EDUCATIONAL AND PSYCHOLOGICAL MEASUREMENT FERRANDO AND CHICO VALIDITY STUDIES DETECTING DISSIMULATION IN PERSONALITY TEST SCORES: A COMPARISON BETWEEN PERSON-FIT INDICES AND DETECTION SCALES

The present study examined whether an internal procedure for assessing the scalability of the response patterns, based on item response theory (IRT), can detect deliberate dis-simulation (faking good) in the Extraversion, Neuroticism, and Psychoticism scale scores of the Eysenck Personality Questionnaire Revised. The procedure is compared to the traditional approaches, which use the Lie and the Social Desirability (SD) scales. A data set was analyzed in which participants were either administered the measures in standard conditions or given special instructions to fake good. The results showed that the IRT-based measures were not powerful enough to detect dissimulation, whereas the Lie and SD scales performed much better. The interpretability of test scores is essential in personality measurement (Hogan & Nicholson, 1988). Personality test scores tend to have weaker psychometric properties than ability scores, perhaps because there are many factors, apart from the item content, that might affect the participant item re-sponses (Fiske & Butler, 1963; Jackson & Messick, 1958). In addition, re-search shows that respondents can intentionally distort scores on a wide vari