Parental food selection questionnaire - Infant version.

Purpose: Complementary feeding practices (CFPs) are associated with health outcomes (e.g., obesity and food allergies). Understanding how parents select foods for their infant is limited. This study's purpose was to develop a psychometrically sound measure of parents' food selection motives for their infant during the complementary feeding period.

Methods: Development and testing of the Parental Food Slection Questionnaire-Infant Version (PFSQ-I) occurred in three phases. English-speaking, U.S. mothers of healthy infants, aged 6-19 months old participated in a semi-structured, face-to-face interview (Phase 1) or a web-based survey (Phases 2 & 3). Phase 1 was a qualitative study of maternal beliefs and motives surrounding complementary feeding. Phase 2 involved adaptation and exploratory factor analysis of the original Food Choice Questionnaire (Steptoe et al., 1995). Phase 3 involved validity testing of the relationships among PFSQ-I factors and CFPs (timing/type of complementary food introduction, frequency of feeding method, usual texture intake, and allergenic food introduction) using bivariate analyses, and multiple linear and logistic regression analyses.

Results: Mean maternal age was 30.4 years and infant age was 14.1 months (n = 381). The final structure of the PFSQ-I included 30 items and 7 factors: Behavioral Influence, Health Promotion, Ingredients, Affordability, Sensory Appeal, Convenience, and Perceived Threats (Cronbach's α = 0.68-0.83). Associations of factors with CFPs supported construct validity.

Discussion: The PFSQ-I demonstrated strong initial psychometric properties in a sample of mothers from the U.S. Mothers who rated Behavioral Influence as more important were more likely to report suboptimal CFPs (e.g., earlier than recommended complementary food introduction, delayed allergenic food introduction, and prolonged use of spoon-feeding). Additional psychometric testing in a larger, more heterogenous sample is needed, along with examination of relationships between PFSQ-I factors and health outcomes.