Causation versus prediction in travel mode choice modeling

This study discusses and analyzes the difference between causal and predictive modeling to model travel mode choice. Causal modeling is expressed through causal discovery and causal inference, used to extract causal relationships in mode choice decision making and estimate causal effects between variables. Predictive modeling is expressed through artificial neural networks. When modeling travel mode choice in three Chicago neighborhoods, we find that both causal and predictive modeling approaches perform well and are useful for their modeling purposes. We also note that the study of mode choice behavior through causal modeling is under-explored while it could transform our understanding of mode choice behavior. Further research is needed to realize the full potential of these techniques in modeling mode choice.