An Exploratory Comparison and Evaluation of Two Two-Step Measures to Identify Transgender People in Survey Datasets

This study compares and evaluates two distinct two-step approaches to identifying transgender people in survey datasets. Traditional two-step methods using sex assigned at birth (SAB) and current gender identity remain dominant. However, they have notable limitations. Gender modality, or the relationship between SAB and current gender identity (e.g., cisgender, transgender, or something else), presents an important alternative item to consider.

Using an online, cross-sectional survey of 952 sexual and gender minority adults in the United States, we conducted an exploratory analysis of categorization divergence/convergence using two approaches: (1) a modified traditional two-step (SAB + current gender identity) and (2) an alternative two-step (current gender identity + modality).

Convergence between approaches was 95%. Rates of refusal for all questions were low, although slightly higher for gender modality. Divergence fell into three categories: (1) individuals grouped as "Questioning" by Approach #2, but not #1 (n=21; 44.7% of divergences); (2) individuals categorizable by one approach, but not the other (n=13; 27.6% of divergences); and (3) individuals whose gender modality differed between approaches (n=13; 27.6% of divergences).

We found preliminary evidence for the utility of an alternative two-step approach, particularly when within-group differences among transgender populations are relevant. Both the traditional two-step model and the alternative we tested have limitations which should be ameliorated through future research. Cognitive testing is necessary to evaluate explanations of divergences. We identify priorities to expand on the relative strengths of our alternative approach and address the remaining limitations and areas of uncertainty it highlights.