The paths from walk preference to walk behavior: Applying latent factors in structural equation modeling

A structural regression model has been developed to explore the relationship among key factors in the explanation of utilitarian walking. The model examines the relationship between and among unobserved, or “latent,” factors that reflect (1) the values and preferences operant at the time of residential selection; (2) the urban form of the neighborhood; (3) the urban form of the residence; (4) the level of auto dependency; and (5) the extent to which the neighborhood is found satisfactory by the participant and those whose opinions he/she respects. The model allows the detailed examination of the paths from initial inclination toward a neighborhood with walkable destinations, through a series of mediating unobserved factors, each of which might either impede or facilitate the adoption of utilitarian walking. Analysis of the model results shows that values and preferences held at the time of residential selection are directly associated with the amount of utilitarian walking undertaken and indirectly associated, through their influence, on the choice of the built environment and the extent of auto orientation. The model is designed to facilitate the observation of the manner in which the various factors interact.