Two contrasting hypotheses concerning the relationship between personality and mood are distinguished. First, broad traits may relate to emotional temperament, and so predict mood across situations. Second, the interactionist approach to personality implies that narrow, context specific traits may sometimes be the most powerful predictors of mood within situations. This article reports correlations between mood and broad and narrow trait measures, the Eysenck Personality Inventory (EPI; Eysenck and Eysenck, 1964) and the Driving Behaviour Inventory (DBI; Gulian, Matthews, Glendon, Davies and Debney, 1989), within the context of vehicle driving. Mood was measured with the UWIST Mood Adjective Checklist (UMACL; Matthews, Jones and Chamberlain, 1990), in two samples before and after a simulated drive. One sample (N = 73) performed a ‘passive’ drive, in which little interaction with other traffic was required. The second sample (N = 93) performed an ‘active’ driving task, in which subjects had to decide when to overtake other vehicles. Results showed that the DBI traits were more strongly related to mood than EPI traits, particularly following the active drive. The DBI Dislike of Driving scale was the strongest single predictor of post-drive mood. Prior to the drive, subjects also rated accident risk, driving skill, and judgement, for themselves and for a ‘peer’ driver of similar age and sex. Analysis of these data in the combined sample (N = 166) showed that the DBI was the more consistent predictor of self-ratings of risk and driving competence, although some relationships between ratings and the EPI were found. Again, the DBI Dislike of Driving scale was the strongest single predictor of self-ratings. Drivers scoring high on this scale seem immune to drivers' general bias towards rating themselves as safer and more competent than their peers. It is concluded that narrow traits are more predictive than broad traits within the driving context. Data are consistent with the transactional model of driver stress, which proposes that dislike of driving is derived from negative secondary appraisals.