Are there population counts for each PUMA-country overlap for the 1960 Census as there is for the 1990 Census?

USA
1

I’m interested in building weights to go from 1960 PUMAs to commuting zones (as defined in David Dorne’s website). In order to build them, I need to have information in the population counts for each PUMA-country overlap. I understand that the Census Bureau provides this files for 1990 and 2000 but I’m not sure if the information is also available for 1960 pumas.

1 Like
2

I’ve attached a spreadsheet that contains the population count of each 1960 puma. This should provide you with the information you need.
Let us know if you have any questions.

Tract_PUMAMINI_pops_.xlsx (1.52 MB)

1 Like
3

Thank you.

Could also provide me the population counts for each PUMA-county overlap for the 2000 Census or is there a website from which I can download these files?

4

I think the information you are looking for can be found in two places. First, this page under IPUMS geographic tools has information on PUMA population counts. Second, you can use the MABLE/GeoCorr12 tool to find population overlap between counties and PUMAs.

I hope this helps.

5

Thanks!

6

I’m trying to calculate the populations by PUMAMINI and PUMA in the file you provided, Tract_PUMAMINI_pops_.xlsx (which I assume is at the TRACT level).

However, many PUMAs have a population well below 100,000, which should not happen according to the description here: IPUMS USA | 1960 PUMAs and Mini-PUMAs

What am I doing wrong?

Here is my Stata code:

import excel “Tract_PUMAMINI_pops_.xlsx”,
egen pop_puma = total(Smpl_or_100), by(puma60)
egen pop_pumamini = total(Smpl_or_100), by(PUMAMINITT)
summarize pop_puma pop_pumamini, d

7

IPUMS USA defined the PUMA and mini-PUMA boundaries for the 1960 5% sample as part of a restoration project (IPUMS USA). 1960 PUMAs with populations less than 100,000 persons are a result of the process we used to delineate mini-PUMAs (PUMAMINI) and PUMAs. The goal of our process was to:

  • Maximize the number of mini-PUMAs we could construct, based on the 50,000 person threshold
  • Make 1960 PUMAs as similar to 2000 5% PUMAs as possible

The main threshold we had to meet was the 50,000 person minimum population for either mini-PUMAs or PUMAs; this differs from the standard 100,000 person threshold imposed by the Census Bureau for all other PUMS data. This revised threshold allowed us greater flexibility to maximize similarity between our defined PUMAs for 1960 and the 2000 5% PUMAs. When our PUMAs exceeded 100,000 persons, we tried to subdivide them into two or more mini-PUMAs to provide as much geographic specificity as possible. When our PUMAs were between 50,000 and 99,999 persons, we could not subdivide them into mini-PUMAs and meet the 50,000 person threshold. PUMAs in the 50k-99,999 range reflect cases where the geographic footprint from the 1960 input units matched a 2000 5% PUMA footprint.

For example, Alabama’s PUMA 2 in the TRACT_PUMAMINI_pops_.xlsx spreadsheet is composed of Dekalb and Jackson counties. In the 2000 census, this PUMA contained 118,378 people. We could create an exact replica of this 2000 PUMA using 1960 counties. However, in 1960, the population of the PUMA was 78,098. Because of the revised 50,000 person threshold and exact footprint match with the 2000 5% PUMA, this PUMA contains fewer than 100,000 people in 1960.

1 Like
8

Fantastic, thanks a lot!

This is incredibly detailed and useful information. I convinced myself yesterday by comparing the full population counts with estimates from the 1960 5% sample; but it’s great to have this confirmed by you :slight_smile: