In Friday’s class, we discussed several issues associated with the four-step transportation planning model. These limitations ranged from specific functional flaws (iterations of variables, difficulty of data collection, etc.) to more general statistical modeling issues (rational choice assumption, complete information assumption, etc.). As the class progressed, the relevance of the four-step transportation planning model came increasingly into question. I will discuss one particular issue associated with the model, the utilization of intangible variables in utility functions.
In the four-step transportation model, several utility functions are used to transition between the mode choice and traffic assignment steps. These functions intend to display how the aggregate values of intangible variables (time value, comfort, aesthetics, etc.) influence an individual’s mode choice. The theory behind mode choice is sound: by using census data, surveys, and other types of measurement systems, transport engineers can provide values to these qualitative variables as well as possible. Certainly, a best-attempt measure is better than a lack of inclusion of these variables altogether. However, it is possible that the current system simply isn’t good enough.
In gathering data for these variables, one must rely on subjective data that doesn’t necessarily reflect population truths. Surveys, in particular, are unreliable measures of qualitative variables. When participating in a survey, it is highly likely that an individual will misrepresent components of their decision-making processes. For instance, many people would likely place a high value on the importance of reducing transportation emissions in a survey, but this does not mean that they would change their transportation patterns. As these types of misrepresentations of values and preferences increase, the coefficients assigned to utility equations become inaccurate. This results in flawed mode choice models, and in turn, flawed traffic assignment models.
The question remains: what can be done to enhance the reliability of the four-step transportation model? It is easy to critique the existing system, but it is much more difficult to create practical solutions. It is likely that there are much more effective manners of data collection than surveying, but these methods would involve huge, unrealistic expenses. While the existing system may not be perfect, it seems to be the best available option with the available resources.
Reference:
Transportation & Tomorrow 