AIM: To develop a quantified picture of the factors that lead people into problem gambling and in particular examine the link between impulse control disorders, such as ADHD, and problem gambling. PARTICIPANTS: The recruitment effort invited people who were self-described social gamblers to participate. However, consistent with previous experience by the investigators, the study attracted gamblers representing the full spectrum of gambling involvement, from non-problem social gamblers to people with severe gambling problems. METHOD: First, responses from self-described social gamblers were collected via an extensive questionnaire. Second, attention span and impulsivity was tested using a computer software package. Thirdly, genetic analysis was performed on blood samples taken from the participants. Data were analyzed using factor analysis, as well as a statistical modelling technique known as LISREL analysis. FINDINGS: Problem gamblers were not significantly more likely than non-problem gamblers to have had an early history of gambling involvement. This finding contradicts the widely held conception that exposure to gambling during the teenage years (or earlier) predisposes individuals to acquire a gambling problem later in life. What appears to be more important in the development of an interest in gambling is the actual experience of winning itself. However, to turn an interest into a problem often also requires the addition of stressful life events, poor coping skills, or impulsivity. Both the timing and the size of the win were associated with problem gambling. The definition of what constituted "big" varied among gamblers, but on average there was very little difference between what problem and non-problem gamblers felt was a big win. CONCLUSIONS: In general, the data do support Blaszczynski's hypothesis regarding the multiple pathways nature of problem gambling development. Evidence was found for emotional, cognitive, and behavioural contributions. The study also made a number of new discoveries concerning the relationships among variables that are important to problem gambling development. While no evidence was found in support of the genetic models, we did find evidence in support of the general notion that impulse control is an important factor in problem gambling, suggesting that physiological models of problem gambling need to be further explored. In addition, we have evidence that some people have no pre-existing disorder, but are basically 'normal' people. We also have found evidence for some of the factors that contribute to the development of gambling problems amongst those who are neither physiologically nor emotionally vulnerable.