Psychometric analysis of the Grit-O and Grit-S scales in a sample of Lima university students

The objective of this study was to analyze the psychometric evidence of the Grit Scale in its two versions, since they are the most widely used instruments to measure this construct. A total of 5417 university students from Lima were selected by convenience, with a first group made up of 4417 students, mostly women (52.6%), whose ages range from 16 to 35 years old. The internal structure of both scales was analyzed by means of a confirmatory factor analysis. The Grit-S was found to be a more coherent and parsimonious measure (CFI =.98, TLI =.97, RMSEA =.08[.07-.09]) than the Grit-O (CFI =.87, TLI =.84, RMSEA =.12[.11-.13]); in both versions the existence of a general factor explaining the variance of the specific factors CI and PE was partially corroborated (Grit-O; LCS =.42; H =.84; FD =.87; Grit-S; LCS =.38; H =.73 and FD =.80). With a second group (n = 1000), it was found that the Grit-S scores presented statistically significant correlations with academic self-efficacy (r =.45**), optimism (r =.40**), growth mindset (r =.23**), and fixed mindset (r = -.23). The reliability of both scales was acceptable. In conclusion, the Grit-S is a more reliable measure than the Grit-O, and the existence of a general factor was identified, which is consistent with the theoretical model underlying the instrument.