Previous research has indicated that safety measures may lead to behavioural adaptation (also termed risk compensation) among road users, partly or completely offsetting the intended safety effects. There is, however, limited knowledge about characteristics of safety measures possibly determining the occurrence of behavioural adaptation. The present study addresses the relationship of driving behaviour to two different kinds of in-car safety equipment, airbags and antilock braking systems (ABS). It is hypothesized that accident-reducing measures like ABS are compensated for to a larger extent than injury-reducing measures like an airbag. On-road unobtrusive measurements of speed, headway, lane occupancy, lane changes, and variability of lateral position were performed on 213 taxis, on the basis of video recordings of traffic travelling to Oslo airport. The behavioural data were matched to questionnaire information collected when the taxis arrived at the airport. In addition to information regarding ABS and airbags, the drivers reported personal background information and answered questions about driving behaviour. Taxis with ABS had significantly shorter time headways than taxis without ABS. There were no relationships with speed, possibly because dense traffic during the observation period may have prevented the drivers from driving at their preferred speed. Simple comparisons also showed fewer lane changes and a lower rate of seat-belt use among drivers of taxis with ABS. However, multiple regression analyses indicated that the latter effects might be explained by driver background factors or by car characteristics other than ABS or airbag. The headway results support the hypothesis of larger compensation for accident-reducing than for injury-reducing measures.