The effect of varying the number of Likert response score points on the psychometric properties of psychological scales: the self-efficacy scale as a model

Abstract

The study aimed to investigate the effect of the number of response points to the Likert scale on the psychometric properties of the psychological scales. The researcher relied on the self-efficacy scale as a model. The study sample was chosen randomly from students of the Faculty of Educational Sciences at Imam Muhammad bin Saud Islamic University, Department of Psychology, from all academic levels. The study relied on the use of confirmatory factor analysis to estimate constructive validity to estimate the difference between the different models. The internal consistency stability was estimated using Cronbach's alpha coefficient for each construct. The results were in favor of the four-response model in the first place, then the five-response, then the triple-response, then the six-response model. Whereas, the seven-level model showed poor matching response. The study recommends relying on psychological tests and measures consisting of a quadruple or six-fold scale in the case of measures whose psychological and emotional characteristics are affected by the nature of the neutral response, or a rejected response affected by social acceptability. The study also confirms the use of the five-response staging because of its statistical immunity.