Study
Cissner and Ayoub (2014) used a cluster randomized control design to evaluate the effectiveness of the Fourth R Curriculum for seventh-grade students across 10 schools in the Bronx, New York. The Fourth R group received the intervention in sex-segregated classes, whereas the control group received a standard seventh-grade curriculum in mixed-sex classes.
Individual-level data was collected at the baseline survey. The Fourth R intervention group (n = 263) was 51 percent male, 75 percent Hispanic, 27 percent Black, 3 percent white, 2 percent American Indian, and 7 percent other. The control sample (n = 248) was 38 percent male, 70 percent Hispanic, 33 percent Black, 4 percent white, 2 percent American Indian, 1 percent Asian, and 10 percent other. The average age in both samples was 12 years. There were more females in the control group, compared with the Fourth R group, but there was no statistically significant difference between the intervention and control groups in previous exposure to program materials or content.
Regression analyses were used to examine the effects of the Fourth R Curriculum. The outcomes included sexual harassment or assault perpetration, sexual harassment or assault victimization, peer violence or bullying perpetration, peer violence or bullying victimization, sexual activity, drug and alcohol use, and prosocial attitudes. The researchers conducted subgroup analyses on gender, student level of risk, and program fidelity (i.e., how much of the curriculum students received).
Study
Crooks and colleagues (2011) conducted a cluster randomized controlled trial to evaluate the effects of the Fourth R Curriculum on violent delinquency 2.5 years after the initial intervention. The study was a follow up to the 2009 study by Wolfe and colleagues (see Study 1) and used the same sample. Violent delinquency was measured using eight items from a self-report inventory developed by the National Longitudinal Survey of Children and Youth (Human Resources Development Canada 2001). From the original sample of 1,722 ninth-grade students, 1,520 (655 control, 865 intervention) completed follow-up measures as 11th-grade students.
The researchers used two-level hierarchical models to estimate the effect of individual and school-level variance on acts of violent delinquency 2.5 years after the initial intervention. The authors conducted subgroup analyses on gender, delinquency at the time of the first study, and previous maltreatment.
Study
Wolfe and colleagues (2009) conducted a cluster randomized controlled trial to evaluate the Fourth R Curriculum. Thirty rural and urban high schools in southwestern Ontario, Canada, were recruited for the study, based on general school populations and the agreement of principals to conduct randomization, teacher training, delivery of the assigned intervention, and evaluation, and to restrict similar programs during delivery. The participants were all students enrolled in the ninth-grade health and physical education curriculum within those schools. Schools were randomly assigned to the intervention or control groups.
Students in the experimental schools received the 21-lesson intervention curriculum, Fourth R: Skills for Youth Relationships. The program was implemented for 1 year prior to the evaluation. Students in the comparison schools received a standard health and physical education curriculum in sex-segregated classrooms. Teachers in the comparison schools were required to cover the topics of the three units being taught in the intervention schools, but without any background or training on these topics or access to a structured curriculum.
Twenty schools, each with a large student body, were split equally between urban and rural locations and had comparably experienced teachers. A total of 1,722 students participated in the study. Participants were predominantly white (88 percent), although some students identified as Asian (4 percent), Arabic (2 percent), Hispanic/Latino (0.6 percent), or other (4 percent). The intervention group (n = 968) was 51 percent female, and the control group (n = 754) was 55 percent female. Both groups had similar levels of risk behaviors, with 1 percent of participants in both groups showing physical dating violence in the past year. There were more girls in the control schools, compared with the intervention schools. There were no other statistically significant differences between the groups at baseline.
A confidential, online survey was given to students in school at baseline and 2.5 years later, at follow up. Peer dating violence was assessed using eight items from the Conflict in Adolescent Dating Relationships Inventory, a self-report measure. Students answered “yes” or “no” to questions regarding their actions during fights or arguments with their significant others within the past 12 months. Although respondents who did not have any dating partners were included in the study, they were given a score of 0 at baseline. Physical peer violence was self-reported, using the National Longitudinal Survey of Children and Youth delinquent behavior inventory. Students were counted if, in the last 3 months, they had injured someone as a result of a fight; had used an object other than their hands in a fight; or had hit, slapped, or physically hurt another teen with the intention of scaring or humiliating them. Substance use was assessed using the National Longitudinal Survey of Children and Youth guidelines, and classified as substance abuse if one of the following four criteria was met: 1) drank alcohol 1 or 2 days a week or more, 2) had five or more alcoholic drinks at one time in the past 30 days, 3) used marijuana 1 to 2 days a week or more, or 4) had tried any other illicit drugs in the past 3 months.
Categorical data for outcomes were analyzed using two-level hierarchical models to account for clustering of students in schools. Separate, subgroup analyses were conducted using the subsamples of students who were dating the year before follow-up. The researchers also conducted subgroup analyses on gender.
