Concordance between the original and short version of the impulsive behaviour scale UPPS-P using an IRT model

Abstract

The UPPS Impulsive Behaviour Scale (with five dimensions of impulsive behaviour) is being widely used. One of the handicaps of this instrument is its relatively long administration time. This has led to the development of a short version: SUPPS-P. There are no studies comparing the relationship between the SUPPS-P scores and the original UPPS-P scores. The objectives of this study, therefore, were to analyse the psychometric properties, concordance of person measures, and efficiency of the SUPPS-P compared to those on UPPS-P, applying an Item Response Theory Model. The UPPS-P and SUPPS-P were administered to 455 undergraduate students. Confirmatory factorial analysis replicated structures reported in previous studies: the five correlated factors structure and the model with two second-order factors (Urgency and Lack of Awareness) with Sensation Seeking dimension as a different factor. Rasch analysis show that both of the instruments presented adequate model-data fit. The results show the measurement for each dimension of UPPS-P offered more precision than SUPPS-P. The structure of items location was maintained in each dimension of SUPPS-P compared to the UPPS-P, but with better person and item separation indices of the UPPS-P dimensions. The concordance analysis reveals high correlations values between scores on both versions. From the standpoint of reducing items, it can be considered that the reduced version is more efficient. This study does not support the equivalence of items on the dimensions of Sensation Seeking and Lack of perseverance.