SHOULD LIKERT DATA BE TRANSFORMED USING SUMMATED RATING SCALE? A CONFIRMATORY FACTOR ANALYSIS STUDY ON THE CONTINUOUS LEARNING

Continuous Learning competence can be measured through self-assessment to minimize interview failure. The problem arises when the Continuous Learning instrument is developed using a Likert Scale. Can the data from the distribution of the instrument be used directly, or must it be converted using the summated rating scale approach? This study aims to compare the results of instrument analysis through Confirmatory Factor Analysis (CFA) based on direct data and conversion data. The method used is descriptive quantitative, involving 281 undergraduate students in Indonesia. Instrument analysis includes estimating reliability, convergent validity, and construct validity through CFA. The study's results show that the reliability and convergent validity of the data converted using the summated rating scale are higher than those of the direct data. However, the direct data produces a better measurement model fit test value compared to the converted data. However, the difference in value between the two types of data is very small and does not have any meaningful difference. Therefore, the data from the instrument can be directly analyzed without converting to the summated rating scale. This study provides insights to researchers and academics in developing a continuous learning instrument. Additionally, it offers insight into how processing Likert data on the Continuous Learning instrument is very straightforward and does not require data conversion through a summated rating scale.