Measurement implications associated with refinement of sexual and gender identity survey items: A case study of the National College Health Assessment.

Objective: To quantitatively measure psychometric impact associated with changes to sexual and gender identity survey questions included on the National College Health Assessment (NCHA). Participants: Respondents included all iterations (31) of the NCHA between the years of 2000 and 2015 (N = 1,202,582). Methods: Secondary analysis using complete NCHA data across 15 years. Results: Scale reliability for cisgender men, cisgender women, gay and lesbian, and bisexual groups improved when transgender was moved to reflect a gender option, rather than a sexual orientation category. Scale reliability for the alcohol protective behaviors measure was consistently higher for heterosexual persons than sexual minorities. Conclusions: Less inclusive, less representative conceptualizations of sexual and gender identities on survey questionnaires can result in poorer measurement of social characteristics and behaviors. Researchers and practitioners should seek to ensure their assessments and evaluations include demographic items that capture important nuances associated with human sexual and gender identities.