Quantitative Analysis of Latent Variable and Its Integration with Freight Mode Choice Behavior Model

This paper explores the quantitative method of latent variable and the impact of convenience on freight mode choice behavior. The structural equation model is established for quantification of convenience. Base model, incorporation model with latent variable, and comparison model are estimated to reveal freight mode choice behavior. These three models are based on the multinomial logit (MNL) and mixed logit (ML) methods. The data comes from the questionnaire experiment in the Yangtze River Delta. Results show that convenience significantly impacts shippers’ choice, and there is a large degree of heterogeneity for convenience among shippers. The model performance is best when the latent variable is integrated with the ML model. In addition, the structural equation model is an effective method to quantify the latent variables, and convenience is negatively correlated with pre- and end- haulage distance, response time, and cargo damage rate.