Psychometric properties of Grit-S scale in university students: An application of the graded response model

Abstract

This study adopted an item response theory to evaluate the Grit-S scale and its dimensions (i.e., Consistency of interest [G-CI] and Perseverance of effort [G-PE]) among university students (N = 450) using two unidimensional Graded Response (GR), multidimensional GR, and bifactor GR models. Differential item functioning (DIF) across two countries (Oman and the United States) was investigated. For G_CI, the findings showed acceptable values for item discrimination, trait thresholds, reasonable precision in both samples, and two uniform DIF. The findings of G_PE showed two questionable item-discrimination parameters (Items 7 and 8) in the two samples, two non-uniform (i.e., Items 5 and 8) and two uniform DIF (i.e., Items 6 and 8). Acceptable precision was found only in the U.S. sample. Item 8 was the source of imprecision in the Omani sample. These results indicated that further modification is needed to improve the precision of the scale in the Omani culture.