Limitations in STEM Identity Surveys to Capture College STEM Students' Marginalizing Perspectives on STEM Personhood

• In Event: Advancing Methods for Evaluating Construct Validity in STEM Education Research

Abstract

We show how surveys validated to measure STEM identity may be limited in their discernment of the ways restrictive and marginalizing criteria may be applied in self-assessments of STEM personhood. We recommend further validation work to develop these tools for purposes aligned with equity efforts in STEM education.

Theoretical Grounding
We use the identity model of Hazari et al. (2010), which postulates STEM identity contributors of interest, recognition, and performance-competence. This work and surveys derived from it uses academically-oriented language and touts its value in the predictive relationship between identity and career aspirations. However, this coupling with academically success and career intentions suggests a limitation in the ability for these surveys to differentiate STEM personhood felt by individuals within a population of STEM-affiliated individuals. This is may be an important area in need of development in these surveys given the attrition of individuals identifying with marginalized racial and ethnic groups in STEM majors (Riegle-Crumb et al., 2019) and research suggesting non-majority students may feel or be seen as less confident or competent in STEM despite better or equivalent aptitude (Eaton et al., 2020; Marshman et al., 2018).

Methods and Methodology
We use explanatory mixed methods (Creswell et al., 2003) to study how college STEM students attending a Hispanic Serving Institution described their STEM personhood.

Our survey reflected the model of Hazari et al. (2010). Validation efforts included expert review, factor analysis, and external validation through prediction of STEM major (Flake et al., 2017).

We selected 20 students for phenomenological interviews (van Manen, 2014) whose survey responses indicated a high sense of STEM identity and represented varied demographics and fields of study. Interviews elicited experiences that contributed to how participants thought of themselves as STEM people. We conducted nine dialogic follow-up interviews (Dinkins, 2005).

Findings
Structural and external validation analysis suggested the survey was appropriate for our purpose of assessing college student STEM identity (Authors, 2022). Yet, although all interviewees rated themselves highly as STEM people, many discussed ways that they fell short of their image of a STEM ideal, particularly females in life science tracts. Participants particularly differentiated themselves in terms of performance-competence, verbally rating themselves below their image of an ideal STEM person when asked to describe how they thought of themselves as a STEM person. Further, interviews indicated a tendency to distinguish positive STEM personhood through comparison to family members perceived as lesser of STEM people due to experiences associated with their background as immigrants.

Significance
Even when their use is supported through established validation techniques, survey instruments may yield results that may easily be misinterpreted or misused, and results of survey analysis may perpetuate, rather than reveal, marginalizing perspectives. In this case, the survey instrument failed to detect differences in STEM identity that respondents verbalized in interviews. We have begun substantive validation of a revised survey instrument by considering items with comparative components, in alignment with how interview participants expressed their assessments of their STEM identities (Wan et al., 2021).

Authors

• Heidi Cian, Maine Mathematics and Science Alliance
  Presenting Author

• Remy Dou, Florida International University
  Non-Presenting Author