Counts, J. M., Buffington, E. S., Chang-Rios, K., Rasmussen, H. N., & Preacher, K. J. (2010). The development and validation of the protective factors survey: A self-report measure of protective factors against child maltreatment. Child Abuse & Neglect, 34(10), 762-772.
Number of participants: Exploratory Factor Analysis (EFA) and validation - N = 249. Confirmatory Factor Analysis (CFA) - N = 689
Summary:
The exploratory factor analysis (EFA) included 45 items theoretically serving as indicators of 4 factors: Family Functioning (FF), Emotional Support (ES), Concrete Support (CS), and Nurturing and Attachment (NA). The most interpretable factor structure emerging from the EFA retained 27 items. Based on a combination of standard factor retention criteria, model fit, and interpretability, a 4-factor EFA solution was chosen as the most appropriate model for the retained items. Additional items were removed due to low loadings, nontrivial cross-loadings, and parsimony, yielding a final scale with 20 items. Coefficient alphas for three subscales were acceptable for FF = 0.94, ES = 0.86, and NA = 0.83. The coefficient alpha for CS (0.63) was below the acceptable range of 0.80. In the validity analyses, all 4 subscales of the PFS were significantly negatively correlated with child abuse potential and stress. In the confirmatory factor analysis (CFA), factor loadings using both maximum likelihood and weighted least square means and variance (WLSMV) solutions supported those of the initial EFA sample. Factor correlations in the CFA also remained consistent with those from the original sample. Results demonstrate that the factor structure generalized well to a new sample.