Factor Analysis of the Childcare Food and Activity Practices Questionnaire (CFAPQ) among Staff Members in Maryland Childcare Centers

Abstract

Objectives

Parental feeding behaviors influence the development of children's dietary behaviors. Less is known about the influence of feeding behaviors used by childcare staff. This study aimed to examine the psychometric properties of the feeding-related subscales of the Childcare Food and Activity Practices Questionnaire (CFAPQ), a scale developed and validated in the Netherlands, for use among staff members caring for 3–5 year olds in 46 childcare centers across Maryland.

Methods

A confirmatory factor analysis (CFA) was performed with childcare staff ( n = 136) to evaluate the goodness of fit of the original 7-factor, 40-item structure of the CFAPQ feeding-related subscales. Because the original factor structure was not confirmed, an exploratory factor analysis (EFA) was conducted to identify an alternative underlying factor structure. Pearson correlations between subscales were examined to identify if the factors related to one another in theoretically expected ways.

Results

The 7-factor, 40-item structure of the original CFAPQ was not confirmed by the CFA (Model fit: RMSEA = 0.14, CFI = 0.72, TLI = 0.69, and SRMR = 0.18). A revised 4-factor model with 22 items was confirmed (Model fit: RMSEA = 0.079, CFI = 0.96, TLI = 0.96, and SRMR = 0.09). The internal consistency of factors in the revised model was moderate to good (Cronbach's α = 0.63–0.81). The 4 factors include: 1) Responsive (8-items; staff encouragement surrounding healthy food intake); 2) Controlling (5-items; staff applying food related pressure or restriction); 3) Indulgent (5-items; staff using foods as reward or to regulate the child's emotional state); and 4) High engagement (4-items; staff involvement in children's snacks and meals). Correlations between factors indicated convergent and discriminant validity in theoretically expected directions. Indulgent was significantly negatively related to responsive ( r = −0.39, P < 0.01). Controlling was significantly positively related to high engagement ( r = 0.22, P < 0.01).

Conclusions

Among this sample of childcare staff, a 4-factor, 22-item structure was a better fit than the original 7-factor, 40-item structure of the CFAPQ feeding subscales. Robust psychometric properties ensure that scales are measuring what they claim.

Funding Sources

National Institute of Diabetes and Digestive and Kidney Diseases.