IDENTIFYING THE PROBLEM DRIVER FROM STATE DRIVER RECORDS

THE PRELIMINARY FINDINGS REPORTED HERE ARE THE RESULT OF A TRIAL ANALYSIS PERFORMED ON A SAMPLE OF 1071 DRIVER RECORDS OBTAINED FROM THE DRIVER FILES MAINTAINED BY THE MICHIGAN DEPARTMENT OF STATE. AS SUCH, THEY REPRESENT ONLY THE FIRST STEP TOWARD ACHIEVING A LONG TERM OBJECTIVE. THE RESULTS OF THIS PHASE OF THE ANALYSIS ARE SHOWN. THE MOST SIGNIFICANT FACTOR FOUND TO RELATE TO NUMBER OF ACCIDENTS WAS TOTAL CONVICTIONS. THE TOTAL SAMPLE OF 1071, HAVING AN AVERAGE OF 0.42 ACCIDENTS PER DRIVER, WAS DIVIDED INTO: A GROUP OF 749 DRIVERS WITH 0 OR 1 CONVICTIONS AND AN AVERAGE OF 0.25 ACCIDENTS PER DRIVER AND A GROUP OF 322 DRIVERS WITH 2 OR MORE CONVICTIONS AND AN AVERAGE OF 0.80 ACCIDENTS PER DRIVER, A VALUE OVER THREE TIMES THAT OF THE FIRST GROUP. THE FINDINGS WOULD SUGGEST THE EXISTENCE OF A GROUP OF GOOD DRIVERS (I.E. NO CONVICTIONS) WHO ARE MUCH LESS INVOLVED IN ACCIDENTS THAN IS THE DRIVER POPULATION AS A WHOLE. OF THE REMAINING DRIVERS, CERTAIN GROUPS HAVE MANY MORE ACCIDENTS THAN DOES THE POPULATION AS A WHOLE. IT WOULD SEEM REASONABLE THAT OUR EFFORTS SHOULD THUS BE CONCENTRATED ON IDENTIFYING THE PROBLEM DRIVERS AMONG THOSE DRIVERS HAVING CONVICTIONS. IF THESE GROUPS CAN BE MADE SMALL ENOUGH, INDIVIDUAL ATTENTION CAN BE GIVEN TO THESE DRIVERS. THE RESULTS OF THE ANALYSIS WERE USED TO FORMULATE AN INITIAL MATHEMATICAL DRIVER PREDICTION MODEL THAT COMPUTES A SCORE WHICH IS RELATED TO A PREDICTED NUMBER OF ACCIDENTS. THUS, DRIVERS CAN BE RANKED ACCORDING TO THEIR RELATIVE LIKELIHOOD OF HAVING AN ACCIDENT. THE GROUP OF DRIVERS IN THE HIGHEST RANKINGS ARE THUS IDENTIFIED BY THE MODEL AS PROBLEM DRIVERS. /AUTHOR/

• Supplemental Notes:
  • No 1, 7 pp
• Corporate Authors:

  Highway Safety Research Institute

  Huron Parkway and Baxter Road
  Ann Arbor, MI  United States  48109
• Authors:
  • Carlson, W L
• Publication Date: 1968-5

Subject/Index Terms

• TRT Terms: Behavior; Crash investigation; Crash rates; Driver performance; Drivers; Forecasting; High risk drivers; Mathematical models; Personnel performance; Problem drivers; Traffic conviction; Traffic crashes
• Subject Areas: Highways; Safety and Human Factors;

Filing Info

• Accession Number: 00222714
• Record Type: Publication
• Files: TRIS
• Created Date: Jun 18 1994 12:00AM