Join MEPS Full Consolidated Files with Event Files

HEALTH SURVEYS
1

TLDR: Is there an example to join the IPUMS MEPS Full Consolidated Files with any of the MEPS Events files?

The IPUMS MEPS Full Consolidated Files, at the most aggregated level, are at an annual level. So for each MEPSID primary key, there are two rows.

What will be the best way to join the Consolidated File and any Events File. I have written two options, please suggest a better one if these are not good enough.

Option1: Can I use a combination of MEPSID+YEAR from the full consolidated files and then DUPERSID+PANEL+xxDATEYEAR from the events file?
Option2: Can I use a combination of MEPSIDRD+ROUNDRD from the full consolidated files at round level and then DUPERSID+PANEL+EVENTRN from the events file?

2

The IPUMS MEPS team is working on documentation on linking IPUMS MEPS to other AHRQ MEPS data (including event files), but it isn’t quite ready yet. I will answer your question directly, but keep an eye on the IPUMS MEPS user guide page for additional information on this soon.

The approach you describe for Option 1 is appropriate; however, I want to share a bit more information and will suggest a slight modification.

First, while I would not expect xxDATEYEAR variables on the event files to include values other than the survey year, rounds (and correspondingly their reference periods) can span more than one calendar year. Because of the complexity of timing information, I would not be surprised if there was occasionally an unexpected value for year in the file. I suggest making your own year of survey variable after reading in the event data and using that to link.

Secondly, AHRQ recently revised DUID (and DUPERSID) to append PANEL to the beginning of the variable. IPUMS MEPS will be updating our documentation and unique identifier accordingly (and retiring the current version of MEPSID–it will still be available under a different variable name). All that to say, review both the IPUMS and AHRQ documentation carefully as the order of the concatenated components of these identifiers is changing/has changed.

3

Thank you @KariWilliams.

I already had the made a new variable related to the year of the data file before stacking of events from multiple years’ events’ files. I used that with the YEAR column from IPUMS MEPS. It has worked from now.

Thank you for replying. Thanks for updating me on the ARHQ changes and IPUMS expected changes of MEPSID.

1 Like