Investigation of instrument validity : Investigate the consistency between criterion and unidimensional in instrument validity (case study in management research)

Purpose The aim of this research is to investigate the consistency between the criterion validity and the unidimensional validity – case study in management research. Design/methodology/approach The investigation of construct validity and criterion validity by using confirmatory factor analysis is presented in this paper. Data were found from three dissertations of Doctoral Program of Management on Brawijaya University and Airlangga University. It consists of seven variables. The first data consist of three variables, i.e. X1, X 2 and Y1; the second data consist of two variables, i.e. X 2 and Y1; and the third data consist of two variables, i.e. X3 and Y1. Findings The measurement instrument of latent variables must be valid. There are many methods of construct validity, i.e. unidimensional validity and criterion validity. The criterion validity is less sensitive than unidimensional validity, that is, it is more simple and easy for application. Research limitations/implications The research is limited to only analyzing criterion validity and unidimensional validity. Practical implications The unidimensional validity is more sensitive than criterion validity, that is, it is more complex and difficult for application. But if structural equation modeling (SEM) is used to analyze data, the unidimensional validity test must be used; this causes SEM to have constraint, i.e. factor indeterminacy. Social implications Majority in Dissertation of Doctoral Program (Brawijaya University and Airlangga University) using criterion validity to investigate the instrument validity, without investigating the unidimensional validity, is included in this paper. Originality/value To investigate the instrument validity, no studies have focused on the selection of instrument validity, particularly in the field of management.