“…Among these are the invariance of measurement, error in measurement per item (not for the entire instrument, as in the CTT), knowing the position of each item on the continuum, and finding the most probable response to each item for a person with a given severity of dependence (Bezruczko, 2005;Conrad and Smith, 2004;McHorney and Monahan, 2004;Prieto, Alonso, and Lamarca, 2003). Some of the studies performed in drug addiction have used these models to analyze the dimensionality of a nicotine dependency scale (Breteler, Hilberink, Zeeman, and Lammers, 2004), the uniformity of substance dependence syndrome elements based on the criteria in the DSM-III-R and ICD-10 (Kan, Breteler, van der Ven, and Zitman, 1998) and discrimination of items on an alcohol dependence scale (Kahler, Strong, Hayaki, Ramsey, and Brown, 2003). The purpose of this work was to apply an IRT model, and more specifically, the Rating Scale Model, which is a polytomous model, derived from Rasch Model, to contribute evidence of the validity of the SDS construct.…”