Rasch-Modeling the Portuguese SOCRATES in a Clinical Sample Paulo Lopes University Luso ´fona of Humanities and Technologies, Ares do Pinhal Gerardo Prieto and Ana R. Delgado Universidad de Salamanca Pedro Gamito University Luso ´fona of Humanities and Technologies He ´lder Trigo Ares do Pinhal The Stages of Change Readiness and Treatment Eagerness Scale (SOCRATES) assesses motivation for treatment in the drug-dependent population. The development of adequate measures of motivation is needed in order to properly understand the role of this construct in rehabilitation. This study probed the psychometric properties of the SOCRATES in the Portuguese population by means of the Rasch Rating Scale Model, which allows the conjoint measurement of items and persons. The participants were 166 substance abusers under treatment for their addiction. Results show that the functioning of the five response categories is not optimal; our re-analysis indicates that a three-category system is the most appropriate one. By using this response category system, both model fit and estimation accuracy are improved. The discussion takes into account other factors such as item format and content in order to make suggestions for the development of better motivation-for-treatment scales. Keywords: addiction, motivation; Rasch Rating Scale Model, SOCRATES. Motivation is a wide term referring to a variety of neurological and physiological factors that initiate, sustain and control behavior (Kandel, Schwartz, & Jessel, 2003). Motivation for change has been described as a condition for the treatment of drug addiction, and the absence of motivation has been used to explain the failure of drug abusers to begin treatment, continue it and succeed in completing it (Miller, 1999). The development of adequate measures of motivation is needed in order to properly understand the role of this construct in rehabilitation. Several scales have been devised to assess motivation for change in substance abusers: e.g., University of Rhode Island Change Assess- ment (URICA; McConnaughy, Prochaska, & Velicer, 1983); Readi- ness to Change Questionnaire (RCQ; Rollnick, Heather, Gold, & Hall, 1992); Treatment Motivation Questionnaire (TMQ; Ryan, Plant, & O’Malley, 1995) and Stages of Change Readiness and Treatment Eagerness Scale (SOCRATES; Miller, 1995). The SOCRATES was developed in the context of the Trans- theoretical Model (Prochaska & Diclement, 1983) and is often used in the assessment of motivation for the treatment of drug addiction. It conceptualizes motivation as a dynamic process and contemplates several stages in which the subject can be included. It is composed of 19 items, which evaluate the three stages of change of the transtheoretical model: Recognition (seven items), Ambivalence (four items), and Action (eight items). Answers are given in a five-point Likert scale (1 = No! Strongly disagree; 2 = No Disagree; 3 = Undecided or unsure; 4 = Yes Agree; 5 = Yes! Strongly agree). If the points are added up, scores range between 7 and 35 for the subscale Recognition, between 4 and 20 for the subscale Ambivalence and between 8 and 40 for the subscale Action. In the studies conducted according to classical test theory, the three subscales show good internal consistency and stability (Miller & Tonigan, 1996; Maisto et al., 1999; Maisto, Chung, Cornelius, & Martin, 2003; Figlie, Dunn, & Laranjeira, 2004). Concerning the SOCRATES factorial structure, both two and three factor models have shown good fit. For instance, in an adolescent clinical sample, a three-factor model had better absolute fit, but factors 2 and 3 of the three-factor model correlated highly and significantly (r = .81), leading Maisto et al., (2003) to choose a two-factor model out of parsimony. The aim of this study is to test the psychometric properties of the SOCRATES subscales in the Portuguese population by means of the Rating Scale Model (RSM; Wright & Masters, 1982). The RSM, according to Linacre (2002), can be expressed as: ln (P nik /P ni(k-1) ) = B n - D i - F k , where: P nik is the probability that person n, on encountering item i would be observed (or would respond) within category k, P ni(k-1) is the probability that the observation (or response) would be within category k - 1, B n is the ability, (attitude, etc.), of person n, D i is the difficulty of item i, F k , the step calibration, is a rating scale threshold defined as the location that corresponds to the equal probability of observing adjacent categories k - 1 and k. The RSM is a type of Rasch Model (RM), and thus it has optimal metric properties: sufficient statistics, specific objectivity, interval data, and statistics for person and item fit (Rasch, 1960; Prieto & Delgado, 2003). A person’s score in logits is the natural Paulo Lopes, Faculdade de Psicologia, University Luso ´fona of Humanities and Technologies, Ares do Pinhal; Gerardo Prieto and Ana R. Delgado, Departamento de Psicologı ´a Ba ´sica, Psicobiologı ´a y Metodologı ´a, Univer- sidad de Salamanca; Pedro Gamito, Faculdade de Psicologia, University Luso ´fona of Humanities and Technologies; and He ´lder Trigo, Ares do Pinhal. Correspondence should be sent to Ana R. Delgado, Facultad de Psico- logı ´a, Avda. de la Merced 109-131, 37005 Salamanca, Spain. E-mail: adelgado@usal.es Psychology of Addictive Behaviors © 2010 American Psychological Association 2010, Vol. 24, No. 2, 355–359 0893-164X/10/$12.00 DOI: 10.1037/a0019080 355