(PS8-55) Measurement Invariance and Mean Differences on the Kessler K6 Screening Scale for Psychological Distress Across Sexual Minority and Heterosexual Individuals

The Kessler K6 non-specific psychological distress scale is frequently used as a screener for severe mental health difficulties in epidemiological studies in the United States. Despite its frequent use, there is a paucity of research examining the measurement invariance of the K6 across sexual orientation. Previous research has documented mental health disparities between sexual minority and heterosexual individuals, including research that has found significantly higher mean K6 scores (i.e., greater psychological distress) in sexual minority individuals relative to heterosexual individuals. However, because measurement invariance of the K6 has not been established across these groups, it is important to determine whether these mean differences are due to artifactual differences in the factor structure of the K6 across sexual orientation or whether they represent true differences in psychological distress. The present study used data drawn from the Well-Being and Basic Needs Survey (WBNS) to examine factorial invariance of the K6 across heterosexual and sexual minority individuals, and to then examine differences in mean K6 score across groups. Data were drawn from the 2017, 2018, 2019, 2020, and 2021 rounds of the WBNS. The present analyses are based on 1,765 sexual minority individuals (7.9% of the sample) and 20,632 heterosexual individuals who completed the K6 and information on their sexual orientation. Testing factorial invariance was conducted using multigroup confirmatory factor analysis (CFA) modeling. Analyses were conducted with the EQS 6.4 program. A CFA evaluating a single-factor structure of the K6 yielded a reasonable fit to the data in the sample of sexual minority individuals. Measurement invariance was then hierarchically tested on the single-factor model of the K6. Overall, results indicated factorial invariance for this single-factor model across groups, suggesting that any mean differences between groups are due to mean differences in the latent underlying construct rather than to mean differences that vary from item to item. Given evidence of measurement invariance of the K6, we proceeded to test for latent mean differences in psychological distress between the two groups. Results indicated that the latent mean for the group of sexual minority individuals was significantly higher than the mean for the group of heterosexual individuals. Finally, we evaluated differences between heterosexual and sexual minority groups on the percentage of people who met criteria for serious mental illness (i.e., the percentage of people who scored ≥13 on the K6). The percentage of sexual minority individuals who reported serious mental illness was significantly higher than the percentage of heterosexual individuals who reported serious psychological illness. Taken together, the results support the measurement invariance of the K6 across heterosexual and sexual minority individuals, and suggest that, as has been found in previous research, psychological distress is higher among sexual minority individuals than heterosexual individuals.