InTransition Magazine
Article URL:
InTransition Magazine : Transportation Planning, Practice & Progress

Archive Edition

Archived editions: 

Point-Counterpoint: Walking Distance to Transit

Editor's Note: An item in last issue’s Research Exchange section titled “Pedestrians Walk Farther Than Expected” discussed an award-winning study that reported commuters were willing to walk more than the quarter- to one-third of a mile distance to rail stations conventionally viewed as the maximum. After reading our coverage and reviewing the study (“How Far, by Which Route, and Why? A Spatial Analysis of Pedestrian Preference”) in its entirety, a transportation consultant wrote to InTransition to challenge its findings. We contacted the study’s authors, who agreed to review these criticisms and offer their response.

Flawed Methods Skewed Results

Recent reports that passengers walk further to transit than commonly assumed were startling.  They implied either that conventional wisdom was wrong or that the obesity generation walks further than their parents and grandparents did. Support for neither conclusion is found in “How Far, by Which Route, and Why.”

The objective of walking distance research is to measure how far passengers are willing to walk. Typically, boarding passengers are surveyed to learn where they originated, usually a residence. The distance walked is then ascertained.

Two separate inputs dominate survey results. One is the passengers’ propensity to walk. The other is surrounding land use.

Land use influences the number of origins in two ways. First is varying density. Origins increase at apartment or condominium complexes and decrease where non-residential uses like parking lots, shopping centers and parks exist. To remove distortions caused by varying density, one divides the number of respondents at each distance by the number of residences at that distance.

No such adjustments were made in the study. Surveys were performed at five rail transit stations. Two stations accounted for 60 percent of the usable surveys. Within the first quarter-mile of one (37 percent of responses), about 40 percent of the land had no residences. Within the first quarter-mile of the second (24 percent of responses), 75-80 percent of the land had no residences. The absence of residences skewed results toward long distances. Had the data been adjusted for density, the aggregated results would have been substantially shorter, in line with expectations.

The second way that land use influences the number of origins is by area. At uniform density, the number of origins increases with the square of distance. For example, the area in the second quarter-mile contains three times that in the first quarter-mile. Survey origins are so widely dispersed at the tail of the walking distance distribution that nobody should expect developments there to be transit-oriented.

Transit market (or modal) share is more useful than walking distance in showing how far transit-oriented development  extends. A proxy for modal share is found by dividing the number of respondents at each incremental distance by the associated area. The new distribution, a narrow, bell-shaped curve, drops off more steeply with distance.

The transit modal share curve reinforces the rule of thumb long used in planning that rail station influence extends only about a half-mile. The new research does not invalidate this rule.

—Robert R. (Bob) Piper, Ph.D., is currently semi-retired as a transportation consultant.

Study Wasn’t Meant to Represent National Average

We appreciate Dr. Piper’s interest in our research. However, he misinterpreted the results and our conclusions in several key ways. Thus, his critiques do not directly apply to our research, though he raises some larger issues important to understanding pedestrian behavior.

First, Dr. Piper claims that our results are skewed towards longer walking distances because at some stations there were few residences directly adjacent to the rail station. However, we have not claimed that our results represent any sort of national average walking distance to rail stations. Instead, we explain that our results show that a substantial number of people who live as far away as a half-mile will walk to a rail station.

Second, Dr. Piper refers to a rule of thumb of a half-mile planning radius for transit stop planning. When one looks to the literature on how far people are willing to walk, there is very little evidence-based information. However, our reading of the literature and knowledge of the transportation planning field suggests that planners tend to think in terms of a quarter mile, not a half mile. Our study provides evidence that the half-mile distance is in fact an appropriate planning radius for pedestrian trips to transit stations, since plenty of people appear willing to walk that distance.

Finally, we do not suggest that the “obesity generation” walks further than their grandparents’ generation—our study did not analyze how walking distances may have changed over time. We do suggest, however, that the distance people are willing to walk to rail is further than commonly thought, and that perhaps if planners plan for these extended distances, communities could make a bigger dent in the obesity epidemic than if planners restrict pedestrian planning to a quarter mile or less.

We invite InTransition readers to review the full report that documents our research findings, “How Far, By Which Route, and Why? A Spatial Analysis of Pedestrian Preference,” which can be found at http://transweb.sjsu.edu/mtiportal/research/publications/summary/0606.html. Also, an article summarizing the survey results is forthcoming in The Journal of Urban Design.

—Marc Schlossberg, Ph.D., Asha Weinstein Agrawal, Ph.D., and Katja Irvin, MUP, are the authors of “How Far, By Which Route, and Why? A Spatial Analysis of Pedestrian Preference.”

Return to this Issue