Pilot phase. Recently, the PDFY curriculum has been tested experimentally with families in rural Iowa. This study is part of Project Family, a series of studies conducted at Iowa State University, and for which Richard Spoth is the principal investigator, in collaboration with the Social Development Research Group at the University of Washington. In the pilot phase of this project (Spoth et al., 1995), all families with sixth and seventh grade children in nine different schools were called and invited to participate (387 families were invited). The schools had been selected from districts meeting eligibility requirements for the federally supported school lunch program. Although not all families were eligible for school lunch benefits, the median annual per capita income was $6,800 ($27,200 for a family of four). A total of 209 families completed the pretest, and 175 (84 percent) of these families completed the final posttest assessment. Each family was offered a financial incentive of approximately $10 per hour per family member for time devoted to study assessments. No monetary incentives were provided for attending the PDFY sessions. Virtually all participants were Caucasian.

Attendance records indicate that most parents assigned to the intervention group attended most of the PDFY sessions; 88 percent of enrolled mothers and 69 percent of enrolled fathers attended three or more sessions; nearly one-half of the mothers (47 percent) and one-third of the fathers (32 percent) attended all five sessions. The mean attendance was 3.9 sessions for enrolled mothers and 3.1 sessions for enrolled fathers.

Figure 4: PDFY and Control Group Improvements in Parenting Outcome Means, by Experimental Condition

Source: Spoth et al., 1995. Note: Significance levels for all analysis of covariance (ANCOVA) tests were significant at p<0.5.

In other analyses examining these data, outcomes targeted by the specific intervention sessions were tested separately, using both the self-reported and videotaped assessments (Kosterman et al., 1997, 1998). Specifically, results indicated that mothers in the PDFY group were significantly more likely to:

• Report that they give rewards to their children for good behavior.

• Communicate rules regarding substance use.

• Punish their children appropriately for misbehavior.

• Restrict their children's alcohol use.

• Expect their children to refuse beer from a friend.

• Express less conflict with their spouses.

• Work at being more involved with their children.

Fathers in the intervention group also reported significantly more communication of rules regarding substance use to their children and more involvement with their children. Observations yielded results that were consistent with the self-reported findings. Mothers in the intervention group exhibited significantly more proactive communication, less conflict, and more improvement in relationship quality or bonding with their children than mothers in the control group (although the latter finding was significant at only p<.06). Fathers in the intervention group also exhibited significantly more proactive communication and better relationship quality or bonding than fathers in the control group. All of these outcomes were goals of specific sessions in the PDFY curriculum.

As a further check on the validity of the findings from the Iowa PDFY study, additional analyses examined:

• Effects of the intervention on outcomes that were superficially similar to targeted measures but that were not targeted by PDFY.

• Increments in improvement when the intervention group was restricted to only those who attended specific PDFY sessions (Kosterman et al., 1998).

These analyses were conducted to demonstrate that specific PDFY objectives were linked with specific outcomes and that these effects were not due to more global causes such as parents merely being aware that they were part of an intervention or desiring to impress the researchers with the "right answer." Indeed, no significant differences between intervention and control groups were found among the six nontargeted parenting behaviors examined in the study but not addressed in PDFY workshops. For example, while mothers in the PDFY group reported being more likely to reward their children (as instructed in PDFY), they were not significantly more likely to receive rewards from their children or to reward or receive rewards from their spouses (not instructed in PDFY). While both mothers and fathers assigned to PDFY reported more involvement with their children (included in PDFY), they did not report more involvement with each other (not included in PDFY). Along similar lines, there was evidence that the subgroup of parents who actually attended specific workshops showed greater improvement on reported parenting behaviors emphasized in those workshops than the entire experimental group, which included nonattendees. Compared with the entire experimental group, the attendees demonstrated more improvement for 19 of 28 (68 percent) targeted behaviors, but only 4 of 12 (33 percent) nontargeted behaviors. These findings help to further link improvement in specific behaviors to attendance at specific PDFY sessions.

Trial phase. Additional results are available from an experimental study with followup assessments also involving rural Iowa families. Most of the critical features of the second experiment were identical to the pilot study, except for the inclusion of a larger sample, a school-based random assignment to condition (i.e., all students at a school were randomly assigned to the same condition), and a longer followup period (1 and 2 years posttreatment). In all, 360 sixth grade students and their parents completed both pretest and posttest measures. Among enrolled families, 93 percent attended three or more sessions, and 63 percent attended all five sessions. Initial findings replicate those of the pilot study. Among parents assigned to the PDFY curriculum, intervention-targeted parenting behaviors showed significant improvement for both mothers and fathers, consistent with PDFY objectives (no differences between schools were found on these measures within the PDFY and control conditions). More recent findings on long-term followup indicate sustained improvements in parent behaviors and subsequent reductions in children's smoking and drinking behaviors (Spoth, Redmond, and Shin, 1998; Spoth et al., 1997, 1998).