Multidimensionality in adolescent eating problems. A two-phase measurement study.

The aims of this study were, firstly, to examine the factor structure of single items from several self-report methods used to measure eating disorder symptoms, and secondly, in search for a short assessment instrument to reduce the number of items within each factor. Factor analyses were employed to identify and confirm the constructs measured by a total of sixty items drawn from these assessment methods. In phase one, 508 secondary school pupils were recruited as subjects, and principal factor analysis identified three factors; body concern, dieting, and loss of control over eating. Twenty-nine items were retained, and in phase two subjected to a second data collection among 4129 secondary school pupils. Confirmatory factor analysis was employed to cross-validate the factor structure from phase one. The three-factor structure was confirmed satisfactorily for boys, but not for girls aged 12-14 years. For girls aged 14-16 years, it was confirmed when a number of items were omitted. The findings are discussed in relation to restraint theory and the continuum hypothesis.