ADVANCING STEREOTYPING RESEARCH: HOW AND WHY TO USE LINEAR MIXED-EFFECTS MODELS IN GENDER STEREOTYPING RESEARCH

We present two studies examining the valence of gender stereotypes using linear mixed-effects models (LMEMs) to demonstrate how they can advance stereotyping research. Although LMEMs are common in some domains of psychology (e.g., developmental and cognitive psychology), they are much less commonly used in research on stereotyping. And yet, concerns about the generalizability of results, inflation of type I errors, and ease of handling missing data are equally relevant to work on stereotyping. In this paper, we first summarize what LMEMs are and how they might be applied to work on stereotyping. We then show how LMEMs can be used to analyze data from studies with researcher-generated attributes (Study 1) and participant-generated attributes (Study 2). We also show that LMEMs are particularly appropriate for designs that employ planned missingness (Study 2). Finally, we discuss how LMEMs may allow researchers to resolve conflicting findings in gender stereotype research and how designs with planned missingness allow researchers more flexibility in answering their research questions.