Home Products My Account Company Census Data Discussion Reviews Support Licensing
P r o d u c t s
Census 2010 /ACS
Market Segmentation
Bundled Packages-Big Savings
Crime Reports
Normalized Data
CensusCD 1970,1980,1990
Business Data
Geocoding, Zip+4, Telco
On-Line Data & Maps
Custom Reports
Special Data Sets
GeoLytics News: Comparing Census Data from Different Years and Different Geographies

Tuesday, April 3, 2007

Comparing Census Data from Different Years and Different Geographies

Comparing data sets collected from different census years has many obstacles for time series analysis. One primary obstacle is the changing geographic boundaries between data sets. The issue of changing geographic boundaries is one every researcher has to tackle if she wants to compare census data from one year or decade to another. For example, census boundaries such as tracts and block groups typically change between censuses, as well postal boundaries such as zip codes. So how does a researcher deal with these changing boundaries if she wants to examine and compare data at smaller units of analysis? The answer is cross-walking data from one set of boundaries to another, so that the data being compared is in the same set of boundaries. This process is called normalizing the data.

GeoLytics offers several normalized data sets that transfers data from one time period cross-walked to the 2000 census boundaries. The Neighborhood Change Database has 1970, 1980, and 1990 data normalized to 2000 tract boundaries. The 1980 in 2000 Boundaries CD and the 1990 in 2000 Boundaries CD have the data cross-walked to all the 2000 Long Form geographies, not just tracts. Thus you can look at the 1980 or 1990 data in the 2000 zip codes, block groups, MCDs and more.

For a list of our individual normalized products and our two discounted product packages, go to:

Another question researchers may tackle is how two geographies relate to one another. For Example, what County is a given Zip Code in? The United States Census Bureau (USCB) defines many geographies as subsets of a larger geography (i.e. counties in the state). But there are also many geographies that the USCB does not define as subsets of each other. The smaller area must be wholly contained within the larger for all parts of the country to allow this subarea definition. Zip codes (or school districts, etc) are areas that are not nested subareas and do not allow smaller geographies (tracts and block groups) to be subareas of them. Thus if someone wants to know what county a zip code is in or what tracts are in a given zip code this is not something that the USCB data will answer directly.

Because of this shortcoming, GeoLytics has created the Area Correspondence File to answer such questions. Different researchers have different research needs so we have built flexibility into this program. There are two types of selections so as to best meet your needs: dominant area vs. all available area (which only lists one dominant area), and weighting by population vs. land area (which gives the percent of coverage for each subarea). We can create a correspondence between any two geographic areas from the 2000 Census. So for example, we can tell you the zip codes in an MSA or all the tracts in a census defined place, etc. For more informaiton on area to area correspondence, go to:


GeoLytics, Inc.
PO Box 10 East Brunswick, NJ 08816 · 1-800-577-6717 · Voice: 1-732-651-2000 · Fax: 1-732-651-2721
E-mail · Web http://www.geolytics.com · Demographic Resources