Transferring Telephone-Based National Household Travel Survey to the Internet

Transfer of the National Household Travel Survey (NHTS) and regional travel surveys to the Internet is inevitable, partly because online surveys offer an efficient means of collecting data. Behavioral surveys are increasingly being offered in multiple media and giving respondents the choice of filling out the survey by using the Internet, the telephone, or other means. This study reports experience and lessons learned from transforming the computer-assisted telephone interview methodology used for the NHTS to an Internet-based method for university students, who have ubiquitous access to the Internet. The study reflects innovations in survey research methodology in the context of surveying students at large universities in Virginia. Comparisons of two rounds of behavioral surveys conducted in 2009 and 2010 are provided, and statistical models that quantify trip underreporting are developed. In an attempt to mimic closely the NHTS instrument, the Internet instrument had a relatively high survey response burden. On the basis of analysis of trip frequencies, trip underreporting was suspected in the first round of surveys. Documented here are improvements in survey design that were intended to lower the survey response burden and reduce trip underreporting. Statistical models are estimated to quantify how changes in the instrument captured more trips. Results show that survey design improvements can encourage students to recall and report 15% to 20% more trips. The study also found that, consistent with earlier studies, discretionary trips and shorter trips were more likely to be underreported. The results from this study are valuable for future regional and national survey implementations.