LONG-TERM EFFECTS OF CONFIRMATION LIGHTS AND FACTORS THAT LEAD TO RED LIGHT RUNNING

Abstract

Red Light Running (RLR) is a safety concern for communities nationwide. The Federal Highway Administration (FHWA) reported a total of 676 fatalities in 2009 were due to RLR. The Insurance Institute for Highway Safety (IIHS) reported that more than half of RLR crash fatalities are other than the driver (pedestrians, occupants, etc.). There are many strategies to mitigate RLR violations that fall in the category of engineering, enforcement, or education. This dissertation focused on confirmation lights which are a low-cost countermeasure that enhance enforcement at four-approach intersections. Confirmation lights were deployed at two intersections in Overland Park, Kansas. Traffic was observed at the treatment sites, nearby signalized intersections (spillover), and control sites. Traffic was recorded before deployment, one-month after, and three-months after deployment. A total 14 intersections were recorded during the morning peak hours (7 a.m. to 9 a.m.) and the afternoon peak hours (4 p.m. to 6 p.m.) for a total of 583 hours of traffic video. A test of proportions showed that overall the confirmation lights did not significantly reduce RLR violations. A violation analysis showed that there was a global increase in RLR violations after deployment, showing that other factors were involved in the increase of violations observed. Time into the red analysis showed that the majority of RLR violations occurred within one second into the red. The negative binomial regression model re-affirmed that the confirmation lights were not a significant factor in the RLR violations observed. The model showed that lane volume, presence of a right turn lane, and traffic movement (left or through movement) were significant factors.