Results from a Cluster Analysis of the Spanish Adaptation of the Strenghtening Families Program (SFP) in a 24 Months Longitudinal Study

The paper presents the main outcomes of the longitudinal analysis of the Spanish adaptation of the Strengthening Families Program (Kumpfer & DeMarsh, 1985). The study overcomes the model process-result, which considers all participants equal and able to be compared directly. The analysis proposes the model situation-context-process-result, which takes into account the applied programs (process), the context and the initial levels of the participants.

This analysis aims to study the influencing factors of the improved medium-term results of families who participated in the Family Competence Program (FCP). It also pretends to ascertain the extent to which the analyses of results are conditioned by the characteristics of the families and the longitudinal design in itself.

The study analyses the results of the implementation of the FCP at seven substance abuse treatment centers and ten Primary Care Social Services centers in Spain. The implementations started in 2009-2011 and completed two years afterwards (2012-2013). 

The analysis investigates the inclusion of the initial level in the factor-scales. Participants’ initial level in the factor-scales enables the control of the unwanted effects of the statistical regression in the classification of families.  

The added-value model focused on longitudinal data set some methodological challenges, both statistical and psychometric. In order to overcome them, we have studied the differential patterns of the latent scores and observed them throughout time, as well as the effect they had on two factors: type of family and maintenance patterns of the differential scores in the children’s, parents’ and family’s levels.

Different statistical models have been applied for the analysis to a sample of 155 families: cluster analysis, multiple regressions and analysis of time series.

The analysis is focused on three key factors: child-parent relationships; family cohesion and positive parenting. They are measured with validated instruments for the Spanish population.

The cluster analysis distinguishes three types of families, according to the process

• 1^st cluster: Families that have undergone a constant improvement
• 2^nd cluster: Families that have kept the achieved levels at the end of their participation in the FCP
• 3^rd cluster: Families that have lost part of the observed improvements

Regarding the longitudinal assessment of positive parenting, two main conclusions can be settled up: the relevance of family type to explain the effects of a preventive program’s and the need to understand correlation patterns among latent variables in order to provide plausible explanations for the empirical results found.