Accident Update
By Edward Heeren
Newly available data from the New Jersey Department of Transportation reveals that the implementation of EZPass on the Garden State Parkway has failed to do anything to reduce the extremely high rate of accidents at Parkway toll plazas.
Accidents in the tenth of a mile located at a Parkway toll booth continue to be 5 times as frequent as on an average section of highway located away from a tollbooth. Accidents one-tenth of a mile from the tolls continue to be 4 times as likely as elsewhere, and accidents two-tenths of a mile from the toll booths continue to be twice as likely as they are away from the tollbooths. Even three-tenths of a mile from the tolls, the accident rate is still 43% to 70% higher than it is away from a tollbooth.
In the 7.7 miles of the Parkway located at a tollbooth or no more than three tenths of a mile away from a tollbooth, accidents continue to be 290% as frequent as they are away from a tollbooth.
From 1997 through 2002, there were 5,290 accidents within 0.3 miles of a tollbooth, but only 1,845 accidents on an average stretch of road farther from the tolls. This represents an excess of 3,445 accidents near Parkway tollbooths over the rate that would expected. Disturbingly, the number of excess accidents has actually increased, from 523 in 1999 before EZPass was implemented, to 566 in 2000 during and just after the implementation of EZPass, to 662 in 2001 and 679 in 2002 with EZPass fully implemented and drivers already familiar with it. The number of total accidents near tollbooths has also increased, from 811 in 1999 to 880 in 2000, to 1008 in 2001 to 1036 in 2002. All of the education campaigns, lane striping, and tollbooth reconfigurations have not reduced tollbooth accidents at all, and there has even been a small increase in such accidents.
Technical data related to this investigation is as follows:
- New Jersey crash data is available at http://www.nj.gov/transportation/refdata/accident/index.shtml
- The Parkway is known internally as Route 444, and this is the route number entered on accident reports. A small number of accident reports omit the route number and report only a road name such as "PARKWAY" or "GARDEN STATE PARKWAY". These reports were not considered in the analysis. The web page at http://web.mit.edu/spui/www/nj/4xx.html#444 is one of several places on the net which discuss the secret numbering of the GSP as Route 444.
- Any accident report data indicating hundredths or thousandths of a mile was dropped, and the mile marker was rounded down to the nearest tenth.
- Tollbooth locations were determined by manually examining the Route 444 Straight Line Diagrams at http://www.state.nj.us/transportation/sldiag/00000444-.pdf . For the Cape May and Great Egg tolls, the mile markers in the press release at http://www.gspkwy.state.nj.us/pr091900.htm were used.
- The Raritan toll plaza is located at mile marker 125.6 southbound and 125.8 northbound. An average location of mile marker 125.7 was used in this analysis.
- The program tabulates accidents according to the year in which they were reported, not the year in which they occurred.
- The Union County 1998 data file contained an unusually large number of records with a missing mile marker. (The Union98.txt file contains 159 GSP records with a blank or invalid mile marker out of 763, as compared to 4 invalid records for 1997, 10 invalid records for 1999, and 2 invalid records for 2000.) This results in a sharp decrease in the number of accidents tabulated for Union County mile markers in 1998 that probably does not reflect reality.
- The number of accidents on a particular stretch of road is of course highly dependent on the amount of traffic on that stretch of road, which results in there being far more accidents in North Jersey than in South Jersey. The amount of traffic on any stretch of a few miles north and south of a particular point should be relatively constant, so I have collected the data for the 2.8 miles north and south of each tollbooth and then aggregated this over the 11 barrier tolls to minimize the influence of any unique factors around a particular tollbooth.
- The minimum distance between two barrier tolls on the Parkway is 5.7 miles, between the Bergen and Hillsdale tolls. For this reason, locations farther than 2.8 miles from a tollbooth were not studied, since in North Jersey, some such locations would actually be closer to another adjacent toll area.
- This accident analysis may actually understate the number of tollbooth accidents, since it only considers accidents near barrier toll plazas, and does not consider accidents near ramp toll plazas. Accidents on the ramp leading to the exit 117 toll plaza, known as route 444R, and on the ramp leading to the exit 105 toll plaza, known as route 444S, are not included in the statistics since the route number is not "444".
- The psychological effects of tollbooth congestion and lane weaving on driver's attitudes and behavior once they have left the tollbooth area also cannot be measured by a study such as this.
