Association for Behavior Analysis International

The Gambling Special Interest Group (G-SIG), previously named the BIG-SIG, is a member-supported organization of researchers, instructors, practitioners, and students who study gambling behavior and the treatment of pathological gambling from a behavioral perspective. The G-SIG publishes two issues of the journal Analysis of Gambling Behavior (AGB) every year, which contains peer-reviewed articles related to the scientific study of gambling and the treatment of pathological gambling. The recent issues contain articles from many distinguished gambling researchers on topics including the modeling of gambling using animal subjects, motivating operations in laboratory gambling environments, delay and probability discounting, and the gambling functional assessment tool. The G-SIG also hosts an annual gambling conference and organizes symposia at the national and international ABA conferences. This symposium contains talks on the recent developments in studying gambling behavior in a laboratory setting and in the assessment of gambling behavioral function within a clinical or laboratory setting.

Simulated-gambling tasks that closely model the temporal game play and contingencies found on real games of chance have been used to collect behavioral measures of gambling under laboratory conditions for more than half a century. These tasks can be played with game credits or chips that are exchangeable for real money, however, for ethical reasons the credits are usually given to participants freely prior to gambling and the researcher pays participants for remaining credits at the end of a session. Whether this standard-stake procedure establishes a game-credits market that simulates a natural game-credits market remains unclear. Computerized gambling tasks have been recently developed to study whether the validity of the laboratory game-credits market can be improved by allowing participants to earn game credits within a session rather than providing staked credits prior to gambling. The findings from two experimental investigations will be discussed.

The Balloon Analog Risk Task is a computerized assessment developed to estimate individual propensity to take risks. We explored the utility of analyzing local response patterns to assess the relationship between BART scores and choice between a guaranteed cash reward and an uncertain reward of the same average value. After random assignment to 1 of 3 choice tasks, young adults played the BART and then were offered a choice between $5 in cash and betting to win $0 to $15. The mean number of adjusted responses per balloon in the BART predicted choice for the riskier monetary option. In addition, specific task offered to participants in each group significantly contributed to the choice between assured and uncertain cash. As a group, betting was associated with higher pumping rate and longer balloon duration. Consistent patterns of responding between explosions and within-balloons differentiated risk takers from those who took the safe choice. Throughout the session, responding became more effectual for both groups, but remained sub-optimal. These results show that the higher number of pumps and explosions that characterize risk takers, result from particular forms of adaptation to the positive and negative outcomes of risk taking behavior seen at a more molecular level.

The Gambling Functional Assessment - Revised (GFA-R) was designed to determine whether the respondent's gambling behavior was maintained by positive reinforcement and/or escape. This measure has been shown to have sound psychometric properties and to be both internally and temporally reliable. Research on the GFA-R, however, has primarily been conducted on university populations. Thus, the present study recruited adult citizens from across the U.S. to complete the GFA-R on Amazon.com's Mechanical Turk. Four hundred participants completed the GFA-R along with two other measures of potential gambling problems. Results demonstrated that the psychometric properties and internal consistency of the GFA-R remained good when tested on this diverse sample. Likewise, the new results also replicated previous ones in that A) respondents endorsed gambling for positive reinforcement to a significantly greater extent than they did gambling as an escape and B) that endorsing gambling as an escape was more strongly associated with the reporting of gambling problems than was endorsing gambling for positive reinforcement.