Hello,
I couldn’t find via a Google search how exactly the census/IPUMS recodes the responses to the census question about educational attainment in the most recent ACS waves to the variables EDUC and EDUCD. Is there a detailed guide somewhere? The reason I want to know is because I would like to ask about educational attainment in my survey and code the responses the same way as the census to maximise the fidelity of the weighting by the census frequencies I’m planning.
Thank you,
Taym
IPUMS obtains the public use microdata sample (PUMS) from the Census Bureau, and recodes a number of variables in order to harmonize them, making them comparable across time. The IPUMS variable EDUC is recoded from the source variable, which in the 2023 ACS is US2023A_SCHL. You can see the source variable(s) associated with each harmonized IPUMS variable by clicking the “source variables” tab on the variable. For example, you can see in the source variables tab for EDUC the source variables that are used to create EDUC in each sample you’ve selected in the extract system.
The easiest way to determine how the source variables from the ACS correspond to IPUMS variables is to create data extract that includes both the harmonized variable and the source variable, and cross-tabulate them. This allows you to see how values of the source variable were recoded into the harmonized variable. Some variables have editing rules that determine in part how the source variable is recoded. You can read about these in the editing procedure tab of each variable. For example, in the editing procedure tab of the EDUC variable, you’ll see extensive information on an editing rule for EDUC that affects the 2005-2015 ACS samples.
Please note that the frequencies displayed on the pages for US2023A_SCHL and US2022A_SCHL, the source variables for EDUC in the 2023 and 2022 ACS, respectively, are incorrect. If you add these variables to an extract you will see the correct frequencies in the extract. We are working to correct the issue with the frequencies posted on the linked pages.
1 Like