Formative Measurement Scale Development: An Example Using Generalized Structured Component Analysis

Abstract

Unlike reflective measurement scales, the steps for development of formative measurement scales tend to behighly subjective and involve mostly the judgment of the researcher. Formative scales have been criticized for this reason.This paper extends Christophersen and Konradt’s (2008) method of joint development of a formative and reflective scale toassess mutual validity of each scale. We utilize a second order method to reduce measurement error in the formative scaleas suggested by Edwards (2011), and test the efficacy of Generalized Structured Component Analysis (GeSCA) for thispurpose. For illustrative purposes, we utilize a sample of formative and reflective job satisfaction survey data both to testour joint formative/reflective scale development technique and to assess which formative aspects of job satisfaction alignwith commonly used reflective job satisfaction scales.