Optimal Reliability and Validity of Measurement Model in Confirmatory Factor Analysis: Different Likert Point Scale Experiment

Abstract

Designing a questionnaire is one of the most difficult challenges in research design, particularly when deciding which level of Likert point scale is appropriate for the instrumentation. Suitable Likert point scale used in the instrumentation able to reduce the risk of facing potential problems of not achieving reliability (indicator and internal consistency reliability) and validity (convergent validity, discriminating validity, and construct validity) and simultaneously preventing the occurrence of multicollinearity. This study compares the performance of reliability and validity of measurement construct in Confirmatory Factor Analysis (CFA) by using Likert 5-point Likert scale, 6-point Likert scale, 7-point Likert scale, 9-point Likert scale, and 10-point Likert scale. The study uses primary data based on a questionnaire data collection method which involves 100 samples from similar population characteristics for each Likert. The data were analysed using Smart-PLS software. The results suggest that expanding the range of the Likert point scale optimizes the performance of reliability and validity of the measurement model. This study offers an insight to researchers in deciding the best choice of Likert point scale to adapt in instrumentation for a better result in the quantitative analysis process.