Spoth and colleagues (2002) used a two-group design with random assignment of 22 rural schools in a Midwestern state which were assigned to intervention and minimal-contact control groups. School selection was based on program eligibility and community population of fewer than 8,500. Schools were blocked on size and proportion of lower-income families and then randomly assigned to groups.
The intervention group families (n=238) received the seven-session Strengthening Families Program (SFP). There were 208 control-group families.
The families were representative of the region and had an average of 3.1 children; 86 percent were two-parent families. Nearly all parents (98 percent of mothers and 95 percent of fathers) completed high school, and about half reported some post–high school education. Average ages of study parents were 37.2 years for mothers and 39.4 years for fathers. Nearly all parents (97.8 percent) were white. Of the 238 families admitted to the study, 79 percent completed the posttest (SFP=189; control=186) and 68 percent completed the 1-year follow-up (SFP=161; control=156).
Two major outcome measures were used. The Young Adolescent Substance Refusal and Substance Resistance measure consisted of four young-adolescent self-reported items and three parent-reported items that measured the likelihood that the young adolescent would refuse a peer alcohol offer, along with items that measured the youth’s response to a specific peer pressure situation. The Intervention-Targeted Parenting Behaviors, a measure of parenting behaviors directly targeted by the SFP program, included 13 parent self-reported items that assessed application of substance-related rules and consequences, positive involvement with the target child, communication, and anger or conflict management. In the case of dual-parent families, mother and father scores were averaged.
The study also included an implementation adherence measure completed by observers three or four times for each team of program implementers. For each of the three program components (youth, parent, family), 25 percent to 50 percent of the total number of observations were checked by a second observer.
Interrater reliability on the adherence measure was calculated, and the intervention group was divided into higher- and lower-adherence groups. Multilevel (mixed model) analyses of covariance were used to account for the nested design of the study (families clustered within schools) and to test for intervention outcome effects in the total sample as well as with the higher- and lower-adherence subgroups.
Gottfredson and colleagues (2006) examined intervention fidelity and effectiveness of the SFP using a four-group experimental design with a sample of 715 predominately African American families in an urban setting. Families were recruited at local events by five organizations in the Washington, D.C., area. One site was a prerelease center for incarcerated parents; the other agencies recruited from the community. Some 1,400 families from high-risk neighborhoods were approached, and 715 (51 percent) were enrolled. Eligible families had a child between the ages of 7 and 11 years and were able to read, speak, and understand English. Participating parents were predominately African American (75 percent) and female (94 percent), and more than half (52 percent) reported a combined family annual income of less than $20,000.
Eligible families were randomly assigned to one of four conditions: child skills training (CT) only (n=176), parent skills training (PT) only (n=177), parent and child skills training plus family skills training (FT; n=188), and minimal treatment (MT) control group (n=174).The FT treatment group received the fourteen 3-hour sessions of the SFP that includes a) 1 hour of preclass activities, b) 1 hour of parents’ skills training and a simultaneous children’s skill training class, and c) 1 hour of family skills training conducted in two multifamily groups, each lead by two group leaders. The other two intervention groups received either the parent training or the child training component of the SFP but not all three components. Pretests and posttests were administered in small groups of parents and children before interventions began and again approximately 1 week after the intervention ended. The overall attrition rate from pretest to posttest was 30 percent, a rate similar to other studies of the SFP.
Three different outcome surveys were derived from previous tests constructed from standardized scales before this study: a 56-item survey for younger children ages 7 and 8; a 138-item survey for older children (ages 9 through 11), and a 195-item survey for parents. Content of the surveys, which included both parent and child reports, consisted of child problem behaviors, child risk and protective factors, and family factors.
The parent- and child-reported outcomes were analyzed with analysis of covariance that included a dummy variable to account for the various sites and pretest scores as covariates in each analysis.