Secondary analysis of loot box data: Are high-spending “whales” wealthy gamers or problem gamblers?

•Novel analysis of open-access loot box survey data, comprising 7,771 loot box purchasers.•Significant, moderately sized correlation between problem gambling scores and self-report loot box expenditure.•No significant correlation between loot box spend and self-report earnings.•Games developers disproportionately profiting from moderate and high-risk gamblers, rather than high earning customers. Loot boxes are purchasable randomised reward mechanisms in video games. Due to structural and psychological similarities with gambling, there are fears that loot box purchasing may be associated with problematic gambling. Whilst monthly expenditure is typically modest (i.e. < $20), the distribution is highly skewed, with a small number of high-level spenders, sometimes referred to as “whales”. It is not known what proportion of industry profits are derived from such players, and whether they are typically wealthy individuals and/or problem gamblers. We used structured literature searches to identify surveys of gamers with open-access loot box data. The resulting datasets were aggregated, and correlations between loot box expenditure, problem gambling and earnings investigated using Spearman’s rho correlations. The combined open-access data comprised 7,767 loot box purchasers (5,933 with self-report earnings). Secondary analysis of this self-report data confirmed that disproportionate revenue appears to be generated from high-level spenders: the top 5% of spenders (> $100/month) represent half of loot box revenue. Previously reported correlations between problem gambling and loot box expenditure were confirmed, with an aggregate correlation of ρ = 0.34, p