A Multisample Item Response Theory Analysis of the Beck Depression Inventory-1A

Article excerpt


The widespread employment of the Beck Depression Inventory-1A (BDI-1A) has spawned a number of practices: 1) The employment of an unweighted total score as a measure of depression; 2) Its use in populations other than that in which it was normed; and 3) The employment of BDI-1A total scores in hypothesis tests about population differences in mean depression. A sequential procedure based on item response theory was employed to assess the validity of these practices for the case of four populations: clinical depressives, mixed nondepressed psychiatric patients, and students from two different universities. The findings suggested that the first practice was not justified for any of these populations, that the BDI-1A was employable only with clinical depressives and with one of the university populations, and that mean comparisons were not allowable.


L'emploi largement repandu de l'inventaire de depression de Beck - Version 1A (IDB-1A) a engendre de nombreuses pratiques: 1) l'utilisation des scores obtenus comme mesures de la depression; 2) le recours A l'IDB aupres de populations differentes de celles pour qui ce test a ete concu ; et 3) l'application des scores totaux obtenus A l'IDB-1A a des tests d'hypotheses portant sur les differences de l'indice moyen de depression dans une population. On a evalue la validite de telles pratiques au moyen d'une procedure sequentielle basee sur la theorie de la reponse d'item aupres de quatre populations differentes : des personnes souffrant de depression clinique, des patients en etablissement psychiatrique souffrant ou non de depression, et des etudiants de deux universites distinctes. Les conclusions de l'etude revelent que le recours a la premiere pratique n'etait justifie aupres d'aucune des populations, que l'usage de Pros-l'IDB-1A n'etait approprie qu'aupres de la population composee de personnes souffrant de depression clinique et que les comparaisons des indices moyens de depression n'etaient pas acceptables.

The Beck Depression Inventory (BDI; Beck, Ward, Mendelson, Mock, & Erlbaum, 1961) is a self-report inventory developed to assess the "depth" of depression of individuals already diagnosed as depressed (Beck, 1972, p.187). Each of the twenty-one items corresponds to a particular, putative symptom of depression, and is paired with a 4-point Likert response scale. In 1979, Beck, Rush, Shaw, and Emery published a modified version of the original Beck in which alternative wordings for the same symptoms, and double negatives, were eliminated. This revised version, the BDI-1A, has become one of the most popular instruments for the assessment of depression (Piotrowski & Keller, 1992).

The widespread employment of the BDI-1A in both research and applied settings has spawned a number of related practices, among these: 1) The employment of a simple, unweighted composite of BDI-1A items (i.e., total score) as a measure of depression; 2) The employment of the BDI-1A in populations other than that in which it was normed; 3) The use of the BDI-lA total score as input into tests of hypotheses about population mean differences in level of depression. Each of these practices is justified by specific kinds of psychometric evidence. For while the scoring of a test as a single composite of the items, and, in particular, an unweighted linear composite, is, in certain quarters, virtually the default, this practice is justified only given that, within the population of interest, the test items behave psychometrically in a particular way. Usually, the requirement is that the items manifest a particular brand of unidimensionality. With respect to the second practice, the BDI-1A was normed in a population comprised of clinically depressed in- and outpatients. Assuming that it had been shown to perform satisfactorily within this norming population, its employment in any other population, say population B, would then be justified by evidence of its satisfactory performance in this population. …