It can be imagined that a reward may be a far more popular policy instrument than the traditional taxation approach towards containing externalities, usually presented in public economics literature. Given the implied policy potential, we conducted an extensive reward experiment in real world conditions on a congested motorway corridor in the Netherlands. In this paper the data collected in the experiment are used to estimate a number of discrete choice models that describe the commuter's behaviour with respect to departure time choice as well as transport mode choice. We apply the traditional scheduling approach where rush-hour travellers trade off travel time for schedule delay disutility, and we study how this equilibrium shifts upon the introduction of a reward. The results of our analysis provide a clear indication that a reward can be used as an effective policy instrument. The participant's behaviour implies that the shadow prices of schedule delay are close to constant over time, a finding which is in line with the classic assumptions in the literature. Preferences for different departure times for car trips within the rush-hour are found to be correlated. This indicates that shifting departure time is likely to be a more important behavioural response to policies for congestion relief than a modal shift or teleworking. If we compare the relative as well as the absolute sizes of the different valuations of schedule delay early, schedule delay late and travel time, we find that they are comparable to past findings in the literature. © 2012 Elsevier Ltd.