Peer influence contexts of alcohol use among first-year college students: Investigating the roles of race, ethnicity, and gender through multigroup measured variable structural equation modeling
Snyder, Kathryn Renee Baird
Komives, Susan R.
Hancock, Gregory R.
MetadataShow full item record
The study purpose was to examine the contributions of peer context variables to the explanation of alcohol use of first-year college students by racial-ethnic group and by gender. Social norms theory and the theories of planned behavior, social identity/self-categorization, and status/status construction contributed constructs. Construct-related scores from sample survey responses demonstrated strong reliabilities ranging from .70 to .97. The following constructs provided measures for the study: Normative perception, subjective norm, affective attitude, cognitive attitude, social identity/self-categorization, status value, perceived behavioral control, intention and alcohol use. Normative perception and subjective norm were combined to create a single scale with stronger reliability than either had separately. Both cognitive and affective attitude were combined to create a single scale. Normative perception and attitude were measured the summer prior to college and in the fall; alcohol use was measured in the fall and in the spring. All other model variables were measured in the fall. Survey data were collected online in three waves and were from a representative sample (N=837) at a large state research institution with a predominantly White (65%) undergraduate student body. Rates of self-reported past month alcohol use and heavy episodic drinking of participants were comparable to those of similar samples in national and in-state studies. Applying multigroup measured variable structural equation modeling, the model explained between 60% of the variance in spring term alcohol use for Asian Pacific American students and 92% for African American/Black students. Data-model fit was acceptable (NFI, CFI > .95, SRMR < .08) for all groups in both analyses. Direct, indirect, and total effects of model variables were identified for each of five racial-ethnic groups in the study (African American/Black, Asian Pacific American, Latino/Latina American, White American, and Multiracial/Biracial American) and by gender for White men and White women. Tests of invariance demonstrated where specific paths in the model were significantly non-invariant (differed) and for which groups. Findings suggest the importance of pre-college intervention, the risk of increased alcohol misuse for first-year students, and the conditional effects of racial-ethnic group and gender.