Although previous research has identified various factors linked to a higher risk of problem sports betting, it is not clear which factors are most uniquely predictive when they are considered alongside each other. In this study, the researchers used two machine learning algorithms, regularized regression and random forest, to identify key factors that are most predictive of problem sports betting among young adults from a broad set of 80 candidate predictors. Data from a total of 221 young adults aged 18–29 years were used in the analysis. Across both machine learning models, sports betting motives and social functioning emerged as the most influential predictors of problem sports betting. Other key predictors included amount wagered and perceived harm from sports betting. But both models had unexplained prediction errors. This suggested some important predictors of problem sports betting were not included in this study.
