On the Need for a Behavioral Analysis of Gambling

Mark R. Dixon and Kimberly Zlomke
Southern Illinois University

While the definition of problem or pathological gambling differs across treatment facilities and information organizations (c.f. Gamblers Anonymous, 2002, National Council of Problem Gambling, Inc., 2002), it is generally agreed that it involves persons who gamble more money than they can afford and incur losses greater than expected. Almost 25% of Las Vegas' gambling revenue is generated by pathological gamblers (Ghezzi, Lyons, & Dixon, 2000), many times at a great financial and psychological cost. Laduoucer, et al. (1994) found the overall indebtedness of gamblers varies considerably, ranging up to $75,000 to $150,000, and that losses may range between $4,500 and $50,000. The average gambler debt reported to a Gamblers Anonymous help line was $35,185 (Council on Compulsive Gambling, 1998).

In addition, there is an increased tendency for other societal problems. Pathological gamblers often face financial, employment, family and psychological difficulties as a result of their behavior, and are more prone to income related crime (Ladouceur, Boisvert, Pepin, Loranger, & Sylvain, 1994). Co-morbidity of substance abuse and depressive disorders is also common among the majority of compulsive gamblers.

A variety of treatments are currently available for pathological gambling but the effectiveness of many is questionable. The most popular is Gamblers Anonymous (GA). Although GA chapters exist throughout the country the program has only achieved an 8% success rate of maintenance of abstinence for at least one year. The same 12 Steps are used in the various addictions 'Anonymous' groups (Gamblers, Alcoholics, and Narcotics) and one may question whether each of addictions serve identical functions in those afflicted with the disorder/addiction. The most effective treatments are those devised to remedy the function of the disordered behavior. Other treatments include psychodynamic approaches, pharmacotherapies, and manual guided therapies (Petry & Armentano, 1999).

Unfortunately, a behaviorally based approach administered by persons versed in the principles of behavior analysis is lacking. So, why the hesitation, neglect, or disinterest of behavior analysts in treating persons with gambling problems? We believe that since behavior analysts pride themselves on practicing empirically validated treatments, the problem arises in obtaining the data needed to guide treatments. We describe two research paradigms to gather data on gambling behavior, laboratory research and naturalistic observation, and some problems associated with each.

Due to the constraints in place by casinos and gambling regulation bodies, natural experimental research on gambling has been nonexistent. Experimental manipulation of games in the actual casino environment is almost impossible due to casino rules on observation and the fact that odds and payoffs for most games are state enforced and cannot be modified. This limits the behavioral investigation of the contingencies, rules, and their interaction, to the laboratory. Many variables that are inherent in the casino settings are not applicable or are obviously contrived in laboratory settings; for example alcohol, cigarettes, noise, and observation by other players. Other problems with the lab approach is the difference between betting and winning non-valuable points or course credit compared to tokens or actual money bet and won in casinos, the use of gamblers' own money, and subjects being allowed to end their gambling session when they want to.

While we could wait for the day that our lab can be licensed as an actual casino, instead we have made initial, albeit somewhat limited, attempts at understanding problem gambling behavior in a controlled laboratory setting. This is the same rationale that Skinner took with his empirical analysis of behavior. In addition, it allows behavior analysis to compete for research and treatment dollars with other professions interested in gambling problems.

To aid in the study of gambling from a behavioral perspective, we have devised computerized video poker (Dixon, MacLin, & Hayes, 1998) and slot machines (MacLin, Dixon, and Hayes, 1999) that allow manipulation of jackpot size and probability. Both programs allow behavior analysts to begin developing a basic understanding of a variety of critical variables involved in gambling, including the simple parameters of probability and magnitude of reinforcement. Skinner's original conceptualization of gambling behavior was that it was based on VR schedules, but conclusive data are still lacking. We believe that other variables, such as duration of play, size of bets, level of risk, and rules given to or generated by the gambler, warrant investigation. Likewise, a more comprehensive analysis of the behavior might require a conceptualization from a matching/maximization perspective, delayed reinforcement, or behavioral economics. The possible interactive effects of alcohol, nicotine, and presence of others could be studied in laboratory as well. Although research on gambling with computer-simulated games will provide many opportunities to study gambling behaviors that were not previously accessible, there is also a need for more naturalistic research and clinical intervention to be attempted.

To study gambling behavior effectively, behavior analytic researchers and clinicians also need to be aware of the functional relations that influence the behavior before attempting to design clinical interventions. We believe that it is necessary to observe gamblers in their natural environment in order to understand the variables and relations that affect gambling behavior.

Also, we should not assume that the reason one person gambles is necessarily the same as the next person. Many distinctions have been made along the continuum of disordered gambling, ranging from pathological gamblers, who meet the criteria of a DSM diagnosis, to problem, compulsive, addictive, and recreational gamblers. "Compulsive' and 'problem' gambling are used almost interchangeably, with both categories of gamblers having growing and continuing problems due to their gambling and disruptions in their life (Council on Compulsive Gambling, 1998). To be diagnosed as a pathological gambler according to DSM-IV standards one must meet five of seven criteria, therefore there is a wide combination of symptoms that may be present in any individual gambler.

The process to investigate function should approximate that of the functional analysis used to assess, and ultimately derive effective treatments for, other types of problem behaviors. One could administer structured interviews or questionnaires to clients to assess possible gambling behavior function, have clients self-monitor antecedents and consequences that surround their gambling adventures, or utilize procedures such as think-alouds while gambling (beginning in the lab) to get an idea of the verbal rules that may be controlling their behavior. Although self-monitoring procedures would allow researchers to observe events and processes otherwise not observable, it adds the possibilities of reactive effects. These reactive effects may confound the empirical research but may be therapeutically advantageous. We are confident that a functional analysis would increase the effectiveness of interventions.

In summary, the population of persons with gambling problems is on the rise. Behavior analysts have made great advances in the treatment of persons with a variety of disabilities. While the gambling issue is relatively a new one, it does not appear that it will be going away anytime soon. The robustness of our science can, hopefully, include the understanding, predicting, and controlling of problem gambling. Although we cannot predict our eventual success rate, the two of us are betting it will be great than 8%.

Note: Since there are increasing numbers of behavior analysts becoming interested in gambling over the past couple of years, the "Behaviorists Interested in Gambling" Special Interest Group of the Association for Behavior Analysis was created in 2001.

References

Council on Compulsive Gambling of New Jersey, Inc. Retrieved April 8, 2002 from http://www.800gambler.org.

Dixon, M.R., MacLin, O.H., & Hayes, L.J. (1999). Towards a molecular analysis of video poker playing. Behavior Research Methods, Instruments, and Computers, 31(1), 185-197.

Ghezzi, P., Lyons, C., & Dixon, M.R. (2000). Gambling from a socioeconomic perspective. In, W.K. Bickel & R.E. Vuchinich (Eds.) Reframing health behavior change with behavioral economics. New York: Erlbaum.

Ladouceur, R., Boisvert, J., Pepin, M., Loranger, M., Sylvain, C. (1994). Social cost of pathological gambling. Journal of Gambling Studies, 10, 399-409.

MacLin, O.H., Dixon, M.R., & Hayes, L.J. (1999). A computerized slot machine simulation to investigate the variables involved in gambling behavior. Behavior Research Methods, Instruments, and Computers, 31, 185-187.

Petry, N.M. & Armentano, C. (1999). Prevalence, assessment, and treatment of pathological gambling: A review. Psychiatric Services, 50, 1021-1027.

Mark R. Dixon is an Assistant Professor in the Behavior Analysis and Therapy program at Southern Illinois University. He heads a gambling laboratory for teaching and research. His interests include choice, self-control, delayed reinforcement, and organizational behavior management. Kimberly Zlomke is a graduate student whose interests include behavior therapy, stimulus equivalence, and addictive behavior.

Return Home