Study
Murry and colleagues (2014) evaluated SAAF using a randomized trial. Nine counties in Georgia were randomly selected to obtain lists of 11-year-old students. Students were then randomly selected from those lists for recruitment to the study. A total of 671 families were randomly assigned to the prevention condition (n = 371) or the minimum control condition (n = 299). Of the 671 families, 620 remained in the sample and completed a posttest. The 65-month follow-up (almost 5½ years) included 326 families from the prevention group and 245 from the control group. Retention was maintained by calls to families on a regular basis from community liaisons. Additionally, families provided contact information and received a semiannual newsletter.
Structural Equation Modeling (SEM) was used to test hypotheses of the study. During each data collection wave, one home visit was conducted for each family. Home visits consisted of self-report questionnaires in interview format. Interviews were private and administered by a data collector to address any literacy complications.
Interviews were conducted separately and privately. Four waves of data were collected, which included pretest, 3-month posttest; 29-month, long-term follow up; and 65-month, long-term follow up.
Study
Brody and colleagues (2006) evaluated Strong African American Families (SAAF) using an experimental design in which eight rural Georgia counties were randomly assigned to receive treatment (SAAF programming) or a control condition (brochures about parenting). Families receiving programming attended seven consecutive weekly classes at local community centers. Classes consisted of family instruction and separate classes for parents and children. Participants were chosen from county schools’ enrollment lists. Eleven-year-old students were randomly selected from these lists; their families were invited to participate in the study.
Out of 521 randomly selected families, 322 participated and completed pretests. The treatment group, 172 families, received SAAF programming. The control group, 150 families, received three informational mailings about parenting techniques. Before treatment, there were no significant differences between the treatment group and the control group. In 53.6 percent of the families the target child was a girl. Of the mothers in families, slightly more than a third were single, another third were married but separated, 23 percent were married and living with their husbands, and 7 percent were living with partners (not married). The median household income was $1,655 per month reflecting the low-income families the intervention targets.
Researchers collected observational data through home visits and used parent and child self-reports of behavior. They collected data from families 1 month before the SAAF intervention, 3 months after the classes had ended, and again 2 years following the end of classes. There was seven months between pretest and posttest. First, Hierarchical Linear Modeling was used to compare group equivalence. Next, Structural Equation Modeling was used for hypotheses testing. Specifically, intervention-targeted behaviors such as parenting style and enhancement of youth protective factors were tested to determine whether the SAAF programming had any direct effects. These analyses were replicated using multilevel analysis of covariance.
