Reduction in car–bicycle conflict at a road–cycle path intersection: Evidence of road user adaptation?

The influence that cycle paths have on accident numbers is not clear, probably because the unique contexts into which each path is introduced have not been sufficiently accounted for. This study aims to explore an important addition to the list of variables that can be used to describe these contexts, that of learning and adaptation by those road users regularly encountering the cycle path. A total of 57 h of video registrations were made of yielding and conflict events between cars and bicycles at a Norwegian road–cycle path intersection, at 2 months, 4 years and 10 years following its introduction. A significant decrease in overall conflict levels was found 4 years after path introduction, with a further decrease 10 years after path introduction, although the decrease between 4 and 10 years was not significant. The decrease in conflicts was mirrored by an increase in the number of car–bicycle yielding events not resulting in conflict. The yielding and conflict incidents registered were classified according to eight possible intersection scenarios, and compared over time to the yielding behaviours of drivers and cyclists in each scenario. Those scenarios in which no conflicts occurred in over 28 h of registrations 10 years following path introduction were the very same scenarios in which changes in driver yielding behaviour were apparent between 4 and 10 years following path introduction. Together these results have important implications both for cycle path evaluations and the modelling of road user behaviour at junctions.

Research highlights
► Evidence of road user adaptation after introduction of a cycle path–road junction.
► Significant reduction over time in the share of near collisions between cars and bicycles.
► Near collision events lowest in those scenarios where change in road user yielding behaviour evident.