Confirmatory factor analysis of the infant feeding styles questionnaire in infant and toddler child care teachers.

Early childhood is a critical time of development. Caregiver feeding style, including of child care providers, has been implicated in influencing children's food preferences and obesity. A better understanding of the determinants of child care providers' feeding styles is important as children increasingly receive daily care and nutrition in child care centers. Valid instruments to measure provider feeding styles specifically for infants are needed. This study's objective was to validate, for use with child care providers, the Infant Feeding Styles Questionnaire (IFSQ), an instrument originally designed for parents of infants and toddlers. Study data came from the Baby Nutrition and Physical Activity Self-Assessment for Child Care (Baby NAP SACC) study, a cluster randomized trial, designed to evaluate the efficacy of a six-month-long obesity prevention intervention in child care centers. The IFSQ was administered at baseline and follow-up to infant and toddler teacher participants (n = 244). Scores for baseline IFSQ teacher responses (n = 219) were calculated for the 13 sub-constructs across the five feeding styles. We tested reliability with Cronbach's alpha coefficients and performed confirmatory factor analysis to examine factor loadings and goodness of fit characteristics, modifying constructs to achieve best fit. Reliability coefficients ranged from 0.42 to 0.82 for all sub-constructs. Goodness of fit indices for the final model ranged from: root mean square error of approximation (RMSEA) 0.000-0.082; comparative fit index (CFI) 0.943-1.000; Tucker-Lewis index (TLI) 0.906-1.000; and standardized root mean squared error (SRMR) 0.006-0.130. In a sample of racially diverse infant and toddler child care teachers, final models of the laissez-faire, pressuring, and responsive constructs and the restrictive-diet quality sub-construct performed well. Overall model fit was poor for the indulgence construct, but good for solely its corresponding belief items.