Help with 1960 census tract fips - best unique identifiers?


I’m trying to make sense of 1960 Census tracts.

I realize, by way of an earlier Q&A posting, that due to two different data sources and a lot of missing data, we should anticipate redunancy in the 1960 tracts. For confirmation, I created a FIPS variable using ‘state + counnty + aggprefix + aggtract + aggsuffix’, but there are many duplicates. Is there any documentation as to how to negotiate this when creating unique 1960 tract FIPS identifiers?

Is there a different ID that’s been created and applied to 1960 census data in the decades since that has become the norm/convention for unique 1960 tract FIPS identifiers (like, GISJOIN, or the like)? If so, is there any documentation explaining this?

Any guidance/references will be greatly appreciated!


With respect to your user forum question, it looks as if you’re using the 1960 conflated census tract shapefile. You created a concatenation of state, county, prefix, tract, suffix, and you did find duplicate values. There is also a GISJOIN value on that shapefile, which doesn’t have duplicate values (except for counties with NODATA tracts).

We create our conflated 1960 tract data by aligning our TIGER/Line 2000-based boundaries to TIGER/Line 2008-based boundaries. After creating the polygons, we used centroids from an older data file to assign the GISJOIN and the aggprefix, aggtract, aggsuffix values to the 2008-based boundaries. The older data file has the correct GISJOIN values, but not necessarily the correct aggprefix, aggtract, or aggsuffix values associated with it. We didn’t keep the aggprefix/aggtract/aggsuffix values up to date over time; instead, we updated the GISJOIN value.

You’ll want to use the GISJOIN value on the shapefile to merge with the population data - I don’t recommend trying to create your own FIPS value by concatenating fields together.