zip code level data?
I am looking for zip code level data from 2004/2005 that has:
- demography: gender, age, race, SES, occupation. Other stuff (like family structure or housing) would be a plus.
- political behavior: party ID + 2004 presidential vote would be more than enough.
- “lifestyle” data: consumption habits, to give me a sense of what zip codes are “creative” or “bourgeois” or “working class.” Think Richard Florida.
The data can come from separate sources, I can always merge. I am looking for free or very low cost data. I don’t have much of research budget. Leave info in the comments or email me at f r o j a s “atsymbol” indiana dot edu. Thanks.
Does pres vote by zip code exist in principle? Is there any relationship between zip code boundaries and precinct boundaries? I suppose someone could have done imputations for precinct boundaries that cross zip code boundaries.
Jeremy
November 12, 2007 at 4:34 am
It might exist from surveys, or imputations. I am hoping that someone would know the answer.
fabiorojas
November 12, 2007 at 4:35 am
larry rothfield and terry clark at chicago have been collecting some very good cultural amenities data at the address level, which includes zip. just google “Cultural Amenities Project”
i think the FEC has data on political donations by zip. various people have put new front ends on this data. in some ways this is more interesting than registration as it measures the especially committed fraction.
finally, a few years ago i bought “zip code download” and have been very happy with it. it’s basically a csv file that lets you create merges between different geographic units. (i’ve mostly been interested in aggregating zips into MSAs but i think you can do a lot narrower than that.)
gabrielrossman
November 12, 2007 at 2:15 pm
The classic collectors of such data are the market research firms. Caci used to collect “ACORN” Claritas was another big data collector.
Arnold Kling
November 12, 2007 at 6:35 pm
Arnold – what’s the cost for such data?
Fabio Rojas
November 12, 2007 at 6:35 pm
Fabio,
They price discriminate. By now, some of their stuff may be on line. The best thing to do is figure out who the researchers are for the companies and get friendly with them.
Arnold Kling
November 12, 2007 at 10:42 pm
YOu might contact Don Beck at spiraldynamics.org
He has some data you might find interesting, and that might work for what you’re doing. He has what are called psychosocial complexity maps.
zatavu
November 13, 2007 at 3:55 am
Have you checked out Sperling’s Best Places? http://www.bestplaces.net/
I’ve been doing various things involving zip code level data in MA and found Sperling has many variables at zip code level for all of the US.
anonymous
November 13, 2007 at 5:29 am
The Census Bureau can give you everything you want in your first point, except that they don’t break the occupational data down particularly finely. (Their occupation categories are things like “Wholesale Trade,” “Retail Trade,” “Construction,” etc.) You can download Census Bureau data in CSV format here: http://factfinder.census.gov/servlet/DownloadDatasetServlet?_lang=en
I suspect that you can get everything else you want through the Inter-University Consortium for Political and Social Research (ICPSR), of which Indiana University is a member. I’m not at work right now to check and I don’t have the ICPSR datasets memorized, but they have pretty much all of the data you could ever want when doing number-crunching in political science. Apparently you have to go through these folks to get acess to the ICPSR data at IU: http://www.polsci.indiana.edu/datalab_dataservice.asp They might also be able to get you exactly the Census data you want without making you wade through the Census download process yourself, and it should all be free to you.
This information brought to you by your friendly neighborhood academic librarian. (Yes, librarians actually know where to find stuff! Who’d've thunk?)
AnonymousLibraryStudent
November 13, 2007 at 12:42 pm
AnonymousLibraryStudent is right that Census provides a lot of basic demographic data at fine levels of geographic disaggregation. (If you’re a GIS user, it’s also available as ArcView GIS shapefiles, which are also usable in the Grass GIS software.)
A minor caveat is that the Census data are aggregated from Census blocks to ZCTAs (Zip Code Tabulation Areas). Real ZIP Code data can be expensive — we don’t get it for free even through work for the Postal Service — though as Arnold says, you may be able to take advantage of price discrimination as an academic. I’ve done some casual investigation of the correspondence between ZIP Codes and ZCTAs, to convince myself that the correspondence is mostly good enough for most applied social science research. The issues tend to be in very fast-growing areas, where ZIP Code boundaries are relatively fluid.
atbozzo
November 17, 2007 at 2:24 pm
Fabio, I’m a little late to this game but given what people are commenting about using Census geographies to construct zip-code-level data, this reference from the Census explains how geographic areas are used in Census tabulations. I have found it to be invaluable in much of the research that I am doing.
Also, I am not sure on the exact release dates, but the American Community Survey (which is replacing the Census long form) will release small-area data as the sample becomes large enough in small areas to censor the data in order to protect confidentiality. Of course, this means that it might not be available for 2004, but should be available before 2010 – and interpolation techniques can produce fairly reasonable estimates.
Mike3550
November 20, 2007 at 10:34 pm