Robert Sampson responds to my short book review

by on April 15, 2012 at 5:10 pm in Books | Permalink

Prof. Cowen -  your Marginal Revolution blog earlier today was brought to my attention…A quick note on your selection/transparency comment, which I found of interest.  One of the ways that traditionally conceived “selection”/individual effects are neighborhood effects is when the former are an outcome of the latter.  It is common in the literature in sociology or psychology at least to see controls for the mediating pathways through which neighborhoods (or really, any context) might plausibly work.   For example, we typically see controls for all kinds of family and individual characteristics (including learning), almost all of which are at least potentially influenced by context.  Controlling them can thus have the result of eliminating the neighborhood coefficient, which is usually interpreted as evidence for selection as the governing process.  But in this example selection factors are themselves neighborhood effects, the basis in part for my reversing a common claim.   A number of recent papers independent of my own work have shown a variant of this process (e.g, http://www.ssc.wisc.edu/soc/faculty/pages/docs/elwert/Wodtke%20Harding%20Elwert%202011.pdf).  Although often technical, behind the development of these models is an important substantive point I think.   Part V of the book also delves into residential migration flows and higher-order structures as another kind of mechanism,  including how changing characteristics of neighborhoods influence residential selection.

More generally, I do not view choice/selection and context as an either/or proposition, and as an economist I am guessing you might agree. (Sociologists are typically structural determinists, but that is another story).  At Chicago I was influenced by Heckman and his arguments on modeling selection and the often misleading faith put on experiments as revealing causality).  Although I tried to examine neighborhood selection seriously, the main motivation of the book was to build up the social science of measuring and conceptualizing the neighborhood and spatial dimensions of social life.  Massey’s recent review of my book I think captures the essence of what I was trying to accomplish in terms of contextualizing human behavior and choice/selection — http://www.sciencemag.org/content/336/6077/35.summary.

Michael G Heller April 15, 2012 at 6:05 pm

Another example of a non-fiction author “fussing over his books even after they were published” (previous post).

Further examples see here:
http://www.project-syndicate.org/blog/utopia–fiction-or-reality-

Also, what’s wrong with sociologists being structural determinists?

Stephen Williams April 15, 2012 at 6:18 pm

” The main motivation of the book was to build up the social science of measuring and conceptualizing the neighborhood and spatial dimensions of social life. Massey’s recent review of my book I think captures the essence of what I was trying to accomplish in terms of contextualizing human behavior and choice/selection”

Why can’t he express himself in clear language? Language such as this closes out most of the reading public. I suppose it would be a good soporific.

The Engineer April 16, 2012 at 9:51 am

I agree. Lots of jargon. Completely unnecessary, if one were actually trying to start a conversation with normal people.

Bill April 15, 2012 at 6:55 pm

If you were to say that neighborhoods have reinforcing network effects, most economists would understand it.

Here is a book review that puts it in better english:\

http://www.press.uchicago.edu/ucp/books/book/chicago/G/bo5514383.html

Jeff April 16, 2012 at 12:03 am

While I agree with Stephen Williams comment above, what Sampson describes sounds a lot like multi-collinearity and/or simultaneity problems. You deal with the first be getting more data or using some kind of shrinkage estimator. For the second, find some instrumental variables.

Comments on this entry are closed.

Previous post:

Next post: