Assessing measurement invariance of scales of likelihood and barriers to bystander intervention for problematic alcohol use among college students

BACKGROUND: Problematic drinking among college students is a serious issue that leads to dire consequences. Bystander intervention (BI) for problematic alcohol use (PAU) is a promising approach that capitalizes on social networks to prevent PAU. Reliable and valid measures of BI for PAU that operate equivalently across demographic groups are needed. PURPOSE: The purpose of this investigation is to assess the measurement invariance of 2 validated scales of BI for Plikelihood to intervene and self-confidence to overcome barriers to intervention. Measurement invariance was tested across 6 groups, including (1) student athletes, (2) Black, Indigenous, and People of Color, (3) Greek affiliation, (4) sexual and gender minority students, (5) first year in college, and (6) men. Each target group was compared to a random sample of people not in the target group. METHODS: Participants (n = 1190) were recruited via Prime Panels to complete the Bystander Alcohol Likelihood to Intervene (BALI) and Overcoming Bystander Alcohol Barriers to Intervene (OBABI) scales. Random subsamples of groups of interest were created to ensure equal size within subgroups. Assessments of measurement invariance occur in a stepwise process: (1) configural, (2) metric, and (3) scalar. RESULTS: Configural invariance was established for each group. Metric invariance was not established, so scalar invariance was not investigated. CONCLUSIONS: Establishing configural invariance means that constructs and subfactors of BALI and OBABI operate equivalently between groups. However, the lack of metric invariance suggests that the items are not endorsed equivalently between these different risk groups. Further research should interrogate where discrepancies lie for items on the BALI and OBABI.