METHODOLOGY
Data used in this study were extracted from an archived computer aided dispatch database (CAD), automatic vehicle location systems (AVL), inductive loop detectors, automatic traffic counters (ATC) and weather archives. After each incident was assigned to a traffic recorder, the average hourly traffic volumes for the specific hour of each incident were extracted from the ATC data and associated with the incident as the normal flow for that time of day and location.

To determine the capacity reduction caused by each incident, the researchers used the Highway Capacity Manual estimates. The length of each incident was determined by the difference in hours between the time when the incident was first confirmed and the time when it was cleared. Vehicle hours of delay was then estimated for each incident. The costs of delay were calculated using fuel consumption and time.

FINDINGS
Because of lack of data prior to the program's inception, it is impossible to determine exactly how much COMET reduces the average duration of incidents. However, the researchers show the vehicle hours of delay, fuel
consumption, time cost and total cost for the actual duration of the incidents. In addition the study shows the costs if the duration of each incident were increased by 1, 5, or 10 minutes.

• Making the assumption that without COMET each incident would increase in duration by 1 minute then the cost of delay increases by $1,422,618, or roughly twice the cost of operating the Comet program for one year.
• Researchers estimate the break even point for the program would be an average reduction of about 30 seconds per incident, or the $711,300 annual operating cost.

The researchers note that there are numerous other benefits derived from the IR program that have not been quantified due to limitations in the study scope and limitations in data collection feasibility. By reducing delay, IR reduces the total exposure of motorists to congestion. Not only is this measurable in terms of delay cost, but also in terms of avoided crashes and pollution. Often the IR responder is first on the scene after a crash and is able to provide first aid that may save a life or reduce the total health care costs associated with an injury.

This study also successfully demonstrates the use of archived data from multiple sources as an evaluation tool.