Machine learning in predicting the effects of personality traits on college students’ grit in Ghana

This study examines whether Big Five personality traits are associated with grit among Ghanaian teacher trainees, applying regularisation assisted econometric estimation within a double machine learning framework. Using survey data from 288 students at Jasikan College of Education, three complementary post lasso estimators were applied, namely ordinary least squares with CHS lasso orthogonalised variables, CHS post lasso orthogonalised variables, and PDS selected variables alongside the full regressor set, each employing lasso purely for covariate selection and orthogonalisation within an OLS framework. A cross validated predictive lasso model corroborated these estimates. Across specifications, openness and conscientiousness consistently predicted higher grit, with openness exceeding conscientiousness in effect size, while extraversion, agreeableness and neuroticism showed no significant association. These specific findings suggest that personality dimensions relevant to grit may depend on resource constrained teacher training conditions, offering targets for interventions strengthening grit among prospective educators in Ghana.