How Grit-S Items Work in Measuring Growth: Focusing on Item Discrimination and Sensitivity to Change

Abstract

This study investigates how sensitively the current grit measure (Grit-S) is able to reflect the differences between individuals (i.e. item discrimination) and change within individuals (i.e. sensitivity to change of items) using longitudinal data. Data from the Korean Children and Youth Panel Study 2018 of elementary (fourth grade) and middle school (seventh grade) cohorts were analyzed. We compared a series of longitudinal factor models of change with an initial trait and a change in grit represented by latent traits. Results showed that the models hypothesizing two latent traits with respect to the two elements of grit (interest and effort) best fit the data. In the elementary school cohort, sensitivity to change was different from discriminating between-person differences at baseline. Meanwhile, in the middle school cohort, sensitivity to change was identical to discriminating between-person differences at baseline. The items having low discrimination and low sensitivity to change resulted in syntactical complexity and did not accurately reflect the trait that was intended to be measured. Despite the Grit-S being widely used for many years, it is suggested that researchers and educators should administer this measure at taking an individual’s change and school-level difference into account.