Freeman 2001 - "The Effects of Sprawl on Neighborhood Social Ties: An Explanatory Analysis"

Freeman, Lance
“The Effects of Sprawl on Neighborhood Social Ties: An Explanatory Analysis.”
Journal of the American Planning Association.
Winter 2001; v.67, n.1; pp69-77
Relevance: medium

To test whether low-density sprawl weakens neighborhood social bonds, Freeman compared survey data on neighborhood social ties with the density and demographic characteristics of the census blockgroups in which the respondents lived.

After controlling for poverty and other factors, he concluded that residential density is not significantly related to the formation of neighborhood social ties; however, such ties are affected by how much neighborhood residents rely on their cars.

Freeman cites studies that suggest that, whereas low densities may undermine social ties, at some point on the density scale higher densities start to have the same effect.  Low density yields too few opportunities for interaction; high density yields too many, resulting in withdrawal from people.

Methods: Freeman measured social ties by the number of confidants (0-3) that participants said they lived in their neighborhood (“Who are the people, other than people living in your household, with whom you discuss matters important to you?  Does that person live in your neighborhood?”).   (More here.) Density was measured by census blockgroup and dependence on the automobile was proxied by the number of individuals who drove to work alone.  Statistical controls were included.

Results: Every 1% increase in the proportion of individuals driving to work is associated with a 73% decrease in the odds of an individual having at least one neighborhood social tie.

Commentary: While this article indicates that commuting alone may significantly influence neighborhood social ties, the numerical relationship likely depends on the limited survey method. In this case, by limiting socail ties to the first three people participants listed, the study may include only best friends and not friendly neighbors.


Clark Williams-Derry

Notes from an email exchange with Dr. Freeman --

On Wed, 6 Jul 2005, Clark Williams-Derry wrote:

When reviewing your 2001 APA article, "The Effects of Sprawl in Neighborhood Social Ties: An
Explanatory Analysis" I was struck by this statement on p. 74:

"Every 1% increase in the proportion of individuals driving to work is associated with a 73% decrease in the odds of an individual having a neighborhood social tie."

The direction of the correlation makes sense to me. But the magnitude seems inordinately steep -- so steep that I was wondering if it's actually a typo.

(And just to make sure I'm interpreting the statement correctly, here's what it seems to mean: Let's say that in a given neighborhood, 50% of people drive to work, and the odds of having a neighborhood social tie are 50%. In another neighborhood with identical demographics, but in which 51% of residents drive to work, then the odds of having a neighborhood social tie decline to (1-.73)*50%=13.5%. I could believe a .73% decrease, or an absolute decrease of .73, but a 73% decrease seems way too large.)

If you have time, could you clarify this? Thanks so much!

Response from Dr. Freeman:

The relationship is steep according to the analyses. But I would not get to hung up on the exact numbers as per your example below. The dependent variable has a restricted range, 0,1,2,3 ties. If you note Table
1 you'll see most of the respondents did not hve any ties. In addition as per Table 1 the independent variable measuring driving alone has a fairly narrow range as indicated by the standard deviation.

So even with small changes in the independent variable we are witnessing relatively large changes in the dependent variable as the range and maximum of the dependent variable is restricted.

Eric de Place

One question to explore: at what densities do we begin to see a decrease in social ties? Are these Manhattan-style densities? Or densities that we have in the NW?

Eric de Place

I worry that the "background" and "evidence" sections of Freeman's article rely too much on old--and quite possibly outdated--studies. This is especially true in his characterization of higher densities' negative effect on social capital. To question the link from sprawl to social capital he cites two studies from 1969, one from 1978, and one from 1938 (admittedly, he also cites two more from 1998 and 1999). It seems entirely possible--even probable--that studies from those times reflected different cultural paradigms than we have today, as well as quite different urban forms. The upshot, is that I think the article may be excessively dismissive of density as a link to soc cap because he is relying on outdated evidence.

The comments to this entry are closed.