Gambling problems contribute to financial ruin, family dysfunction and divorce, job loss, fraud, and embezzlement. Respecting consumer choice, economists are concerned with negative externalities of gambling. The external costs of adult gambling in the US are at least $5 bn per year, since many gambling-caused harms cannot be easily computed in terms of income. This raises social concern about gambling prevalence and intensity, the most popular casino games, and the manner in which the US Federal, State, and Tribal governments regulate casinos.

This thesis is dedicated to the analysis of slots play in an economics laboratory and the role of game design on gambling behavior. Chapter I reviews academic literature on gambling prevalence, gambling disorder assessment, and slot machine play. Unfortunately, many gambling studies use questionable metrics of gambling prevalence. At the same time, the gambling industry popularizes self-reported prevalence studies which use clinical criteria for gambling disorder diagnostics that were never meant for population studies. The numerous shortcomings in existing gambling research call for novel approaches in gambling studies, such as slot machine experiments in the economics lab.

In Chapter II, I review controlled studies of slot machine play and propose an experiment to investigate the differences in gambling behavior by varying the degree of “fun” in the slot machine user interface. The fun treatment features the interface of Harrison, Lau, Ross, and Swarthout (2015) with celebratory sounds, recognizable slots symbols, and spinning reel animations. The dull treatment maintains the same paytable, but presents a minimal user interface with none of the sounds, symbols, or animations. Unlike most controlled studies with slot machines, this experiment allows subjects to earn a cash reward, $49 on average, and then decide how much of it to spend on slots play, if at all. Finally, I present the data on subjects’ decisions to gamble with their own money and betting behavior. The results show that 86 of 200 subjects choose to play slots, rather than leave the “laboratory casino,” using their own pre-slots earnings to gamble. Conditional on the choice to gamble, subjects, on average, make 64 total spins, betting 34 cents per spin. Additionally, I stratify the gambling decisions by treatment, risk preferences, and gambling disorder screen reports, and perform statistical significance tests.

Chapter III is devoted to the effects of the fun slots design on subjects’ slot machine play behavior and their beliefs over payouts. Analyzing spin-by-spin bets, I find that the fun slots design increases subjects’ out-of-pocket bets by 93% of their mean. Using belief elicitation, I find that the fun treatment is associated with pessimism over expected payouts. The evidence of the greater bets despite additional pessimism, suggests that the fun slots design treatment dominates the considerations of bets and payouts. These results stress the importance of game design in gambling devices and explain slots operators’ massive investments in user experience research and state-of-the-art visual technologies, exemplified by slot machine terminals released in the last decade.