STEM is an important part of who I am: The Development of the STEM Identity Scale

Given that the underrepresentation of women and racial and ethnic minoritized groups in STEM continues to be an issue, scholars have sought to understand the potential beneficial effects of STEM disciplinary identity on the persistence of underrepresented groups in STEM pathways. In line with this goal, Chapter 1 utilized a meta-analysis to investigate the strength of the relationship between STEM disciplinary identity and persistence. Across 31 studies, a positive and medium to strong relationship between STEM disciplinary identity and persistence emerged, such that students with stronger levels of STEM disciplinary identity reported greater perseverance in STEM pathways. Further, this relationship was moderated by the percentage of Black students included in the studies, with directions of effects suggesting a weaker connection between science identity and persistence for Black students compared with other students. Lastly, the review revealed several limitations associated with the instruments used in past research to capture STEM disciplinary identity. Thus, Chapter 2 utilized best practices in measurement development to create a new multidimensional self-report measure of STEM identity. Item development was informed by expert review and cognitive interviews with undergraduate students majoring in a STEM field. An exploratory factor analysis (EFA) was conducted with a racially diverse sample of 191 undergraduate students. The EFA resulted in a five-factor solution that captured Interest, Competence-Performance, Recognition in STEM, Recognition and Support Outside of STEM, and Self-Recognition of STEM Identity. Additional exploratory analyses comparing subgroups found statistically significant differences between students majoring in a STEM field and non-STEM students in all five dimensions of the STEM Identity Scale, suggesting evidence of discriminant validity. No significant differences were found on the SIS comparing students using sex, race, ethnicity, and generation student status. The implication of these findings and suggestions for future psychometric evaluations of the STEM Identity Scale are discussed. In conclusion, the development of the SIS is a critical contribution to the STEM identity literature by improving on the limitations of past instruments