Developing and Establishing Psychometric Properties of Measure of Bystander Intervention for Problematic Alcohol Use

Background: Alcohol consumption is a common occurrence among college students, with roughly two thirds of college students reporting drinking alcohol. Alcohol is often used to facilitate social interactions among college students. Thus, there is potential for harm reduction via bystander intervention. Currently no adequate measures to assess bystander intervention for problematic alcohol use (PAU) exist. The purpose of this project is to develop and validate reliable measures of bystander intervention for problematic alcohol use, including measures of likelihood to intervene, barriers to intervention, and actions taken in response to situations specifically regarding PAU.

Methods: We developed initial measures of bystander intervention for PAU, based on both college student (n=79) and expert feedback (n=8), that include the constructs of likelihood to intervene, barriers to intervention, and actions taken in response to situations specifically regarding PAU. In the current study, we sought to confirm the factor structure of the initial measures using confirmatory factor analysis (CFA). We administered the measures in a college student sample (N = 1,270) recruited through Prime Panels. Participants were required to be between the ages of 18 and 25, enrolled full-time as a college student in the United States, and report being around a friend who was drinking in a concerning way in the past 3 months.

Results: We estimated models using maximum likelihood, and evaluated model fit using guidelines for RMSEA, CFI, and SRMR. CFA analysis resulted in the following two scales: 1) Alcohol Likelihood to Intervene (14 items), 2) Overcoming Bystander Alcohol Barriers to Intervene (22 items). Of note, the Bystander Alcohol Actions (22 items) was formative, no CFA was conducted. Additionally, scales demonstrated acceptable convergent and discriminant validity, with significant but non-overlapping correlations with constructs such as IPV Bystander Behaviors, Alcohol Communications, or Party Safety behaviors. The Actions measure was added to the model in order to estimate factor correlations among the three main factors and we found that the hypothesized model displayed significantly superior fit to all other models. The Alcohol Likelihood to Intervene scale consisted of 3 subscales: 1) looming risk of bodily injury; 2) signs or risks of overconsumption; and 3) risk of sexual violence perpetration. The Overcoming Bystander Alcohol Barriers to Intervene scale consisted of 3 subscales: 1) situational factors; 2) fear of external consequences; and 3) social-cognitive factors. The Bystander Alcohol Actions scale includes items related to actions taken before drinking, as well actions taken in reaction to observing PAU. Measurement invariance analyses indicates that the measures are hold up equitable across subgroups of students.

Conclusion: These measures capture important intervention constructs for PAU. These scales can be used to inform and evaluate bystander intervention programs aimed at reducing rates of PAU among college students.