important issue in research of stereotype threat is whether it artificially lowers performance of stereotyped groups on real-life cognitive tests. The effects of stereotype threat on cognitive test performance may be viewed as a source of measurement bias (Wicherts, Dolan, & Hessen, 2005). As such, these effects have implications for cognitive test scores that can be formalized in psychometric models in order to shed light on the generalizability of stereotype threat beyond the social-psychological laboratory (e.g., on the actual SAT or GRE). Confirmatory factor models have been used to detect several scenarios of stereotype threat effects at the scale level (Wicherts et al., 2005). In addition, stereotype threat effects may have implications for group differences in validity coefficients (Wicherts & Zand Scholten, 2010). Although the results of differential prediction analyses have been proffered both in favor of and against the generality of stereotype threat effects (Cullen, Hardison, & Sackett, 2004; Walton & Spencer, 2009), these effects have no necessary implications for the outcomes of differential prediction analyses (Wicherts & Millsap, 2009). On the other hand, Arbuthnot (2009) recently found a relation between stereotype threat and Differential Item Functioning (DIF), suggesting that stereotype threat results in a failure of measurement invariance at the level of items. In this talk, I will consider stereotype threat from a latent variable perspective, discuss several scenarios of stereotype threat effects at both the scale and item level, show some empirical results, discuss implications, and propose some future research in this area.
Start Time: 02:00
Date: 2010-05-24