If you like your data clean, pretty and predictable, you probably don't want to read the rest of this post.
Researchers from my hometown here in Ottawa recently published a study ahead of print in the journal Obesity. The paper, Relationships Between Neighborhoods, Physical Activity, and Obesity: A Multilevel Analysis of a Large Canadian City looked at a number of different built environment variables and their impacts upon the probability of both leisure time physical activity and overweight and obesity in the population.
The statistical models controlled for: Age, education, household income, smoking status, and, given we live in a city where there's a lot of snow, season of data collection.
The variables that were considered as potential neighborhood based influencers of physical activity and obesity were: Total bike and walking path length, number of free or minimally expensive indoor and outdoor recreational facilities, park area and green space, social cohesion, neighbourhood safety, number of grocery stores, fast food chains, convenience stores, specialty food stores and full service restaurants.
In men, almost nothing mattered. In fact with weight, nothing at all mattered. Number of fast food joints, convenience stores, grocery stores, park space or recreational facilities - none of these things seemed to impact upon a man's weight. There was however one variable that correlated with increased physical activity. Crime rate. Yup, the more dangerous the neighbourhood, the more active the man, to the tune of a 14% increase in their odds of being physically active for every standard deviation increase in crime rate. Running away from the bad guys?
In women? With women for every standard deviation increase in park area they were found to be 17% more likely to be physically active. So far so good, but check this, for every same standard deviation increase in park area women were also 15% more likely to be overweight or obese. In fact the increased risk to overweight and obesity seen with increasing park size in women was only ever so slightly eclipsed by the 17% increased risk conferred to their weights by a standard deviation increase in numbers of convenience stores. The riskiest thing for women's weights according to this paper? Fast food restaurants with a 38% increased risk of weight for every standard deviation increase in those.
So what do these numbers mean? So many questions. Should inactive men be encouraged to live in dangerous neighbourhoods? Does nothing in a man's environment really matter to their weights? Should women rethink that new home they're considering beside the park?
I don't think so.
I think all these numbers mean is that the reasons why people gain weight and exercise, they're stupidly complicated and trying to suss out the impact of a dozen or two different criteria on this sort of ridiculous complexity, might in fact be turn out to be a bit of an exercise in statistically meaningful futility.
Prince, S., Kristjansson, E., Russell, K., Billette, J., Sawada, M., Ali, A., Tremblay, M., & Prud'homme, D. (2012). Relationships Between Neighborhoods, Physical Activity, and Obesity: A Multilevel Analysis of a Large Canadian City Obesity DOI: 10.1038/oby.2011.392