Abstract
Clinical features have low predictive value for the outcome of eating disorders. This study focuses on the possible predictive value of gender ideals. Data was from adolescent girls thoroughly assessed both at presentation and at a three-year follow up. Few variables differed between the outcome groups. However, perceptions of gender ideals, as measured by the Three Questions of Ideals, were found to contribute to the prediction of outcome. A discriminant analysis showed that gender ideals, maturity fears, sexual debut, and vomiting were significant discriminators in a set of factors, which correctly classified the outcome for 75.3% of the sample.
The authors are grateful to Hans Arinell for statistical and data processing support. Thanks also are due to the participants and collaborators of the COEAT project. The study was supported by Stiftelsen Sven Jerrings Foundation, the Swedish Research Council, Professor Bror Gadelius' Memory Foundation, the Gillbergska Foundation, the First of May Flower Annual Campaign, HRH Crown Princess Lovisa's Fund for Child Health Care, the Swedish Society for Medical sciences, and Uppsala University.
Notes
Note: aEating Disorder Inventory for Children.
bYouth Self-Report.
cn.s. not significant p > .05.
asemi-structured follow-up interviews with adolescents (A), parents (P), of medical aspects (M).
bEating Disorder Inventory for Children.
cYouth Self-Report.
Note: asemi-structured follow-up interviews with adolescents (A); (EDI-C) Eating Disorder Inventory for Children; (TQI) Three Questions of Ideals; (YSR) Youth Self-Report.
bn.s. not significant p > .10.
Note: a(A) semi-structured follow-up interviews with adolescents; (EDI-C) Eating Disorder Inventory for Children; (TQI) Three Questions of Ideals; (YSR) Youth Self-Report.
bLeave-one-out cross validation was used. Each case classification was based on discriminant function generated by all other cases.