SAN DIEGO - A new study from DriveCam underscores the cost of poor driving habits. In the study, drivers with multiple collisions represented only 0.4 percent of all transit drivers. But more than 70 percent of these drivers were consistently cited for four poor driving behaviors.

Additionally, multi-collision drivers showed a significantly higher frequency across all poor driving behaviors. By recognizing these behaviors as leading indicators of future risk, fleet safety and risk managers can coach and train drivers to avoid costly outcomes, DriveCam said.

DriveCam specializes in addressing the causes of poor driving by combining data and video analytics with real-time driver feedback and coaching.

The study, part of the company's "Driving Insights" series, focused on the analysis of over 20,000 drivers in the transit industry who were active between June 2009 and June 2010. The behaviors with the largest observed variance between multi-collision and non-collision transit drivers were:

• Following Too Close - 1.25 to 1.75 seconds (58 percent of multi-collision drivers, 33 percent of non-collision drivers). Additionally, 1 second or less (24 percent of multi-collision drivers, 10 percent of non-collision drivers)

• Failed to Keep an Out -- 30 percent of multi-collision drivers, 11 percent of non-collision drivers

• Not Looking Far Enough Ahead -- 64 percent of multi-collision drivers, 48 percent of non-collision drivers

• Rolling Stops -- 32 percent of multi-collision drivers, 19 percent of non-collision drivers

DriveCam's video event recorder provides the ability to identify when a collision has occurred. DriveCam does not analyze collision events for legal reasons; however, extensive video event review of a driver's prior non-collision poor driving behaviors allows DriveCam to better understand those behaviors that act as leading indicators of collisions. Identifying these behaviors provides safety managers a focused direction in coaching and training.

DriveCam's "Driving Insights" series taps the company's extensive database of driving events from over 3 billion driving miles.