Latent trait standardization of the benzodiazepine dependence self-report questionnaire using the Rasch scaling model

Abstract

The aim of the present study was to obtain standardized scores that correspond with the raw scores on the four Rasch scales of the Benzodiazepine Dependence-Self Report Questionnaire (Bendep-SRQ). The eligible normative group for standardization of the Bendep-SRQ scales consisted of 217 general practice (GP) patients, all using benzodiazepines. Two standardization methods were used and compared: "classical standardization," which transforms raw scores into standard scores on the unit normal distribution, and "latent trait standardization," which transforms raw scores into latent trait scores. The latter requires the Rasch model with the additional assumption of a normally distributed latent trait, which held true for the scales "problematic use," "lack of compliance," and "withdrawal," but not for "preoccupation." The observed unequal item spacing on the "preoccupation" scale was hypothesized to induce a response tendency of nondeviation, causing a local violation of the assumption of a normally distributed latent trait. Nevertheless, comparison of the results of the two standardization methods revealed such a high degree of resemblance, that latent trait standardization could be used for "preoccupation" just as well as classical standardization. The presented standard scores and corresponding percentile ranks make raw Bendep-SRQ scores clinically interpretable in relation to the normative GP sample. Incorporation of the Rasch scaling methodology into the development of the Bendep-SRQ marks the adoption of the item response theory in the field of applied test methodology. In this process, it appears that equal item spacing has to be taken into account to prevent local violations of the Rasch model with the additional assumption of a normally distributed latent trait.