Different Approaches to Modeling Intersectionality: A QuantCrit Exploration of Stereotype Threat Patterns

• In Event: Intersectional and Critical Approaches to Examining the Experiences and Success of Minoritized Undergraduates in STEM

Abstract

Objectives & Theoretical Framework. Our aim was to explore two questions: (1) What are the developmental patterns in stereotype threats (STs) among undergraduates majoring in STEM? (2) How do modeling decisions regarding participant ethnicity and gender steer analyses toward an anti-racist agenda? To explore patterns in STs from an anti-racist perspective, we employed tenets of QuantCrit, a quantitative framework informed by Critical Race Theory (Garcia et al., 2018). Past work on stereotype threat (Steele & Aronson, 1995) has explored gender and ethnicity STs separately, but it is also important to investigate patterns that correspond to intersectional stereotypes—those associated with overlapping identities that uniquely combine to contribute toward discrimination (i.e., gender and ethnicity; Crenshaw, 1989).

Methods & Results. Data were from a broader longitudinal study of 4,261 undergraduates’ motivation in STEM, which included measures of gender- and ethnicity-based stereotype threat (Steele et al., 2002). We tested three growth models of STs that used increasingly nuanced approaches for examining gender and ethnicity. Model 1 (M1) used a multigroup latent growth curve to model STs, arranging students into four groups by gender (female vs. male) and ethnicity-status [minoritized in STEM (Black/AA, Latinx, Multiracial) v. non-minoritized (White, Asian)]. Model 2 (M2) used a more granular approach where 10 groups were formed by combining gender and ethnicity (e.g., Asian female students) and revealed patterns not captured by M1. For instance, M2 revealed that Asian female students had increasing ethnicity-based STs, a pattern not evident in M1. Both models estimated common experiences of students of shared gender and ethnic identities, but with different assumptions about the role of ethnicity in STEM contexts. In contrast, M3 applied a growth mixture modeling approach, challenging the assumption that all members of a demographic group would have the same pattern of ST, but still exploring whether students of a particular identity were more likely to follow a certain pattern in ST development. Some patterns in the M3 solution were similar to those found in M1 and M2. However, M3 also estimated new patterns (6 total); for instance, a pattern that had among the highest initial level of ethnicity and gender ST was comprised of participants with heterogeneous ethnic and gender identities. This suggests that M3 captured heterogeneity within groups of participants defined by ethnicity and gender that prior models did not, demonstrating the strength of the approach for expressing nuances of developmental processes among groups who experience oppression (Suzuki et al., 2021).

Significance. This study provides evidence that ethnicity-based ST varies by gender and vice-versa, suggesting an intersectional approach should be adopted in future research on their developmental patterns. Moreover, our comparison of models M1-3 makes transparent how researcher decisions regarding race and gender provide different affordances in estimating heterogeneity within and between participants of different identities. Researchers wanting to dismantle deficit models for minoritized students should consider modeling decisions and what insights can be uncovered by purposefully comparing multiple approaches. Doing so transparently allows us to interrogate narratives supported by different models and provide critical counternarratives.

Authors

• John Keane, Michigan State University
  Presenting Author

• Alexandra Lee, LearnPlatform by Instructure
  Non-Presenting Author

• Garam A. Lee, Michigan State University
  Non-Presenting Author

• Tony Perez, Old Dominion University
  Non-Presenting Author

• Lisa Linnenbrink-Garcia, Michigan State University
  Non-Presenting Author