Dimensions of Implementation As Predictors of Effects of New Beginnings Program

Introduction: A full theoretical test of a program addresses not only did the program work, but how did it work? Careful and systematic study of the implementation of the program demonstrates the extent to which processes outlined in the program’s action theory led to hypothesized changes.

Method: Using data from the 470 parents assigned to the intervention condition of the New Beginnings Program effectiveness trial for divorcing parents, we test predictions based on the multidimensional model of implementation (Berkel, Mauricio, Schoenfelder, & Sandler, 2011). The family courts in Arizona serve a highly diverse demographic population of divorcing parents. In contrast to tightly controlled efficacy trials, this produced sufficient variability in parent characteristics (e.g., ethnicity, gender of parent) to test how different aspects of implementation predict outcomes for different sub-groups of the sample.

Results: Preliminary analyses demonstrate that in general home practice of program skills led to improvements in parenting and that this accounted for the relation between attendance and parenting outcomes targeted by NBP (i.e., acceptance, rejection, communication, and discipline). However, these effects were not uniform across outcomes and subpopulations (i.e., fathers vs. mothers and Latinos vs. non-Latino White parents). For example, in some instances, attendance was not associated with home practice. We hypothesize that the fidelity and quality with which providers delivered the program would influence participants’ likelihood of attending and practicing assigned skills with their children each week. Moreover, these processes may differ by gender and ethnicity. We collected observational ratings of fidelity and quality for activities across all 10 sessions of the NBP across 50 groups. Using mediational and moderational analyses in Mplus, we will assess how fidelity and quality predict attendance and home practice, and in turn, improvements in parenting outcomes from pre- to posttest. We hypothesize the following effects:

1. fidelity will be associated with home practice, but not attendance
2. quality will be associated with attendance
3. quality will moderate the association between fidelity and home practice
4. attendance and home practice mediate the associations between fidelity and outcomes and quality and outcomes

We will also examine these models for four subpopulations: mothers, fathers, Latinos, and non-Latino White parents.

Conclusions: These analyses will inform how these processes work across cultures and consequently how preventive interventions can be optimized to make strides against health and social disparities.