ConnectEd: A Randomized Controlled Trial Connecting Through Educational Training

Rates of sexual violence (SV) or dating violence (DV) remain high among college students (22% in the past year for female undergraduates). Binge drinking commonly occurs (25% of undergraduates report binge drinking in the past month) and is a strong predictor of sexual risk taking and violence victimization and perpetration. Given connections between SV/DV risk and alcohol use and abuse, many advocate prevention programming to jointly address SV/DV and alcohol abuse. Bystander interventions, recognized as promising violence prevention strategies, are unique in their engagement of all community members to 1) recognize situations that may become violent and 2) learn to safely and effectively intervene to reduce violence risk. Based on their promise, the Campus Sexual Violence Elimination Act (SaVE) now requires all publicly-funded colleges to provide bystander intervention in the hopes of reducing sexual violence. No randomized intervention trial has compared bystander interventions focused exclusively on SV/DV prevention with those additionally addressing alcohol abuse prevention within the bystander context. We propose such an experimental trial at the University of Kentucky (UK), a nationally recognized leader in addressing sexual violence through bystander intervention programming. The purpose of this randomized intervention trial is to evaluate the relative efficacy of bystander training elements. In this study will be randomized to one of the following: 1) Online Bystander Training (Haven), 2) Online Substance Abuse Training (AlcoholEdu), 3) In-person Bystander Training (Green Dot Intensive Training), or 4) In-person Bystander/Substance Abuse Training (ConnectEd). Aim 1: Evaluate methods to capture behaviors and diffusion of training through social media communications (e.g., Facebook, Twitter, Instagram). Aim 2: Determine the relative efficacy of the three bystander training conditions to a) increase bystander efficacy and behaviors by domain (alcohol abuse and violence prevention), b) reduce acceptance of SV/DV, alcohol's role in sexuality, and alcohol use / abuse, and c) reduce risk taking behaviors by domains of sexual behavior and alcohol use or abuse. Aim 3: Determine the relative cost of implementing bystander approaches and their cost effectiveness. Design: Incoming UK undergraduates will be recruited and randomized to one of the four conditions and followed for 9 months, one academic year (n=4000, total over 2 academic year cohorts). A mixed-method approach will be used to measure intervention fidelity and costs over time and by condition. Using the most rigorous design available, we seek to generate new understanding of how novel bystander intervention approaches work to reduce not only SV/DV risk taking behaviors but also reduce alcohol abuse.

Incoming students will be invited to participate in a randomized trial testing bystander program efficacy. Those consenting to study participation will be randomized to one of the four conditions. These conditions include:

1. In-person Green Dot Intensive Bystander-based Sexual Violence Prevention Training (GreenDot);
2. In-person Green Dot Intensive Bystander Training combined with Substance Abuse Prevention Training (WellAware);
3. Online Bystander-based Sexual Violence Prevention Training (Haven), and;
4. Online Substance Abuse Training (AlcoholEdu).

All incoming students (17-24) agreeing to study participation will be invited to complete surveys at baseline (prior to training), and at end of the fall and spring semesters. Additionally, students will be asked to complete micro-surveys using the myUK app.

Student recruitment will be conducted as students are entering UK during their summer orientation (June and July). We anticipate that at 45-50% of students will agree to study participation and provide data; thus our final participant pool is expected to consist of 4,000 participants from academic years 2016 and 2017.

Continuous variables will be summarized with descriptive statistics (n, mean, standard deviation, median, first and third quartiles, and min and max); categorical variables will be described with counts and percentages. Change scores and percent change scores will be calculated from baseline (prior to start of first academic year) as well as from end of Fall to Spring within each academic year. Numerical and graphical summaries will be provided overall, by subpopulation within college communities (e.g., by gender, sexual attraction, fraternity/sorority or athletic team membership) and most importantly by training condition. Simple comparisons of groups will be made using ANOVA for continuous variables and chi-square tests for categorical outcomes. Comparisons between groups will be conducted as randomized; simple comparisons of continuous variables between groups will be performed using ANOVA and chi-square tests of independence for categorical variables. Although groups will be randomly assigned, potential confounders will be examined with bivariate analyses and comparisons requiring covariate adjustment will use regression modeling (e.g. ANCOVA, logistic regression). Planned subgroup analyses include analyses by gender group, by sexual attraction, by Greek life participation. These will also be conducted as randomized and as received. Effect modification will also be assessed by including interactions in regression models (linear, logistic, and mixed models depending on outcomes and the presence of longitudinal measures). SAS v9.4 will be used for all data management and statistical analysis; a significance level of 0.01 will be used for all statistical tests.