Assessing the Effectiveness of Our Program

At this point we can share some of the preliminary results from our violence prevention efforts. After administering our program to a group of 14- and 15-year-old freshmen in the fall semester of 2014, we did a quick post-test. We then followed up in the spring semester of 2015 gathering survey data from participating students in the fall as well as other students who had not participated in the program, asking about both dating behaviors and attitudes.

We measured students' attitudes on violence using the Attitudes Towards Female Dating Violence (ATFDV) and Attitudes Towards Male Dating Violence (ATMDV) scales (Price, Byers, & the Dating Violence Research Team, 1999) as well as select items from the Illinois Rape Myth Acceptance (RMA) scale (Payne, Lonsway, and Fitzgerald (1999.) The ATFDV and ATMDV ask students whether or not they find certain behaviors acceptable in a relationship. These behaviors could include punching or kicking a partner, pressuring him or her for sex, insulting or belittling a partner, and other forms of jealous, controlling and abusive behavior. These scales are a good way of gauging what kinds of behavior students might find normal, and whether or not they can recognize abusive behavior as abusive. The RMA asks students about their beliefs in common rape myths, like the belief that a woman who wears a short skirt is asking for trouble, or that men commit rape because they just get "carried away." The scale is useful for understanding whether or not students understand consent and to what extent they might blame the victim of a sexual assault. In each of these scales, higher numbers indicate greater acceptance of violence and unhealthy behaviors, so an effective program would lead to lower scores on each of these scales.

To gauge the effectiveness of the program we estimated an OLS regression model using results from the student survey data collected in Spring 2014. The regression model can measure the effectiveness of the program while controlling for factors beyond the classroom. (In our case we controlled for student dating experience, family background, religious background, race and gender.) We can then use the results of the model and input specific values to estimate the substantive impact of the program while holding other variables constant. The program had a statistically significant effect and reduced acceptance of violence across all scales. The figure below illustrates the substantive impact of the program. In the figure we estimate the impact of the program on male students across each of our measures. The program reduced acceptance of violence and rape myths by about 8% across all scales. (Though the effect on female students was not as large, it was still in the appropriate direction.)

 

These results are especially encouraging because we are measuring program effectiveness 6 months after the completion of our sessions. In other words, students didn't simply forget everything we taught them in the fall. Six months later, they still displayed significantly healthier attitudes about relationships and dating violence.

We're off to a good start, but we are still digging through the data both to better understand where we can improve the program, and to get a better handle on the general phenomenon of dating violence. In addition to data on dating violence, we collected data on other social behavior including social media habits, texting and real-world social interactions as well as other factors like television viewing habits and religious beliefs. We'll keep posting as we investigate how these factors contribute to dating violence and interact with our prevention efforts.