Exploring Heterogeneous Structural Relationships Between E-Shopping, Local Accessibility, and Car-Based Travel: An Application of Enriched National Household Travel Survey Add-on Data

This study analyzes structural relationships between information and communication technology (ICT), e-shopping, local accessibility, and travel intensity in the U.S. state of Georgia. Beginning with the 2017 National Household Travel Survey (NHTS), we enriched the data by exploiting the add-on program. After obtaining contact information for the Georgia NHTS participants who expressed willingness to take a follow-up survey, we conducted another survey with those participants. In addition, we appended land use characteristics associated with respondents' home locations. To test the hypothesis of possible heterogeneity in structural relationships, we applied a mixed structural equation model. We corroborated our hypotheses that the use of ICT has a positive impact on e-shopping frequency, whereas local accessibility reduces e-shopping frequency. In addition, we identified two meaningful segments exhibiting different structural relationships. The smaller segment (12.7%) showed complementary effects of e-shopping on car-based travel intensity, but local accessibility did not have a significant impact on car-based travel intensity. The dominant segment (87.3%) presented a negative impact of local accessibility but no meaningful effect of e-shopping frequency on car-based travel intensity. A post-hoc analysis revealed meaningful differences in demographics between the two classes. The study suggests that an assumption of homogeneous structural relationships between e-shopping and car-based travel intensity and between local accessibility and car-based travel intensity may not hold. The paper discusses benefits and practical issues concerning fusion of the NHTS data with other data sources. In addition, it stresses the potential of the NHTS add-on program and notes some recommendations.