fredag 8 november 2013

Some criticism on an article about Social disorganization theory and foreclosures

Last week google scholar alerted me to a new article that could be of interest to me, an online first article in Housing studies published 21/10 2013 by Jaclyn Schildkraut* & Elizabeth Erhardt Mustaine 
Movin', But Not Up To The East Side: Foreclosures and Social Disorganization in Orange County, Florida
The authours argue that social disorganization theory can and should be expanded to a broader scope of issues than the normal crime, delinquency, disorder and health related issues it is. In their study they study the relation of social disorganization theory and foreclosures, short sales or normal sales of homes in Orange County, Florida. So far so good, it is an original and refreshing idea to try something new. But in my view the article has some very serious flaws, in fact so much that i was surprised it passed peer review. Below i will outline some of my criticism.

Social disorganization theory is a classic within criminology/urban sociology stemming from the Chicago school of sociology with early works by Park & Burgess (1925) and notably Shaw & McKay (1942). The basic premise is that there is more crime, delinquency and disorder in some parts of a city, and that this largely can be explained by social disorganization in those neighborhoods. Social disorganization is conceptually pretty much what it sounds like, a lack of social organization, but in classic social disorganization theory usually studied by looking at correlates in the form of poverty (or concentrated disadvantage which is a broader concept), (ethnic) heterogeneity and residential instability. In a neighborhood with lots of poverty, a heterogeneous population and a high population turnover the theory states that we can expect a high degree of social disorganization and thus a high crime rate. More recently surveys have often been employed to directly measure the social disorganization (ie, measure social networks, organizations, trust and shared expectations), but it is still fairly common to use the correlates of social disorganization rather than direct measurements.

In the article the authours bring up the normal references in their discussion of social disorganization, but as soon as we reach the "Measures" (page 9) section some question marks start to pop up. I have numerous minor issues with the article, but i will here focus on two major and inter-related problems. As mentioned above the standard three variables to use as indicators of social disorganization are poverty, heterogeneity and residential instability. The authours of this study replace heterogeneity with "social capital". They have no discussion on social capital in the theory section of the paper. They measure social capital by using four variables; share of males without bachelor degree, share of females without bachelor degree, share hispanic, share minority.
This is problematic in a number of ways.
1) They include two variables (hispanic and minorities) that fit the "heterogeneity" normally measured in relation to social disorganization. But in including education and re-labling it "social capital" they distort the measurement into something not fitting the theory. 
2) No theoretical discussion on social capital in relation to social disorganization, or even to social capital in general. They do however heavily cite Robert Sampson (the foremost contemporary researcher on social disorganization) who often bring up social capital as a broader concept in his theoretical discussions. In essence a low social capital is the same as social disorganization. If we are to consider this measure as social capital then it could be argued that they try to establish the impact of social capital (social disorganization) on foreclosures after controlling for poverty and residential stability. But the authors make no such argument.
3) Arguably education is human capital, rather than social capital, at least on the individual level. The authors argue that educational attainment measure "residents having the social capital to improve their circumstances"(p10), a statement that sounds plausible, but they cite no studies showing such a connection.
4) The authors cite a number of studies to support their construction of social capital this way, correctly describing those studies as bringing up "immigrant concentration" and "socioeconomic status" - but those studies do not discuss education! To be fair at least one of the studies have included one education variable in their measurement of socioeconomic status (Sampson & Groves 1989) - but other studies cited do not use education at all (ie, Sampson et al 1997)! And, even the study that does include one education related variable includes it on the "poverty" variable, while they put it in a "social capital" variable (in part resembling the heterogeneity variable) instead of their poverty variable.

To sum up then, the authors construct a new measurement of social capital with no theoretical discussion, in which they include the often used heterogeneity concept combined with education while referencing studies that does no such thing. Normally the three main correlates or sources of social disorganization - poverty, heterogeneity and instability - are used to indicate social disorganization, and the authors appear to do the same thing with social capital instead of heterogeneity rather than treating social capital as actual social disorganization.

In their results then they find that in relation to foreclosures (ie, forcible sale of someones home) poverty and instability is insignificant while social capital is significant. For short sales (ie, selling home to a low price to pay loans) none of the three variables are significant, and for normal sales again only social capital is significant.

The authors then conclude (page 14): 
"Conversely, communities with greater proportions of real estate sales that are foreclosures also have quicker sales, less affluent residents moving into the community, and greater amounts of social disorganization. This is very supportive of the thesis that social disorganization and troubled real estate markets are related at the community level" (my emphasis)

I have a very hard time making the same analysis. Two of the measurements of social disorganization (poverty and instability) are insignificant. "Social capital" is significant, but measured in a way that is not at all consistent with social disorganization literature (at least not the literature they cite). If the authors are considering their measurement of social capital as de facto social disorganization rather than as a correlate of social disorganization, similarly to how for instance Sampson et al (1997) discuss collective efficacy then their conclusion would be correct. But. The authors make no such claims in their study that i can see. And, Sampson et al measure collective efficacy (social capital) directly with 10 survey items before controlling for poverty, heterogeneity and instability, and can thus show that social capital mediates much of the effect the social disorganization variables have on crime. Making such claims on the data used in this article would seem very strange.

I would conclude the opposite of what the authors do. Social disorganization and troubled real estate markets are not related (in their analysis). Now, i havent mentioned that they also control for several other variables, among which is "proportion affluent households".This is an interesting finding, and the authors could have argued that social organization (ie, affluence) is more important for foreclosures than social disorganization (ie, poverty). However, lets take note of what it actually means - places with more rich people gets less foreclosures.

Finally then, although i disagree with much of the method and conclusions in the article the authors do bring up several interesting points on the relation of social disorganization, foreclosures and short sales in their discussion that are worth keeping in mind (p17).
"Many foreclosed houses fall into disrepair, yards are unkempt, and units may be left vacant for
extended periods of time. Flipping foreclosed homes quickly can result in high residential
turnover, which can impede interactions and attachment between neighbors. Selling
homes for much less than what they may be valued impacts the economic worth of other
homes in the neighborhood, which in turn is likely associated with the social capital of
incoming neighborhood residents. All of these circumstances are factors highlighted by
social disorganization theorists as important in the relationship between neighborhood
organization and the presence of social problems."

Inga kommentarer:

Skicka en kommentar