Development and calibration of an instrument measuring attitudes toward statistics using classical and modern test theory

Abstract

Assessment of attitudes towards statistics [ATS] is needed to support the success of statistics education in tertiary institutions, so measuring instruments with high accuracy is required. However, existing instruments to measure ATS have not considered the use of technology as an essential variable affecting success in statistics education. The current study sought to fill this gap by developing a standardized instrument to measure ATS and considering aspects of technology use as a necessity for statistics education in the modern era. The study involved 367 students from various study programs spread across several universities in Indonesia as participants. To examine the quality of the instrument, we performed factor analysis, reliability estimation, and item calibration. We calibrated items based on classical test theory [CTT] and item response theory [IRT] using the graded response model [GRM]. Exploratory factor analysis [EFA] indicated three main factors (i.e., interest, difficulty, and value) for measuring attitudes toward statistics. Factor loading of each factor component > 0.45, indicating that all items contributed to the main factor. Cronbach's alpha coefficient of the three factors ranged from 0.784 to 0.929, indicating that the instrument was reliable. Item calibration based on CTT and IRT-GRM indicated that item performance was satisfactory regarding item endorsement and discrimination. In addition, the information function indicated that the instrument accurately measures attitudes from very low to very high levels. Overall, the psychometric properties of the instrument indicated that the instrument was valid, reliable, and feasible for use in practice and research in the field of education.