A telephone survey of 103 Sydney general practitioners (GPs) was conducted to assess the extent of agreement or disagreement with 15 statements originating from an earlier, focus groups study, concerning patients' drug and alcohol problems. Another aim of the present study was to determine whether the results provided evidence for a typology of general practitioners. A cluster analysis indicated the presence of three groups: Interactive Problem Solvers; Traditionalist Healers and; Distant Technologists. However, a subsequent principal components analysis and re-examination of the distributions of scores lead to the preferred explanation that the survey was measuring a continuous dimension--"Willingness To Intervene" (WTI). Overall, the GPs' responses suggested high levels of willingness to intervene, although older doctors showed less willingness to intervene than younger doctors. Implications of these findings for medical educators are discussed in relation to evidence that, in practice, doctors often fail to intervene in patients' drug and alcohol problems. Willingness to intervene is viewed as one of several necessary factors, such as knowledge, clinical skills and self efficacy, none of which are sufficient alone to guarantee intervention. Arising from consideration of the willingness to intervene dimension, a general model of probability of medical intervention is outlined.