|►◄ Reverse Zone|
Complete List of Posts
Mon, 22 Sep 2008
In his famous book Bowling Alone, Robert Putnam launched the new field of Social Capital by showing that Americans have less social interaction with each other in recent decades than at any other point in their history. The assertion is well documented with copious quantitative data. One of the culprits, according to Putnam, is urban sprawl. Social interaction measures are significantly lower in the suburbs. Various other researchers have confirmed this.
But a couple of months ago, an opinion piece in ArchNewsNow Bowling Alone in Urbanistaville says that no one had ever actually checked this and that in fact the opposite was true: social interaction goes up when density goes down. I wondered what quaint little California or Texas town was called "Urbanistaville".
The arguments in the research paper itself are startling not just by the assumptions that have to be made to reach the final conclusions, but some of the off-hand offensive stereotypes that are proposed to explain the results.
The paper starts out saying (if you read carefully enough) that in fact one does measure higher social interaction at higher density. In fact it says it is necessary to assume that higher density increases social interaction in order to prove the opposite. Bear with me. It also assumes that everyone wants to have as much land as possible such as you get in the outer suburbs because, well, who wouldn't? The only thing that would stop you from having an even bigger lot is that the size of your lot would reduce the entire neighbourhood's density and you would then miss out on all the benefits of higher density. "In choosing space consumption, a household would consider the direct gains from having more room, along with the negative effect on the social interaction it enjoys, caused by the drop in neighborhood population density due to its larger residence."
This is where self-selection comes in. Unusually gregarious people would then choose to live in high-density census tracts. And who are these gregarious people who choose not to live in the suburbs and give up the advantages of having bigger lots? Blacks. Hispanics. Asians. Foreigners. Unemployed people.
Given that being gregarious is the only reason these people live there, we can then calculate the self-selection bias from this. A bit of a circular argument. Having corrected for this bias, we now see the coefficients reverse their sign. It turns out that interaction is greater in lower density after you remove the bias caused by people's choice of census tracts.
What a tragedy for all these poor misguided people who selected where to live on this basis. It turns out that all these gregarious people unwittingly chose a location with lower social interaction, not higher! As soon as they realize their mistake surely they'll move to the suburbs too.
The paper then goes on to ask why is it that these blacks and hispanics have so little social capital, considering. The first theory: there are so many museums and theatres that there is no need to interact with real people. That must be it; they are too busy going to the opera every night. Another one is that high density is associated with criminal activity, which makes people suspicious of each other. The authors actually expand on this reason at length, trying to figure out how they could test not whether crime is a cause of the lower social interaction, but presuming that crime is a necessary consequence of density and testing by what mechanism it makes people interact less. The paper then seems to gloat that they are "undermining an important line of attack used by critics of urban sprawl."
The ArchNewsNow article goes even further, saying that it's in our genes that density causes fighting, domestic violence, child abuse and neglect, soaring infant and maternal mortality, psychosis, gay bashing, and so forth, as a mechanism to reduce this density. Darn. Do our genes also cause the writing of pro-sprawl articles and the lobbying of municipal politicians to prevent zoning for semidetached and townhouses?
In all seriousness, it would be interesting to see real studies that do not presume that the social effects of density are linear with the number of units or people per acre. There is a world of difference between interactions among residents of high-rise apartment apartment buildings and those of compact older neighbourhoods with row houses or small lots. It's not just the continuation of the line from large lots to small lots. It is entirely possible that both extremes of density have low social capital scores, and that there is a happy medium somewhere in between that Goldilocks would find just right.