Validity of the Short Mood and Feelings Questionnaire in late adolescence

Studies examining the validity of the Short Mood and Feelings Questionnaire (SMFQ; Angold, Costello, & Messer, 1995) have largely focused on selected or clinical samples in childhood (6-11 years) or early to midadolescence (12-16 years) and have not investigated misclassifications or how the SMFQ relates to adult depression measures. Using data from the Avon Longitudinal Study of Parents and Children (2012), we assessed the validity of the SMFQ in relation to an adult depression measure administered in late adolescence (age 17-18 years). We also investigated sociodemographic and clinical variables previously shown to affect misclassification on short self-administered questionnaires compared with more detailed assessments of depression. We assessed construct validity using factor and item response theory analysis. To investigate content validity, we tabulated SMFQ items against the International Classification of Diseases (ICD-10; World Health Organization, 1992) and Diagnostic and Statistical Manual of Mental Disorders (4th ed.; American Psychiatric Association, 1994) depressive symptoms. Criterion validity was examined using receiver operating characteristic (ROC) analysis. Potential misclassifications were investigated using logistic regression and multiple-indicator multiple-cause modeling. Factor analysis produced high loadings, low residual variances, and appropriate model fit indices. Seven of the 10 ICD-10 depressive symptoms were covered by at least 1 SMFQ item. The discriminatory ability of the SMFQ for meeting ICD-10 diagnostic criteria for depression was very high (area under ROC curve = 0.90). Individuals with anxiety symptoms, females, and less well-educated individuals overreported depressive symptoms on the SMFQ in relation to ICD-10 depression. We conclude the SMFQ is a valid instrument capturing a latent trait of depression in a community-based sample in late adolescence. Further work should be carried out to increase understanding of variables associated with misclassification.