Observed and self-reported assessments of caregivers' feeding styles: Variable- and person-centered approaches for examining relationships with children's eating behaviors.

BACKGROUND: Demandingness and responsiveness are dimensions used in general parenting as well as the feeding literature to measure parent attitudes and behaviors. These dimensions can be treated as continuous variables (variable-centered) or used to form groups of parents based on scores on each dimension (person-centered). Research focusing on these dimensions in the feeding domain has relied exclusively on maternal self-reports; observational studies of feeding styles have yet to be conducted.

OBJECTIVES: The purpose of this study was to investigate the effectiveness of variable-centered versus person-centered analyses in examining relationships between maternal feeding styles, child eating behaviors, and child weight status using both self-reported and observational methods for assessing parental feeding style.

METHODS: Participants included 137 low-income Head Start mothers and their preschoolers. Mothers completed the Caregiver's Feeding Styles Questionnaire and the Children's Eating Behavior Questionnaire. Parent/child interactions were observed during dinner meals at home and were coded using a previously developed coding system.

RESULTS: Maternal demandingness during feeding was negatively associated with child BMI z-scores and positively associated with slowness in eating and satiety responsiveness. Maternal responsiveness was associated positively with enjoyment of food and associated negatively with food fussiness. Significant demandingness by responsiveness interactions showed that children of authoritarian mothers showed the highest levels of food fussiness and the lowest enjoyment of food. Overall, the findings were stronger for the self-report than for the observational measures and the variable-centered approach was clearly superior to the person-centered approach.

CONCLUSIONS: The current results indicate that for research purposes the variable-centered approach may be the most powerful for examining relationships between maternal feeding and child eating. However, for clinical purposes, the person-centered approach might be most informative.