lmbett
commented
6 years ago Great questions. The definition has been discussed in the sociology literature, but there is no accepted gold standard. As you suggest, I though about using measures of spatial variation to define neighborhoods, but you may get a definition for income, another for education, another for race and ethnicity etc... it would be interesting to explore though.
RE: Information, Spatial Selection and the Statistics of Neighborhoods
First of all, thanks for presenting!
You discussed your use of Census Block Groups (CBGs) to operationalize “neighborhood”. Could you talk a little about the methodology the Census uses to draw and define these groups? Do you think the government methodology for selecting the CBGs had any significant effects on the information you found?
Additionally, given a smaller “neighborhood,” do you think your results would be show more or different effects? Perhaps there are smaller block effects that are missed in this research. The middle income areas, for example, might appear mixed at the CBG scale, but have very clear differences at a small scale (at the street block or the building level). I assume the data for this would be difficult to find, but I wonder what information is lost through this minor “operation of coarse-graining.”
Also, I wonder why some of the faster growing cities have high income segregation by neighborhood. You state that this is due at least in part to recent residential decisions. Do you think the neighborhood segregation of these areas is not just symptomatic of higher growth, but partly causal/predictive?
“We see that Dallas TX, followed by New York City and New Orleans, LA show the highest I(y,n) and that many cities of Texas show in general strong income segregation by neighborhood. This is particularly interesting because these cities are currently among the fastest growing in the nation so that at least some of the observed income segregation is the result of recent residential choices.”