Item response theory applied to the Beck Depression Inventory.

The Beck Depression Inventory (BDI), a scale that measures the latent trait intensity of depression symptoms, can be assessed by the Item Response Theory (IRT). This study used the Graded-Response model (GRM) to assess the intensity of depressive symptoms in 4,025 individuals who responded to the BDI, in order to efficiently use the information available on different aspects enabled by the use of this methodology. The fit of this model was done in PARSCALE software. We identified 13 items of the BDI in which at least one response category was not more likely than others to be chosen, so that these items had to be categorized again. The items with greater power of discrimination were sadness, pessimism, feeling of failure, dissatisfaction, self-hatred, indecision, and difficulty of work. The most serious items were weight loss, suicidal ideas, and social withdrawal. The group of 202 individuals with the highest levels of depressive symptoms was comprised by 74% of women and almost 84% had a diagnosis of a psychiatric disorder. The results show gains resulting from use of IRT in the analysis of latent traits.