Validating the Academic Writing Creativity and Self-Efficacy Scale: A Rasch Model Analysis

Abstract

The current study aimed to examine the psychometric characteristics of an instrument used to evaluate academic writing creativity and self-efficacy scale among pre-service teachers, instructors, and school teachers. Using a convenience sampling, an adapted and expanded version of  Mitchell’s (2021) situated academic writing self-efficacy (SAWSES) questionnaire was distributed to 655 participants with three different occupation statuses, such as pre-service teacher (N = 315, 485%), TVET instructor (N = 80, 12%), and school teacher (N = 260, 40%). Unfortunately, due to the presence of outliers, of 655 data, only 582 data were used for Rasch statistical analyses, including the analysis of (a) unidimensionality, (b) reliability for the item and the person separation, (c) rating scale, (d) item bias, and (e) item differential function. The analysis of Differential Item Functioning (DIF) resulted that the academic writing creativity and self-efficacy scale could potentially be biased when employed to collect the data from participants with differing teaching statuses and educational backgrounds. These findings thus suggest that caution should be taken when using the Academic Writing Creativity and Self-Efficacy Scale, as highlighted by the results of the current study.