Hi. I am using the following crosswalks for best-matching 2010 PUMAs to 2013 MSAs:
2012-2021 ACS and PRCS samples:
I have a question about the match summary file. I am seeing PUMA-approximated MSAs whose populations add up to much more different than the match summary suggests. For example, for Eau Claire, Wisonsin (2013 MSA 20740) there are two PUMAs that best match the MSA: 55101 and 55103.
In the match summary file, it says that the 2010 population of Eau Claire MSA is 161,151, and the 2010 population of the Best-Matching MSAs is 152,254, with an omission error of 5.5%.
However, when looking at the crosswalk file, I see that 152,254 is the population of just one of the two PUMAs listed for this MSA, 55103. The other PUMA for the MSA is 55101 and has a population of 133,662 and includes only 5.5% of the MSA. This makes it seem like though it listed both PUMAs, that only one is truly used as the “best matching PUMA.” Is there some other crosswalk where it only includes the best approximation of the MSA, because when using this crosswalk it has me massively over-estimating MSA populations relative to the match summary suggests.
Thank you very much for any advice.
Haleigh Tomlin, Research Analyst
You’re correct that the crosswalk file isn’t limited to “best-matching” PUMAs. It includes a record for every spatial intersection between a 2013 MSA and a 2010 PUMA. The best-matching PUMAs are only those where a majority of the population resides in the MSA.
To identify the records in the crosswalk that contain the best-matching PUMAs, you can just filter on Column K (“Percent PUMA Population”) to select only the records where the percentage is greater than 50%.
For example, in the Eau Claire case, the crosswalk file includes two records (illustrated below), indicating that the MSA intersects these (and only these) two 2010 PUMAs. Only one of these two PUMAs (55103) has a majority of its population (100%) in the Eau Claire MSA. The other PUMA (55101) has only 6.66% of its population in the MSA, so it isn’t a best-matching PUMA. You could exclude it by filtering out all records where the Column K value isn’t greater than 50%.
The match summary file summarizes how the best-matching PUMAs correspond with the MSA. In the Eau Claire case (illustrated below), Column C indicates that there are 2 PUMAs that intersect the MSA, and Column D indicates that the best-matching set includes only 1 PUMA, which is consistent with what we see in the crosswalk.
1 Like
Thank you for this helpful explanation!
1 Like
