Built Environment Vs. Self-Selection
Study after study shows that people who live in higher-density, mixed-use neighborhoods tend to walk more and drive their automobiles less, thus reducing the environmental impact of automobile dependence.
But the study "Examining the Impacts of Residential Self-Selection on Travel Behaviour: A Focus on Empirical Findings" wondered whether residential self-selection might confound the idea that the built environment overrides the natural disposition of people leave where they are comfortable. In other words, maybe you could take the people out of the suburbs but not as easily take the habits of the suburbs out of the people.
The paper identified nine approaches used in previous research to empirically address the issue of residential selfselection, and reviewed the empirical findings of 38 studies using those approaches.
"If the key question is, ‘Does the [built environment] have a distinct influence on [travel behavior] after self-selection is accounted for?’, then based on the empirical evidence to date, the answer would have to be a simple and resounding ‘yes’," the authors concluded. "... It is more difficult, however, to assess the strength of the autonomous influence of the [built environment] relative to the influence of self-selection, or even to ascertain whether that autonomous influence is ‘large enough to matter’ on its own terms."
In the end, the researchers found they were unable at this point to specify the nature and extent of the influence of the built environment on travel behavior. That doesn't negate the need to make changes in the built environment that have benefits outsite the travel behavior of residents such as increasing the diversity of available housing options, the authors note. But ignoring self-selection is problematic.
"Given the extensive evidence that has accumulated on the impact of self-selection, we believe it is misleading to present empirical results that do not take that impact into account," the researchers conclude. "Such faulty findings are likely to result in flawed policies, and/or an overestimation of their effectiveness."
This study has been added to the Best Practices.