We have examined whether practitioner ratings (immediately post-intervention) or other recorded characteristics of a single-session 1-hour motivational intervention were predictive of 3-month cannabis use outcome. In the context of a cluster randomized trial involving 200 non help-seeking illegal drug users (age range 16-20), 105 were randomized to the intervention, of whom 97 (92%) were interviewed for followup at 3 months, 96 of whom were current cannabis users at study entry. Six intervention characteristics and seven practitioner ratings as well as patterns of self-motivational statements were investigated in relation to substantial change in use, (which was defined as cessation or reduction by more than 50%). Both practitioner ratings post-session, and also the subject's own elicited self-motivational statements, were found to be predictive of outcome 3 months later. The strongest predictor of substantial change, however, was simply whether change had been discussed during the session. On the basis of the above findings, it does indeed appear possible for outcome to be predicted by the motivational interviewing practitioner immediately following delivery of the intervention, on the basis of simple observations and ratings. This area warrants more specific study.