Predicting preschool children's eating in the absence of hunger from maternal pressure to eat: A longitudinal study of low-income, Latina mothers

Abstract

Early work by Klesges et al. (1983, 1986) suggested that mothers who frequently prompt their children to eat have children at greater risk for obesity. This is consistent with the hypothesis that controlling feeding practices override children's responsiveness to their internal fullness cues, increasing the risk of overeating and obesity (e.g., Johnson & Birch, 1994). Subsequent cross-sectional research on pressure to eat, however, has been inconsistent. Most studies have shown that maternal self-reports of pressure to eat are negatively associated with childhood obesity, and observational studies showed inconsistent relationships with child weight status. In the present study we examined the association between low-income, Latina mothers' pressure to eat and their preschool children's eating in the absence of hunger using both self-report and observational measures of feeding practices. A longitudinal design examined eating in the absence of hunger over 18 months; children's BMI at the initial timepoint was statistically controlled to address the tendency of mothers of underweight children to pressure their children to eat. At each timepoint, mothers completed the Child Feeding Questionnaire (Birch et al., 2001) and were observed feeding their child a meal in a laboratory setting. Eating in the absence of hunger (Fisher & Birch, 1999) was assessed at both timepoints as well. A cross-lagged panel model showed that observed maternal prompts to eat a different food at time one predicted kcal consumed in the absence of hunger at time two (controlling for kcal consumed in the absence of hunger at first timepoint: beta = 0.20, p < 0.05). Results suggest that pressure to eat alone may not be what contributes to eating in the absence of hunger, but that the nature of that pressure may be more important.

Keywords: Eating in the absence of hunger; Hispanic preschoolers; Maternal feeding practices.

Figure 1

Cross-Lagged Panel Analysis: Statistically Significant Paths (p < .05) Between Study Variables (All Time 1 Variables Were Allowed to Correlate with One Another)