LAUSR

Abstract: Grit, defined as passion and perseverance for long-term goals, has received much attention over the last decade. However, its dimensionality is still very much up for debate. The present study examines whether the bidimensional structure of the construct is a statistical artifact caused by item reversal. The sample consisted of 1,115 participants ( Myears = 38.92; SD = 15.08), of whom 67.3% were women. Several confirmatory models were specified with different versions of two grit scales. A bifactor model with a grit-general factor and two method-specific factors (one direct items, the other reversed items) demonstrated better fit according to all indicators (Original Grit-S: CFI = .995, RMSEA = .024; Mixed Grit-S: CFI = .966, RMSEA = .068; Mixed EGO: CFI = .985, RMSEA = .051) than a bifactor model with a grit-general factor and two theoretical factors (perseverance of effort and consistency of interests, Original Grit-S: CFI = .933, RMSEA = .085; Mixed Grit-S: CFI = .944, RMSEA = .088; Mixed EGO: CFI = .982, RMSEA = .056). Grit is a unidimensional construct, and the two dimensions identified in previous studies (consistency of interests and perseverance of effort) were the result of a statistical artifact due to item reversal. In addition, redirecting the items modified the factor structure of the Grit-S scale. The practical implications of the study are discussed.