1 / 22

Measuring Exposure to Vehicle Accident Risk Using the American Time Use Survey

Measuring Exposure to Vehicle Accident Risk Using the American Time Use Survey. Lee Giesbrecht and Jonaki Bose Bureau of Transportation Statistics RITA, U.S. DOT. What Do We Mean by Exposure. Exposure measures form the denominator of accident rates

DoraAna
Download Presentation

Measuring Exposure to Vehicle Accident Risk Using the American Time Use Survey

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Measuring Exposure to Vehicle Accident Risk Using the American Time Use Survey Lee Giesbrecht and Jonaki Bose Bureau of Transportation Statistics RITA, U.S. DOT TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  2. What Do We Mean by Exposure • Exposure measures form the denominator of accident rates • Defining the numerator may be simpler – number of accidents, fatalities, injuries • However, defining exposure or the opportunity to be involved in any of the numerator incidents is not that clear cut – using total population counts is obviously a poor choice TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  3. Defining Exposure • Vehicle miles traveled • Numbers of registered drivers • Numbers of registered vehicles • Time spent on roadways/outdoors TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  4. Why Bother with the Denominator? • Better understanding of the nature of accidents • Are there variations in accident rates of different subgroups? • Are there interactions between subgroup status and factors such as time of day, type of vehicle, urbanicity, etc. • For example even though there are similar numbers of men and women in the population, men tend to drive longer and further, thus increasing their potential exposure TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  5. Literature • Travel survey data has been explored as a denominator measure • Data on miles traveled used in New Zealand and Australia to compute crash rates • Data on miles traveled and travel times used in Canada to compute average speeds for subgroups • Miles traveled from the NHTS and its predecessors has also been explored TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  6. Combination of Two Previous Studies • Bose, J. and J. Sharp. “Measurement of Travel Behavior in a Trip-Based Survey Versus a Time Use Survey: A Comparative Analysis of Travel Estimates Using the 2001 National Household Travel Survey and the 2003 American Time Use Survey”. Presented at the American Time Use Survey Early Results Conference, December 8-9, 2005. • Giesbrecht, L., J. Bose, and M. P. Cohen. “Developing Exposure Measures to Evaluate Vehicle Accident Risk”. Proceedings from the Survey Research Methods Section, American Statistical Association, 2005. TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  7. Our Objectives • Continue to highlight the NHTS as a source of data, especially since it includes – • children under five • vehicle occupancy • time of travel, distance and mode • Compare the ATUS with the NHTS as a source of exposure data TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  8. Background on the 2001 NHTS • Nationally representative RDD survey • 26,000 households, 60,000 persons • Daily and long distance trips, vehicles • All household members • Randomly assigned travel day • Approximately 250,000 daily trips, 45,000 long distance trips (in the national sample) • Data collection within 6 days • Proxies accepted TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  9. Background on the 2003-4 ATUS • Continuous survey • Nationally representative telephone survey based on area probability sample (retired CPS panels) • 21K people 15 years old and above • Data for one household member • Data on household and personal characteristics, as well as on all activities • For each activity data collected on start and end times, detailed nature of activity, who else was part of the activity • Travel was considered a primary activity TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  10. Similarities • Nationally-representative surveys • Data collected over the phone • Data collected for a 24-hour period (4 am start) • Data on travel activities/trips collected at a person level • NHTS defines trips as movement from one address to the other. In the ATUS movement from one address to the other is coded as travel, regardless of any other activity TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  11. Differences • Trip-based (NHTS) vs. activity-based (ATUS) • NHTS data collected for 2001, 2003-4 for ATUS • NHTS allowed proxies; ATUS did not • NHTS used a travel diary; ATUS used recall • Assigned reporting day constant in NHTS; could rotate to other weeks in ATUS • 6-day collection period for the NHTS; 1 day for ATUS • NHTS collected data for ALL household members including children; One person 15+ for ATUS • More transportation-related details collected for NHTS; more contextual activities in ATUS • RDD sample used for the NHTS, area probability for ATUS TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  12. ATUS is More Limited than NHTS • No information on household vehicles (vehicle type, age, mileage) • Larger sampling errors due to smaller sample size • Lower estimates of time spent in travel, but this may not necessarily be a limitation • Bounding activites may lead to better estimates • Sample day substitution may lead to bias toward days with less travel TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  13. A Few Details on the Analysis • Used data from FARS for the numerator for illustration purposes • Appropriate weights were used in estimates (used national NHTS sample and weights) • Complex survey design was taken into consideration when estimating sampling error TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  14. Minutes of Vehicle Travel per Day TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  15. Minutes of Vehicle Travel per Day by Age TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  16. Minutes of Walking per Day by Age TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  17. Minutes Bicycling per Day TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  18. Fatalities per Million Minutes of Vehicle Travel TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  19. Fatalities per Million Minutes of Vehicle Travel by Sex and Age TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  20. Pedalcyclist Fatalities per Million Minutes Bicycling per Day TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  21. Possible Reasons for Differences • Use of travel diary in NHTS; specific probe for walk and bike trips • Reconciliation of trips among household members • Exclusion of exercise and dog walking trips from ATUS • Substitution bias in ATUS– people might be more likely to respond when they are less busy • Lack of other activities to bound trip times TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

  22. Summary • These were just a few results to illustrate the utility of ATUS in this area • Also points out the higher utility of the NHTS for this purpose because of richer data and a larger sample – the ATUS is not a very good replacement TRB 2007 Annual Meeting – Travel Survey Methods Committee Meeting

More Related