CATEGORIZATION OF COMPOSITE SCORE AND TOTAL SCORE OF LATENT CONSTRUCT

Abstract

The categorization of composite scores and total score is an important step to achieve the aim of some quantitative research involving latent construct, but the required understanding about how the categorization should be done aptly based on the features of the data is lacking. The purpose of this study is to guide researchers on how categorization of composite scores and total score are done based on the features of the data. Consideration of the measure of central tendency of the data, the presence of outliers in the composite score and total score are factors that are necessary for consideration in order to validate the categorization. Three different approaches of categorization were deliberated to show that winsorized composite scores yielded an unbiased mean of composite score which can represent the overall response of a target population in research involving latent construct.