GPS-Based Household Interview Survey for the Cincinnati, Ohio Region

Past Household Travel Surveys (HTS) in the United States have only piloted small subsamples of Global Positioning Systems (GPS) completes compared with 1-2 day self-reported travel inventory diaries. The Ohio Department of Transportation Research Division, in cooperation with the Ohio-Kentucky-Indiana (OKI) Council of Governments, in 2009 initiated the first exclusively GPS Household Travel Survey (HTS) (e.g., no diaries). In addition, the ODOT GPS-only HTS is one of a few initial US studies to successfully use a personal devise GPS. Most past subsample studies have installed GPS in vehicles for recording and comparison of auto travel only. Additionally, this study was one of the first completed address-based HTS, which includes cell only households, with Internet recruit available. This research documents the protocols used for GPS deployment to all members of a household over 12 years old for a three-day recording period. Simplified diaries were provided for children under age 12. The main survey commenced in August of 2009. Recruitment of over 5,000 households progressed over a twelve month period, with proportional distribution. The highly stratified address-based sampling method used is described as well as the forms and materials, response rates for recruitment, and the compliance rates and experiences with households. Also described are the logistical challenges of getting 4+ persons households to all carry a GPS unit on a concurrent day and the challenge of loss rates on GPS units not returned. This research documents the Internet-based prompted recall (PR) verification method used and its outcomes. The PR was essentially used to impute mode of travel and trip purpose from GPS recordings. These imputation methods and findings are described. Also presented and discussed is the depth and degree of accuracy of core trip variables obtained using GPS-only methods, including modes, travel time and distance, and purpose at locations. Lessons learned are recorded. The primary conclusion to be drawn from this research is that it is feasible to undertake a GPS-only household travel survey, achieving a high standard of representativeness for the sample, while imputing mode and purpose at a sufficiently accurate level to support modeling work. The high level of accuracy attained in this survey for imputing mode and purpose with 96 percent on mode and around 90 percent on activity (other than detailed breakdowns of the “other” category) is far superior to self-report surveys. The richness of the “ground-truthing” of time, location, distance, speed, and route information from this survey surpasses what can be achieved from any other form of survey.