A within-Person Mediation Model of a Skills Based Family Prevention Program

Family skill-based interventions often aim to improve parenting by teaching parents new skills. The assumed process of change is that parents will learn the skills while attending the intervention, practice them at home, and show better parenting.

We collected daily intensive longitudinal web surveys with 33 parents assigned to attend the Strengthening Families Program 10-14 over seven weeks of one-session-per-week intervention delivery (7 to 52 days of surveys per parent). Attendance was parent’s presence (1) or absence (0) for each weekly session. Skill practice: parents evaluated the degree to which they used each skill taught at that week’s session today (0 to 100, items averaged for scale). Items differed week to week to map onto that week’s content. Parenting was reported daily using an average of three items assessing intervention-targeted parenting skills. Each item tapped a subscale of the original instrument used in previous evaluations of SFP 10-14 (support, guidance, and anger management) and asked parents to assess whether they parented that way today (0 to 100).

Two- and three-level (day-week-person) unconditional means models were tested, and a two-level mediation model was used (day-person; Bolger & Laurenceau, 2013; Hayes, 2014). Within-person mediation illuminates the process of change over the days and weeks of the intervention: on days when parents attended that week, we expect that they practice more, and on days that parents practice more, we expect that their parenting is better. Significant mediation (ab) would indicate that the effect of attending the intervention on parenting occurs through increased use of intervention-taught parenting skills.

There was an effect of skills practice on parenting (b): on days when parents practiced more, parenting was better (g = 0.30*), and parents who practiced more reported better parenting (g = 0.73*). However, attendance in the intervention program did not predict practice (a): on days when parents had attended that week, practice was neither higher nor lower (g = -3.10 NS), and parents who attended more than average practiced neither more nor less (g = 2.50 NS). Thus, the mediation effect (ab) was not significant. There was no direct effect of attendance on parenting (c’). Follow-up analysis will be considered.

Examining interventions at this micro timescale illuminates the process by which parents and families change. Results of the model demonstrate that attendance at the intervention did not impact parents’ practice/incorporation of parenting skills at home in that week. Because practice was related to better parenting at the within-person level, these results suggest interventions could be more impactful if each session instigated day-to-day practice of the skills that are taught.