Background <p>Loot boxes are randomized reward mechanics in modern video games that share features with conventional gambling products. Research studies have begun to test longitudinal patterns (“migration”) from engagement with loot boxes to gambling behavior. This study investigated such effects at a 6-month follow-up in an online sample of young adults that play video games (aged 19—25) from British Columbia, Canada.</p> Methods <p>Participants were stratified into two subgroups at their baseline assessment: 83 reported they did not currently gamble and 43 reported they currently gamble, after cleaning. At baseline, participants provided responses to the Risky Loot Box Index (RLI) and estimates of their past year spending on both randomized (i.e., loot boxes) and non-randomized (“direct purchase”) microtransactions. Microtransaction spending and RLI scores at baseline were tested as predictors of self-identified gambling initiation and spend at follow-up. We tested a set of frequentist regressions and a corresponding set of Bayesian regressions.</p> Results <p>At baseline, participants who reported gambling showed higher levels of engagement with both randomized and non-randomized microtransactions. Among non-gambling participants at baseline, loot box spending and RLI predicted gambling initiation at the follow-up, in a Bayesian logistic regression with informed priors. Loot box spending and RLI at baseline predicted gambling expenditure at follow-up, in both the frequentist and Bayesian linear regressions. Spending on direct purchase microtransactions did not predict gambling initiation in either set of models when controlling for loot box spending, underscoring the role of randomized rewards.</p> Conclusions <p>These data provide further prospective evidence for gambling ‘migration’ in a sample recruited in Western Canada, indicating that young adults who spend money on loot boxes are at elevated risk for real-money gambling.</p>