A comparative analysis of cross-sectional study and natural experiment in rail transit-travel behavior research: A case study in Wuhan, China

Abstract

There has been a global increase in investment in rail transit, driven by its potential to enhance transportation efficiency, reduce air pollution, and stimulate economic growth. Both cross-sectional studies and natural experiments have contributed to the growing body of evidence supporting these claims. While natural experiments are commonly preferred for evaluating the impact of rail transit, cross-sectional studies remain popular due to their ease of data collection. However, there is a scarcity of studies that compare these two approaches using the same dataset to assess the robustness of cross-sectional studies. Using a two-wave panel dataset from Wuhan, China, this study used both cross-sectional and natural experimental analyses to examine the relationship between urban rail transit and travel behavior. The study attempted to enhance the credibility of the cross-sectional analysis by controlling for confounding variables and by combining it with the propensity score matching (PSM) method, respectively. The results revealed that the cross-sectional analyses could produce similar results, when setting a more stringent significance level. The findings suggested that well-designed cross-sectional studies can be reliable and represent a cost-effective alternative to resource-intensive natural experiments.