Data Analysis: Structural Equation Models

Structural equation models (SEMs) include a broad range of data analysis techniques to study patterns of (inter)relationships among several variables, including correlations, structural dependences and loadings of indicators on latent constructs. The flexibility of the tool can be exploited to quantitatively and empirically test virtually any theory or research hypothesis, and it is at the basis of the popularity of the method in social sciences. SEM started being used at the end of the 20th century in the transport sector and it is now commonly applied in different fields, ranging from travel-activity budgeting to travel satisfaction, safety behaviors, and travel demand management. Building on a discussion of the advantages and drawbacks of the methods also in relation with other analytical approaches, a discussion of SEM best ambits of use and likely future developments for transportation research is developed.