Calculating Quarterly Unemployment Rates by Race for CA

I’m trying to calculate quarterly unemployment rates by race in California for the following quarters: Dec 2019 - Feb 2020, Mar 2020 - May 2020, Jun 2020 - Aug 2020, and Sep 2020 - Nov 2020.

I used the following code in SAS to weight my estimates:

Proc surveyfreq data=CPS.ca6;
table race_eth_code * month * unemployed / row;
weight WTFINL;
ods output CrossTabs=test1;
title “Unemployed by Race”;

The numbers I get are extremely large. For example, when I pool months Dec 2019 - Jan 2020, I get a total unemployed population of 2,548,528. This seems extremely large for a 3-month period in California. Am I doing something wrong?

I believe your problem is because you need to adjust the weights to account for pooling multiple samples. Each sample’s weights sum to the full population. So if you’re pooling three samples, you need to divide the weights by 3.

Great, thank you!