Use of an Observational Comparative Strategy Demonstrated Construct Validity of a Measure to Assess Adherence to the Satter Division of Responsibility in Feeding

Abstract

BackgroundCurrently developed parent feeding measures do not exclusively measure behaviors compatible with theoretical underpinnings of the Satter Division of Responsibility in Feeding for Children Aged 2 to 6 Years (sDOR.2-6y). A measure of adherence to sDOR.2-6y has been constructed and shown to have translational validity. ObjectiveTo examine the construct validity of the 15 sDOR.2-6y items. DesignObservations from in-home mealtime video-capture in a household with a child aged 2 to 6 years were compared with parent sDOR.2-6y responses. One sDOR.2-6y item about mealtime regularity was compared with mealtime data provided in the Automated Self-Administered 24-Hour Dietary Assessment Tool that had been completed by a larger sample of parents that included the smaller sample of video-capture participants. Participants/settingParents of preschool children aged 2 to 6 years in central Pennsylvania participated in Automated Self-Administered 24-Hour Dietary Assessment Tool online dietary assessment (n = 61) from May to November 2013 or in-home mealtime video-capture (n = 20) during October 2013 or both (n = 16). Main outcome measuresConsensus of three trained sDOR coders of plausible parent sDOR.2-6y item responses after video viewing, parent responses to sDOR.2-6y items, and mealtime regularity from three Automated Self-Administered 24-Hour Dietary Assessment Tool recalls. Statistical analysisData analysis included identification of codeable 5-minute segments, coding by three sDOR experts, intraclass correlation coefficient, and determination of coder congruence to establish plausible sDOR.2-6y responses and comparison between coder-observed and parent sDOR.2-6y responses. ResultsVideo-capture participants were mostly women, White, and overweight with a mean ± standard deviation age of 35.3 ± 6.4 years. Postvideo debriefing supported the observed meal as typical in location, timing, and composition. Of 273 coding decisions, coder congruence was inadequate or dissonant for 3.7% and intraclass correlation coefficient was 0.81 (0.77 to 0.84; P < 0.001). Parent and congruent coder responses were in agreement for nearly 70% of the possible comparisons. Of the original 15 items, three were deemed not valid, nine of the remaining 12 were considered valid, and three could not be invalidated. ConclusionsCongruence between observed and self-reported behaviors supported additional validation and scoring studies with the 12-item sDOR.2-6y to assess adherence of parents to sDOR.