Car drivers’ perceptions of the quality of alternative travel modes have been identified as a barrier for including these alternatives in their choice sets. The present study investigated the accuracy of car drivers’ perceptions of public transport (PT) travel time and the potential effect of these perceptions on choice sets. A sample of car drivers was intercepted on the main corridors to Amsterdam, the Netherlands, using video recognition of license plates, and was sent a questionnaire asking (among other questions) whether they could have made the specific trip by PT and their estimate of the door-to-door travel time by PT. Objective travel times were obtained from route-planning software. 21,335 questionnaires (31\%) were returned. About 10 percent did not report PT travel time for their car trip, largely car drivers who did not perceive PT as an alternative. The mean ratio of perceived travel time by PT to reported travel time by car was 1 : 2.3. About half the difference was due to distorted perceptions, and the ratio reported depended strongly on their PT use. Analysis of associations between choice set and characteristics of traveler and trip showed that if perceived PT travel times were more accurate a substantial number of car drivers would include PT in their choice set. Actual changes in behavior might be much smaller.