(PS7-B38) The Structure of Distress: Understanding the Relationships Between Depression, Anxiety, and Stress in Trans and Non-binary Individuals

The Minority Stress model posits that elevated distress in transgender and non-binary (TNB) people results from unique experiences of discrimination/prejudice (Su et al., 2016; Testa et al., 2015). While we can identify social causes of distress in TNB people, the phenomenology of distress in this population remains unclear. Specifically, we do not know whether this distress is solely elevated in magnitude, relative to cisgender people, or if it differs structurally from distress in other populations. Measures of distress that have been conceptualized and normed in cisgender (or unselected) samples, such as the widely used Depression, Anxiety, Stress Scale -21 (DASS-21; Lovibond & Lovibond, 1996), present a barrier to understanding TNB distress. While one prior analysis (Lindley & Bauerband, 2022) supported the original 3-factor structure of the DASS-21 in a TNB sample, this work relied on a confirmatory factor analysis and thus was not positioned to detect any potential better-fitting factor structures unique to TNB participants.

We conducted an Exploratory Factor Analysis of the DASS-21 to identify potentially unique distress phenomena in TNB individuals. We hypothesized that due to the minority stress they experience, our TNB sample would show more overlap between the Anxiety and Stress subscales, compared to the traditional structure developed in cis/unselected samples, in which these scales are distinct. We used DASS-21 data from 271 TNB participants (ages 18 – 76, M_age = 32.88) from an online cross-sectional study of U.S. LGBTQ+ adults (N = 893).  A principal axis factor analysis was run (quartimax rotation), and 2-, 3- and 4- factor models were examined for model fit and construct validity.

All models had fit indices in the acceptable range. While RMSEA was significantly greater as number of factors increased (4 > 3 > 2 factor), there were no significant differences in fit indices. The 4-factor model (RMSEA = .055, SMSR = .03, TLI = .943) was discarded due to suspected overfitting. The 3-factor model (RMSEA = .068, SMSR = .04, TLI = .911) was also discarded. Several items (over-react [8], making a fool of myself [11], no enthusiasm [18]) had significant cross loadings (i.e., > 0.35 on >1 factor). In addition, the third factor had only 3 unique items (2 cross-loadings) with low face validity as a standalone construct. The two-factor model was chosen to balance concerns about model fit (RMSEA = .082, SMSR = .05, TLI = .872) with construct validity. The first factor, accounting for 30% of variance, was a combination of the DASS-21 stress and anxiety items. The second factor, accounting for 19% of variance, mirrored the original DASS-21 Depression factor. Item 5 (difficulty with initiative) loaded onto both factors (0.31 – 0.35), indicating that motivational difficulties may be related to both a more general stress response and low mood. Our finding of a combined stress/anxiety factor provides preliminary evidence for a unique distress phenomenology in TNB individuals relative to cisgender people. Implications will be discussed not only regarding the validity of common clinical research tools, but also for their potential to inform diagnosis, intervention choice, and outcome monitoring for common mental health conditions in TNB people.