Violent and sexual victimization among American and Canadian college students: Who is more at risk and are the risk factors invariant?

ABSTRACT

The victimization of college students is well documented. Comparative research on the extent of victimization and the risk factors for it is less prevalent. Although some cultural similarities exist, it is possible that American and Canadian college students differ in their experience of victimization. These differences have not been fully explored. To do so, we use data from the American College Health Association’s National College Health Assessment II, which assesses health and risk behaviors of college students. In 2013, this survey was also administered to Canadian college students. We find that Canadian college students have higher risks of both violent and sexual victimization, likely due to their higher levels of several risk factors. In addition, key differences in the risk factors for violent and sexual victimization emerged. These findings have implications for the development and implementation of risk reduction and prevention programs on college campuses.

KEYWORDS:

• Victimization
• canada
• sexual victimization
• college

Notes

1. Those individuals who indicated they were transgender were coded as missing on the gender variable because there were so few (n = 144, .02%) who identified this way.

2. The final sample size for the multivariate analysis is 49,823 for the models predicting violent victimization and 49,661 for those predicting sexual victimization. The sample size decreased from the original due to listwise deletion. We performed multiple imputation through chained equations using Stata 14.2 to impute missing values on 10 sets of data for the total sample, the American sample and the Canadian sample. We then reran our models with the imputed data. The results were not substantively different than those presented. Because of this similarity, we chose to present the models with listwise deletion so that our descriptive and bivariate analyses would correspond with our multivariate results. Stata does not have functions to perform bivariate analysis (t-tests or chi-square) using the MI estimate commands.