Document Type

Journal Article

Publication Date


Subject Area

operations - traffic, infrastructure - vehicle, ridership - drivers, ridership - behaviour, place - urban, mode - bike


Urban areas, Signalized intersections, Signalised intersections, Red light running, Motor vehicle operators, Michigan, Late exit, Human behavior, Drivers, Detroit (Michigan), Countermeasures, Clearance interval (Traffic signal cycle), Behaviour, Behavior


Red-light running and the associated risk of severe crashes at signalized intersections have been an ongoing concern to many safety professionals. The proportions of crashes that occur due to such driver behavior represent a substantial number of crashes at urban and suburban signalized intersections in Michigan. In the year 2001, crashes related to red-light running in Michigan represented about 28% of the crashes of all severities and 40% of fatal and serious injury crashes occurring within signalized intersections. Many initiatives to discourage red-light running have been used in the United States with various degrees of success. Such programs include public awareness campaigns, automated enforcement programs, and some engineering countermeasures. These initiatives have resulted in some reduction in the overall crash statistics in the communities where such programs have been implemented. A series of evaluation studies was performed in Michigan to test the effectiveness of implementing change and clearance intervals calculated according to ITE guidelines on late exits and red-light violations at nine signalized intersections in the Detroit metropolitan area. The study described used four approaches at four test intersections where engineering treatments have been applied (16 total test sites) and four approaches at five control intersections. Driver behavior was examined at test and control intersections by observing and quantifying the number of vehicles that did not clear the intersection when the cross-street signal light turns green (late exit) and red-light violations. Upon analysis, red-light-running data at the test and control sites did not exhibit a significant difference. However, the test sites showed significantly lower late exit rates compared with the control sites, thus lowering the risk of right-angle crashes.