“Walkability” Critical to Health — or Maybe Not

“The built environment really does matter to health,” says Lawrence Frank, the author of several reports that find that people who live in walkable neighborhoods are less obese than people who live in neighborhoods that lack sidewalks and other walkable amenities. Frank was “the first one to make a connection between land use and obesity,” says an admirer.

Walkable or not, the photographer who lives in this neighborhood is “in love with living in Atlanta.”
Flickr photo by rhagans.

So reports of his latest study are particularly revealing. Looking at Atlanta neighborhoods, he found that people who prefer to exercise have similar obesity rates whether they live in walkable neighborhoods or not. Meanwhile, people who prefer to drive have somewhat higher obesity rates, but they too are similar whether they live in walkable neighborhoods or not.

To the Antiplanner, this means that Frank’s previous studies were not measuring the effects of the built environment on health, but instead measured the choices people made based on their desire to exercise. People who want to exercise more will tend to live in walkable neighborhoods, while people who don’t won’t care whether they live in walkable neighborhoods. Thus, if walkable neighborhoods have lower rates of obesity, it is because people who exercise more tended to prefer those neighborhoods.

The CNN report of this study quotes a real-estate professor who says that homes in walkable neighborhoods can cost up to three times as much as homes in neighborhoods without sidewalks. Great! So let’s pass an ordinance requiring all neighborhoods to be walkable. The boost in housing prices will keep out many of the riff-raff and fatties who don’t want to walk or ride their bikes everywhere. But, contrary to Frank’s prior claims, it probably won’t improve people’s health.

The truth is that America has been suburbanizing since the nineteenth century, and the most rapid rate of suburbanization took place between 1945 and 1970. Yet the obesity “epidemic” apparently began in the 1990s, suggesting that something other than the suburbs was responsible. Many think it was the availability of low-cost high-fructose corn syrup and other processed sweeteners, which were first widely used in the early 1980s. This supposedly led to a transformation of eating habits, at least among some people, which made them fat. While no one cause is likely to be the sole explanation, this makes more sense than the suburbs.

Frank’s new study does find that the amount that people drive is influenced by neighborhood design, even if their health is not. Both exercisers and sedentary types drive about 26 miles a day if they live in walkable neighborhoods, he found, and about 40 miles a day if they live in car-friendly neighborhoods.

But this may be the result of some other choice people are making. Maybe people in the car-friendly neighborhoods like or need to drive more for some reason, and they selected those areas because they wanted to drive in less congested areas. Unless we control for all such preferences, we won’t know whether it is the neighborhood design or some other factor that is leading people in the car-friendly neighborhoods to drive more.

Unfortunately, I can’t find Frank’s actual study on line — just the news report. If anyone knows where to get it, I hope they will leave a comment.

Meanwhile, in the correlation-equals-causation file, more news reports cite another study claiming that raising gas prices by a dollar will reduce obesity by 15 percent. This study is available on line. It’s by an economics graduate student whose methodology was simple: he compared the rising obesity rates in the 1990s with gas prices which (after adjusting for inflation) were declining.

The same student also has studies “proving” that obesity was caused by Americans’ increasing work hours. However, he thinks that increased cigarette prices limited the rise in obesity.

The problems with economics is that you can’t ethically do experiments. So if you want to study a problem you have to collect a lot of data and control for a lot of variables. If someone compared obesity rates in two different places, one which had high gas prices and one low, then you might find legitimate evidence of a relationship between the two.

For example, gas prices are much higher in Europe than in the U.S. So have Europeans avoided the American obesity epidemic? Hardly. Obesity rates have tripled in Europe in the last two decades. Reported rates in Norway and other European countries (scroll to chart at bottom) are almost identical to those in the U.S.

And European cities are so walkable too. So much for the theory that higher gas prices will reduce obesity.

Update: MSetty kindly sent me a copy of Frank’s paper (which he wrote with three co-authors). Significantly, the file name is “Frank Self Selection Paper” and the title of the paper is “Stepping towards causation: Do built environments or neighborhood and travel preferences explain physical activity, driving and obesity?”

The paper finds that, for people who prefer not to walk a lot, the design of the neighborhood makes little difference to their physical activity or health. But people who want to walk do exercise a little more in walkable neighborhoods. The authors suggest that we need more walkable neighborhoods so people who want to walk can choose to live in those neighborhoods.

The paper concludes,

“Lastly, there are likely many factors that inluence neighborhood selection and preference that were not measured in the present study, including availability, cost, andother neighborhood characteristics. Although many factors enter into the choice of residential location, cost, job location, quality of public schools, transportation choices, and opportunities for physical activity are likely to be important. Understanding this myriad of factors and their relationships will better allow for the determinationof preferred built environments and independent effects of built environment on health behaviors.”

The key word here is “myriad,” which the dictionary in my word processor defines as “so many they cannot be counted.” In other words, there are just too many for government planners to account for, so they are bound to simplify, which means they will oversimplify. Instead of using subsidies or regulations to try to force the construction of more walkable neighborhoods, planners should keep their hands off and let people live the way they want (and can afford) to live.

Share

10 thoughts on ““Walkability” Critical to Health — or Maybe Not

  1. Dan

    The same student also has studies “proving” that obesity was caused by Americans’ increasing work hours. However, he thinks that increased cigarette prices limited the rise in obesity.

    Matt Kahn had some intelligent objections to the methodology; compare them with the above muddled objections for context and insight into what kind of people use ‘prove’.

    DS

  2. Dan

    So have Europeans avoided the American obesity epidemic? Hardly. Obesity rates have tripled in Europe in the last two decades.

    So what. What are the comparative, actual numbers?

    Randal conveniently omits these numbers. I wonder why.

    Nonetheless, the built environment is a disamenity, something that exacerbates a condition.

    The built environment as a disamenity means that if you want to walk off your calorie-dense meals, its harder to do. If you want to walk around to walk off your cheap corn meal , but spend 2 hours a day in your car (partially voluntary, per Plantinga and Bernell [whom I’ve also linked here multiple times]), its harder to do because you don’t have the time. If you want to walk off your supersized meal, lack of pedestrian safety may scare you off from doing so. The September 2003 AJPH goes into this concept of built environment as health issue in-depth**, and Pollan’s latest book talks about the cheap food as health issue in depth.

    DS

    ** Much of the front matter is not behind a paywall, and many of the articles are free too (it’s an important issue, despite what Randal wants to sell you).

  3. Dan

    The problems with economics is that you can’t ethically do experiments. So if you want to study a problem you have to collect a lot of data and control for a lot of variables.

    And you allow yourself to be billed as an “economist” when convenient. This argumentation is comedy, not analysis.

    That is what dependent and independent variables are for, and factoring them, then using Ts and rs to judge their robustness. Then using this to express the chance of something being true.

    Sheesh.

    DS

  4. StevePlunk

    Correlation does not mean causation. Why do people forget this? It appears they forget when convenient and it helps their political cause.

    I’m pretty sure our affluence has more to do with obesity than anything else. No, I have no study, no statistics, and no expertise. But for gosh sakes sometimes the obvious doesn’t need those things.

  5. Dan

    It appears they forget when convenient and it helps their political cause.

    Despite Randal’s mischaracterization of the nature of the study (it was a dissertation, not a thesis – [imagine that, Randal not understanding the issue]), you have no evidence that your assertion is true in this case. More likely his committee didn’t give good enough guidance.

    I’m pretty sure our affluence has more to do with obesity than anything else.

    Pollan’s latest book linked above explains the food portion of our obesity problem (hint: more complicated than affluence). The AJPH link above explains the built environment portion of our obesity problem (hint: mo’ complicated than affluence). Here and here** are overviews explaining the land use portion of the American obesity problem (hint: mo’ complicated than affluence).

    In short: affluence is indeed a component in all factors but an indirect causative factor, despite the efforts of people who want to limit our freedom of transportation choices to only one choice**** (one with a high correlation to obesity).

    Anyone with empirical evidence showing otherwise feel free to share it.

    DS

    ** http://tinyurl.com/346oz7

    **** http://www.obesityresearch.org/cgi/content/abstract/10/4/277

  6. msetty

    The reason Randal couldn’t find Frank’s study is because it hasn’t been posted yet!

    Dan, email me if you want a copy of the latest, which is NOT any of the papers you list.

  7. Tad Winiecki

    I am concerned with the children and their physical, mental and spiritual health. My ideal would be close to what I had in my childhood – I walked or rode my bike to school. After school I participated in school sports, band, clubs, Little League, YMCA or played with my friends. I could walk or ride my bike to a grocery store if I needed to pick up something for my mother. I never rode a bus or had my parents drive me to school.
    Bad cases now in the central city are children who walk home after school and try to avoid thugs. Then they sit and look at a TV or computer screen until their parent returns. In the suburbs many students ride the school bus and miss out on afterschool activities. Others ride with parents and participate in activities if their parents have time to take them. If both parents work outside the home children may be left alone with their babysitters (the TV, computer and video game). Some of these media are bad influences – encouraging violence, greed, lust, bad eating habits and other sins and discouraging exercise.
    If it is safe, easy and not too far to walk or bike to school and park or other play area children will get more exercise and will become more self-reliant and build self-esteem.
    Through streets and sidewalks good. Cul-de-sacs, deadends, media encouraging sin bad.

  8. Dan

    Mike et al.:

    Apologies. I see that Larry has a new one, I was pre-coffee this morning and thought we were discussing the last one in the SMARTRAQ series, the summary report, which I’ve linked to here a number of times. Here is a good overview and discussion of those findings and their limitations.

    DS

Leave a Reply