Predictors of Sexual Harassment Using Classification and Regression Tree Analyses and Hurdle Models: A Direct Replication

ABSTRACT

Sexual harassment affects a large percentage of higher education students in the US. A previous study identified several risk factors for sexual harassment using hurdle models and classification and regression tree (CART) analyses. The purpose of the present study was to assess the robustness of these findings by replicating the analyses with a new sample of students. Secondary data analysis was conducted using data from 9,552 students from two- and four-year colleges. Hurdle model coefficients were assessed for replicability based on statistical significance and consistency of the replication effect size relative to the original effect size. Kotzé et al.’s findings were robust, with 91% of all tested effects meeting at least one of two replication criteria in the hurdle models and 88% of the variables replicating in the CARTs. Being younger, consuming alcohol more frequently, attending a four-year college, and having experienced more prior victimization and adversity were important predictors of peer harassment whereas being LGBQ+ was an important predictor of sexual harassment from faculty/staff. These findings can inform targeted prevention and intervention programs. More research is needed to understand why certain demographic and contextual variables are associated with greater harassment risk.

Acknowledgments

We would like to thank Moin Syed and Etienne LeBel for helpful feedback on an earlier version of this manuscript.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Data Availability

All code associated with this project can be found on the Open Science Framework: https://osf.io/zu26t/?view_only=b324e903040b4e389fe39f7c65e3f5ff. Data associated with this project can be accessed by contacting Katherine Lust ([email protected]).

Additional information

Funding

This material is based on work supported by the National Science Foundation Graduate Research Fellowship Program under Award No. [2237827] (for KAH). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. The authors received no other financial support for the research, authorship, and/or publication of this article.