Incorporating Variability in Measuring Alcohol Perceptions in School Settings

Social norms and environmental perceptions influence adolescent behaviors in schools. Past research focused on overall school perceptions of alcohol use and binge drinking, little work incorporates the variability of student perceptions. The amount of agreement between students at a school could determine the influence perceptions of norms have on drinking behaviors. A generalized variance approach accounts for variability across the full distribution of categories; the current study uses this measurement strategy to account for the variety of alcohol perceptions within schools. Thus this paper will allow for an understanding of the relationship between both majority perceptions as well as consistency of perceptions across student populations and alcohol use in high schools.

Data are taken from the Maryland Safe and Supportive Schools Climate Survey, which aims to systematically measure school climate. Students in 58 schools (n=26,034) completed a school climate survey during Spring 2013. Perception of alcohol use norms was measured with two items (how much of problem students’ use of alcohol is at their school; how risky is having a drink daily). The first item was also used to create a measure of variability, using a formula derived from Simpson’s Diversity Index where responses were used to create a score representing the probability two randomly selected students will have different perceptions. Scores range from 0 – 1, with higher scores reflecting a greater diversity of alcohol perceptions reflected in that school’s population. Two separate multi-level regression models examined the relationship between these variables and two outcomes – amount of days of alcohol use and amount of days binge drinking within a month’s time – controlling for clustering within schools and known school and individual correlates.

 At the individual level, students’ perceptions of alcohol norms were significantly related to both use and binge drinking with more approving drinking norms predicting increased behavior. At the school level average norms did not predict use or binge drinking, however both perception variability and the interaction between variability and norms predicted both use and binge drinking. Consistent with hypotheses, variability was negatively related to outcomes meaning that more diversity of perceptions predicted less use. Further analysis shows a consistent gap in drinking behavior between low and high variability of norm perceptions schools across approving alcohol norms, with drinking behavior. The findings suggest that modeling variety, not just strength of perceptions is important in capturing the true effect of social influence in high schools. Implications of these findings for data-based decision-making are considered.