Study 1
Edwards and colleagues (2021) conducted a quasi-experimental design to evaluate the efficacy of the IMpower program on reducing sexual assault victimization among American Indian (also known as Native American) girls. Youth participants were American Indian middle and high school girls in grades 6–12, who were living either on an Indian reservation (in the Northern Great Plains region of the United States) or in a nearby city in the same state as the reservation. Specifically, girls on the reservation were recruited from one middle school and two high schools, while girls living in a nearby city in the same state were recruited from five middle schools and three high schools. Community and school partners felt it was an ethical concern to withhold intervention from some girls in the school. Because of this, and because of potential contamination effects, the intervention group comprised girls living on the reservation who participated in the 6-session, 2-hour classroom-delivered IMpower curriculum, and the comparison group comprised girls in a nearby city who did not receive any type of school-based sexual assault prevention programming.
A total of 984 girls were eligible for the study (intervention group = 599; comparison group = 385). Of these 984 girls, 255 consented and completed the 6-month follow-up survey (intervention group = 74; comparison group= 181) and were included in the study sample. Of the girls in the intervention group, 66.7 percent identified as heterosexual, and 33.3 percent identified as a sexual minority (i.e., identifying as bisexual, lesbian/gay, or another non-heterosexual identity). The average age of girls in the intervention group was 14 years. Of the girls in the comparison group, 80.1 percent of girls identified as heterosexual, and 19.9 percent identified as a sexual minority. The average age of girls in the comparison group was 13.7 years.
Propensity score weighting was used to minimize selection bias and aimed to statistically equate treatment groups. Using t-tests and chi-square analysis to establish covariates to use in the propensity score model (i.e., binge drinking, grades, experience of sexual harassment), propensity scores were calculated to use as weights in the causal analysis to correct for bias in the probability of being assigned to a treatment group. There were no statistically significant differences between girls in the intervention and comparison groups, based on baseline characteristics. Although not statistically significant, an attrition analysis found that girls in the intervention group who endorsed alcohol use at pretest (i.e., baseline) were less likely to participate in the posttest survey and girls in the comparison group who had higher grades and were older at pretest were less likely to participate in the posttest survey.
Survey data were collected at baseline and at the 6-month posttest. All survey items were responded to with a “yes” or “no” regarding whether the behavior had occurred during the past 6-months. Outcomes of interest included sexual assault, physical dating violence, and sexual harassment. Sexual assault was measured using four items, with three items taken from Cook–Craig and colleagues’ (2014) sexual violence instrument, which measured:
- Sexual coercion (“Another student had sexual activities with you although you did not really want to because… the student threatened to end your friendship or romantic relationship if you didn’t”),
- Physically forced sex (“Another student forced you to do sexual things that you did not want to do, such as kissing, touching, or being physically forced to have sexual intercourse?”), and
- Incapacitated sex (“You had sexual activities when you did not want to because you were drunk or on drugs?”).
One-item from the Youth Risk Behavior Survey (Kann et al. 2014) was used to measure sexual dating violence (i.e., “Someone you were dating or going out with forced you to do sexual things that you did not want to do, including things such as kissing, touching, or being physically forced to have sexual intercourse?”). Physical dating violence was measured using one item from the Youth Risk Behavior Survey (i.e., “Someone you were dating or going out with physically hurt you on purpose, such as being hit, slammed into something, or injured with an object or weapon?”). Finally, sexual harassment was measured using two items from the American Association of University Women’s Survey on Sexual Harassment (Lipson 2001), with questions related to unwanted sexual comments (i.e., “Another student made sexual comments, jokes, gestures, or looks about you?”) and sexual rumors (i.e., “Another student spread sexual rumors about you?”).
Poisson regression and logistic regression models were used to estimate the intent-to-treat effect to determine differences between girls in the intervention and comparison groups, at the 6-month follow-up. The study authors did not conduct subgroup analyses.