September 19, 2003 | WebMemo on Smart Growth
A new report from Smart Growth America and the Surface Transportation Policy Project, Measuring the Health Effects of Sprawl, links growing obesity concerns with sprawl. The report's findings, however, fall short of supporting this conclusion.
Instead, this is another attempt by the report's sponsors to spin research showing only trivial weight differences between city and suburban residents into a national crisis requiring land use restrictions.
From Boston to Canberra, the headlines told the message: "America's expanding obesity tied to sprawl." The study purports to demonstrate that people living in more sprawling, suburban counties are fatter than people who live in more dense central cities. The principal cause of this reported difference is that people walk more in cities than suburbs, where there is greater dependency on cars.
But what may be most revealing about the report is not the report's conclusions, but the apparent desperation of the sponsors who felt obligated to publicize research showing that there are only trivial weight differences between cities and suburban residents. A better headline would have been: "Researchers demonstrate statistically significant insignificance."
Using survey data from the Centers for Disease Control from 445 counties around the country, the researchers then applied a county sprawl index to the data and developed a computer program to predict the weight and obesity of people in the counties by degree of sprawl, and then compare the prediction to the actual. After so manipulating the data, their findings proved to be "statistically significant", a mathematically determined condition meaning that the formula they developed appears to have a high probability of being reliable for predicting purposes.
But statistical significance is not enough: the results must also be material, which in this case they are not. The researchers predict, for example, that the average Cook County (Chicago) resident weighs 0.9 pounds (15 ounces) less than suburbanites in Lake County, Illinois. (see chart at end) In effect, sprawl in Chicago can be said to add less than a pound to a suburbanite's weight. The researchers discover similarly trivial findings across the country. For the country as a whole and comparing citizen weight in the 25 counties at either extreme of sprawl and compactness, 19.2 percent of residents in the least sprawling communities were obese, while 21.2 percent in the most sprawling were obese. Even the Rutgers University researcher had to admit - with or without pun intended -- that "Those aren't huge differences."
But the panic driven headlines, wide coverage, and superficial stories beneath them indicated that few reporters bothered to look beyond the sponsors' press releases. Washington Post Writers Group columnist Neal Peirce characterized the report as research "scientifically linking the United States' pattern of highway-driven, sprawling, road-dependent development with the alarming epidemic of rising weight and obesity that the country's been experiencing."
But as public relations experts have learned, impressions are often more important than reality. So to market findings of a connection between sprawl and obesity, however insignificant, establishes a public policy connection that is anything but insignificant. Indeed, the project sponsors can even get away with characterizing their finding of "small impact," with the confidence that reporters and columnists will never read far enough to see the qualifying statements. This "big headline, little story" model is an effective strategy for an anti-sprawl lobby that would force us out of our cars and make us live on top of one another.
And that's precisely the point. Most Americans live in urban population densities of 4,000 per square mile or less. The "walkable communities" that the anti-sprawl lobby want us to create and live in exist principally in the four densest boroughs of New York City (Manhattan, the Bronx, Brooklyn and Queens). Here, population densities are from 20,000 to 70,000 per square mile. This is where the Smart Growth America-STPP researchers predict the lowest obesity rates.
Even there, however, the predicted weight difference between Manhattan (70,000 per square mile) and Queens (20,000 per square mile) is more than the difference between Queens and the most sprawling suburban county in the New York metropolitan area. Indeed, according to New York City's Department of Health and Mental Hygiene, 24 percent of the adults living in the walkable Bronx were obese compared to 21 percent nationwide, which is also precisely the proportion of obese people found in walkable Brooklyn.
The report, however, goes on to suggest that planners should design urban areas for walkability to address the problem of rising obesity. But what they don't say is that Manhattan style walkability cannot be achieved without Manhattan densities. If urban areas were the playthings of planners who could decree a corner grocery store here, a wine shop there and a Starbucks across the street, perhaps it would be different. But it is not. In fact, Manhattan is a wonderful place to live for people who want to live there. But so are suburban communities like Simi Valley, Maple Valley, Chesterfield and Apharetta. Approximately 1.5 million people choose to live in Manhattan, and 7.5 million in the four dense boroughs of New York City. At the same time, more than 30 times as many people choose to live in places that have far more room. Public policy should not assume the nanny-role of requiring community designs that conspire to increase walking, particularly when it leads to weight differences of such trivial magnitude.
And, while the report's findings are trivial, the analytical gaps are big enough to drive a subway train through. Researchers did not, for example, consider the relationship between income and weight, despite some evidence that the connection between the two is more significant than the one between weight and suburbs. In New York City, for example, 22 percent of people earning less than $25,000 per year were obese, compared to 15 percent of those earning over $50,000. This omission might have been justifiable if the data were not available, but it is, in the very same dataset the researchers used. Inclusion of the income variable is likely to have made the sprawl/weight results even less consequential than they are already.
Then there is the temporal connection. How can the latest iteration of sprawl (the previous one began with the advent of electric trolleys), which began the day after World War II ended, be a principal cause of the obesity explosion, which has largely emerged in only the last 10 or 15 years? Over the past decade America has not sprawled more, indeed, there have been population increases in some of the densest urban areas. Americans are not driving much more than they did a decade ago, nor is there any evidence that they are walking less. Not many more people are avoiding transit today than 10 years ago. Yet, over the past decade or so, obesity, by CDC data, has nearly doubled. And as the researchers' findings reveal, sprawl probably had very little to do with it.
See the study here.
Wendell Cox, Principal of the Wendell Cox Consultancy in St. Louis, Missouri, is a Visiting Fellow at The Heritage Foundation, and Ronald D. Utt, Ph.D., is Herbert and Joyce Morgan Senior Research Fellow in the Thomas A. Roe Institute for Economic Policy Studies at The Heritage Foundation.