Abstract Background Alcohol‐induced blackouts are associated with heightened risk for alcohol‐related consequences among young adults. However, to date, the only validated measure of blackouts (the Alcohol‐Induced Blackout Measure‐2 or ABOM‐2) has only been studied in a college student sample. To ensure blackout measures are valid for use across young adult populations, this study evaluated the structural validity, convergent validity, and measurement invariance of the ABOM‐2 among noncollege young adults. Methods Young adults aged 18–29 years ( N = 610; 50.5% female; 42.1% white; 43.9% Black; 32.8% Hispanic) who were not enrolled in and never completed a 4‐year college degree participated in an online survey. The survey assessed hazardous alcohol use, negative alcohol consequences, blackouts, and intentions/willingness to blackout. Results Confirmatory factor analyses supported the correlated two‐factor (fragmentary and en bloc blackouts) structure of the ABOM‐2, demonstrating excellent model fit (CFI/TLI = 0.997/0.995, RMSEA = 0.068) and high factor loadings (>0.87). ABOM‐2 subscales had strong associations with hazardous alcohol use and consequences ( r s = 0.59–0.66) and moderate associations with intentions/willingness to blackout ( r s = 0.32–0.48). The en bloc subscale was significantly associated with both hazardous alcohol use and consequences, while the fragmentary subscale was only associated with consequences. The ABOM‐2 demonstrated configural, metric, and scalar invariance across sex (female and male), with minimal decrements in fit with additional model constraints. Subscale latent means across sex were not significantly different. Conclusions Findings support the validity of the ABOM‐2 in research contexts when a detailed assessment of blackout experiences is necessary. Our results showing the ABOM‐2 is valid among young adult males and females who did not complete 4‐year college support use of this measure outside of research conducted only among those in college. Future work should evaluate measurement invariance longitudinally and across additional identities (e.g., racial and ethnic identities).
Strowger et al. (Sun,) studied this question.