Tyler Cowen links to an article by Jan Brueckner and Ann Largey that finds that people in suburbs have more social interaction than people in cities. Here’s the abstract:
Various authors, most notably Putnam (2000), have argued that low-density living reduces social capital and thus social interaction, and this argument has been used to buttress criticisms of urban sprawl. If low densities in fact reduce social interaction, then an externality arises, validating Putnam’s critique. In choosing their own lot sizes, consumers would fail to consider the loss of interaction benefits for their neighbors when lot size is increased. Lot sizes would then be inefficiently large, and cities excessively spread out. The paper tests the premise of this argument (the existence of a positive link between interaction and density) using data from the Social Capital Benchmark Survey. In the empirical work, social interaction measures for individual survey respondents are regressed on census-tract density and a host of household characteristics, using an instrumental-variable approach to control for the potential endogeneity of density.
I’ll have to read through this more carefully. To start with, I’d like to see some simple comparisons and scatterplots (average # friends of people who live in rural area, towns, suburbs, and cities). I’m sure there’s a good reason for the regressions in the article, but, as always, I’d understand and trust results such as their Table 3 much better if I could first see some data patterns. It looks like the data are publicly available, so it shouldn’t be a problem to do these analyses.
We’re actually planning to do similar analyses with the data we’ve collected using the General Social Survey. One of our conjectures is that people living in cities have more acquantainces but fewer people they trust, compared to people who live in smaller communites. I’d always thought this would be consistent with Putnam’s work on cohesive communities.
That's a frustrating paper: no plots, no tabulations of simple things that are immediately relevant…just regression tables with coefficient estimates for dozens of variables (and only for regressions that include all of the variables at the same time). I'm sure they looked at a lot more than that; I wish they had shown just a little bit of what they looked at.
My gut feeling is that I'm suspicious of these kitchen-sink regression models (in all contexts). There are just so many possibilities for unexpected relationships between variables to screw things up. This doesn't mean the results are "wrong" or that there is a problem in this particular case, just that I'm cautious about taking regression coefficients at face value.
In this case, there's an additional reason to be suspicious: I grew up in the 'burbs, and since then I've lived in cities, and I think people in cities have a more social interactions, hang out with friends more often, etc. Of course I could be wrong, for lots of reasons. For instance, I grew up in the suburbs twenty years ago, and I'm living in a city now; perhaps the suburbs have changed (on average). Or perhaps I and my friends are very unusual city-dwellers, and/or my parents and my parents' friends (when I was growing up) were unusual suburbanites. But I also think it's possible that the "density" regression coefficients can't be interpreted the way the authors of the study are interpreting them, perhaps because of interactions between variables or for some other reason.
So, interesting study, but it leaves me with a lot of questions.
The OLS regressions show that people living in more dense census tracts actually hang out with friends more often than those living in less dense census tracts. But, as the authors point out, people may self-select into census tracts in a way that is confounded with sociability (e.g., more social people want to live in denser parts of town — city rather than suburbs).
So they use metropolitan area as an instrumental variable to estimate the effect of population density on social interaction. Critically, this assumes that people don't self-select into metropolitan areas in a way that is confounded with sociability. But the authors don't provide evidence to justify this assumption. It seems very likely to me that the sort of people who choose to live very dense metro areas like New York City differ in many ways from those who choose to live in less dense metro areas like Albany, NY. And these places differ from each other in lots of ways besides population density.
The results essentially say that people who live in dense metro areas have less social interaction than people who live in relatively sparse metro areas. But more dense metro areas also tend to be larger ones. The characteristics that differentiate large cities from small cities (the metro-area level of analysis) are not the same as the characteristics that differentiate the center of a city from its suburbs (the census-tract level of analysis). This paper elides that distinction.
One could argue that it's not population density that is the key variable here, but suburb vs. city lifestyle. The metro area is a combination of a city and it's suburbs. If the most dense metro areas (the biggest cities) have the highest percentage of people living in suburban "sprawl" then these data would support the opposite conclusion: that sprawl is associated with less social interaction.
I read this post with great interest since a co-author and I have been using the same data to look at a related question: how does sprawl impact various forms of political participation? Our first move is to differentiate "sprawl" into five separate components, and then to evaluate the impact of each separately. What is interesting from the point of view of this Brueckner and Largey article is that low-density environments do encourage certain political activities–but that other aspects of sprawl have negative effects. The paper is available here