USADA SCIENCE SYMPOSIUM PART THREE:
A Model of Perceptual Deterrence
By Dr. Larry Bowers, USADA Chief Science Officer
The United States Anti-Doping Agency’s 14th annual Symposium on Anti-Doping Science took place from Oct. 3-5 at the Lansdowne Resort in Leesburg, Va. Located just 35 miles from our nation’s capital, the symposium was comprised of a global network of anti-doping experts, including scientists from both WADA accredited laboratories and National Anti-Doping Organizations, as well as academic experts representing institutions throughout the United States and United Kingdom. Tasked with exploring the critical challenges currently facing the anti-doping community as well industry wide best practices, this year’s symposium was centered on the role of deterrence and how it can be used to cultivate a more effective global anti-doping effort.
Even when athletes recognize that doping is cheating and cheating is inherently wrong, it is possible they may not directly apply that knowledge into their own actions. It is in these circumstances that the threat of anti-doping organizations both detecting the use of performance-enhancing drugs and methods, and sanctioning the athlete for this use, can actually deter athletes’ behavior as it pertains to doping. It can be said then that perceptual deterrence provides insights on how to achieve compliance with anti-doping rules using all the tools available to anti-doping organizations (rule-making, sample collection, testing, education, adjudication, and formal and informal sanctions) and thus is the keystone of anti-doping. Ultimately, the concept of perceptual deterrence is based on the athlete (and his support personnel) making rational decisions on whether or not to dope based on the perceived costs and benefits.
USADA has been working with RTI Health Solutions for several years to build a model based on an expected net benefit approach. The model considers both the financial and non-financial rewards and costs of doping, the perceived likelihood of reaching a higher rank in sport with or without doping, and the perceived probability of being detected. A decision tree approach is then used to compute whether or not individuals within a cohort of athletes would dope or not based on the relative rewards and risks. The inputs can be varied to consider the impact of programmatic changes on the number of tests and resources needed to achieve a desired deterrence. As would be expected, increasing numbers of tests result in increased perception of being caught – hence, increasing deterrence. Other factors, such as a tipping point with little deterrence below a certain level of testing, act to require increased test numbers to achieve a particular level of deterrence.
The model includes a fairly large number of inputs, including some which require knowledge of athlete perceptions. USADA and RTI Health Solutions conducted a survey of athletes’ perceptions using a web-based survey instrument that included questions using the double-bounded contingent-behavior method. Athletes believe that performance-enhancing drugs (PEDs) work and that competitors have used them in and out of competition without being detected. Athletes also feel that health risks associated with PED use are not overstated. Like non-athletes, the athletes felt that informal sanctions (loss of status and reputation imposed by teammates, competitors, and significant others) were at least as “damaging” as periods of lost eligibility. We were also able to make some quantitative assessments of athlete attitudes and behaviors.
The results of the survey were used to inform the model. For example, 39% of athletes would not use PEDs regardless of the rewards and costs. These are athletes that have strong moral inhibitions. The perception that the collection and testing methods can detect the doping behavior significantly decreases the deterrence, even with twice as many tests per athlete. This demonstrates the importance of well-timed collections and scientific research to increase detection windows. The effects of many other components of the testing program can be evaluated with the model. The model can also be run to determine the best test distribution at a fixed budget.
While this year’s theme was not entirely focused on physical or biological science, this was a chance to expose scientists to the world of social science and its role in anti-doping and deterrence. Symposium attendees left with a better idea of how other pieces of the puzzle fit with and complement their scientific testing efforts in order to best support clean athletes.