An Analysis of Driver Inattention Using a Case-Crossover Approach On 100-Car Data: Final Report

Using the 100-Car Study database, two analyses were conducted: re-calculation of odds ratios (ORs) using a case-crossover baseline, and characterization of secondary task engagement in real-world environments. First, ORs were recalculated for drowsiness, secondary task engagement, and total time eyes-off-road (TEOR) using conditional logistic regression. The results suggested that drowsiness (OR 38.7; CI 26.4 – 56.8), tasks with >2 eyeglances away from the forward roadway or >2 button presses (OR 2.3; CI 1.3 – 3.1), and tasks with 1-2 eyeglances and/or 1-2 button presses (OR 1.4; CI 1.1-1.7) significantly increased crash/near-crash risk. The results also indicated that total TEOR of 2 s or greater during a 6-second task period increased crash/near-crash risk (OR 1.6; CI 1.3 – 2.0) and a 3 s or greater total TEOR over a 15-second task period significantly increased crash/near-crash risk (OR 1.3; CI 1.1 - 1.6). These OR point estimates are lower than the results obtained using a case-control; however, they are still statistically significant in both analyses indicating that these behaviors increase risk. The second analysis assessed secondary task duration, frequency, and the relationship of task duration to total TEOR. Results indicated that drivers in the 100-Car Study engaged in secondary tasks 23.5 percent of the time that they were driving, approximately 40 percent higher than indicated in previous research. Secondary tasks that were found to be both of long duration and with a high percent of total TEOR (such as applying makeup) had crash/near-crash risk ratios that were not significantly greater than 1.0. In contrast, analysis of all secondary tasks of long duration, including those with lower total TEOR (such as talking with passengers), had OR values significantly less than 1.0. The results from both of these analyses suggest that in-vehicle display designers need to assess and be cognizant of the total TEOR for in-vehicle displays for two reasons: 1) a brief total TEOR will increase risk for drivers, and 2) total TEOR is associated with involvement in crashes/near-crashes. Assessment tools like the “15-second rule” developed by the Society of Automotive Engineers (SAE) or the ‘2-second rule’, developed by the Alliance of Automobile Manufacturers (AAM) have not been shown to be associated with or predictive of crash/near-crash risk for any type of task. Thus, the authors argue that total TEOR should be included in the list of assessment tools for in-vehicle display designers.

• Record URL:
  http://www.nhtsa.gov/DOT/NHTSA/NVS/Crash%20Avoidance/Technical%20Publications/2010/811334.pdf
• 
• Corporate Authors:

  Virginia Polytechnic Institute and State University, Blacksburg

  Virginia Tech Transportation Institute
  3500 Transportation Research Plaza
  Blacksburg, VA  United States  24061

  National Highway Traffic Safety Administration

  Human Factors/Engineering Integration Division
  Washington DC,     20590
• Authors:
  • Klauer, Sheila G
  • Guo, Feng
  • Sudweeks, Jeremy
  • Dingus, Thomas A
• Publication Date: 2010-5

Language

• English

Media Info

• Media Type: Web
• Edition: Final Report
• Features: Appendices; Figures; Glossary; References; Tables;
• Pagination: 148p

Subject/Index Terms

• TRT Terms: Attention lapses; Drivers; Drowsiness; Highway safety; Risk assessment
• Uncontrolled Terms: Case-control studies
• Subject Areas: Highways; Safety and Human Factors; I83: Accidents and the Human Factor;

Filing Info

• Accession Number: 01357954
• Record Type: Publication
• Report/Paper Numbers: HS-811 334
• Contract Numbers: DTNH22-00-C-07007
• Files: HSL, TRIS, USDOT
• Created Date: Dec 1 2011 9:56AM