[Parental assessment of children's eating behaviors with a short form of the Anorectic Behavior Observation Scale].

OBJECTIVE

To determine whether the Anorectic Behavior Observation Scale (ABOS) originally developed for clinical purposes can be used in epidemiological studies to assess parents' rating of their child's eating behavior.

METHOD: In a study of 1,057 children aged 10-17 years, the children completed the Eating Disorder Inventory, Child Version (EDI-C), while the parents of 922 children completed the ABOS. The ABOS consists of 30 items, divided into three subscales--unusual eating behaviors, bulimia-like behaviors, and hyperactivity.

RESULTS

The data revealed a number of psychometric problems of the ABOS, including variance restriction and extreme distributions. This is due to the checklist design of the ABOS, which generates few "yes" responses in nonclinical populations. Using principal components analysis and confirmatory factor analysis, 10 items were selected that had better psychometric characteristics than the 30-item ABOS while maintaining the same level of validity. This resulted in three subscales similar to the original ones, with 4, 2, and 4 items--eating behavior, bulimia-like behavior, and level of activity/fitness.

CONCLUSIONS

The short form of the ABOS, called Eating and Activity Questionnaire for Parents (EAQP), is a parsimonious instrument for screening parents' assessments of their child's eating behavior.