CPS: ASEC vs "Core" vs March Basic Monthly

Are the variables labeled “Core” under the Select Data -> Select Variables menu from the ASEC or the March Basic Monthly? If my samples include all the ASEC samples but have not selected the March Basic Monthly samples, will I still get these variables?

Does the March Basic Monthly include additional variables beyond the list in Core plus ASEC? If so, is there a straightforward way to get just the Basic Monthly variables incremental to the Core, omitting any variables that are duplicated?

Id also like to observe that there are groupings under the Core and the ASEC that have the same name, and that references to variables in these groupings are somewhat confusing and ambiguous. See, e.g., “WORK VARIABLES – PERSON,” whicch is the designation of both the fourth group under Core and the first group under ASEC.

Thanks for asking this question, as I can understand how these details may be a bit confusing. The “core” variables are variables that are part of both the Basic Monthly samples and ASEC samples. Therefore, if your extract contains only ASEC samples (and no Basic Monthly samples), these variables will still be available.

The March Basic Monthly file is just like any other Basic Monthly file. The ASEC includes all March Basic Monthly Survey respondents as well as oversamples from other months (see more information on oversamples). If you want to include both the ASEC and the March Basic Monthly Survey, note that both samples have a MONTH value of 3. To make it simple for users to distinguish data from one file versus the other, we have created ASECFLAG.

Finally, I do understand that the design of the CPS is a bit confusing. In particular I agree with your note about the two variable groupings called “WORK,” one under the Core category and the other under the ASEC category. This is because there are a group of work variables that are included in both the Basic Monthly and ASEC samples as well as a group of work variables that are only included in the ASEC samples. We have tried to design the IPUMS CPS extract system to help reduce some of this confusion. If you first select the samples you intend to include in your extract, then only variables included in those samples will show up in the variables drop-down menus. Of course, if you have any suggestions for how we can improve the IPUMS extract system please feel free to send us an email at ipums@umn.edu.

I would suggest that “Core” and “ASEC” appear in the variable group names, as well as in the menu.

And as long as we are on the subject of menus, let me again plead for quality variables to show up under Q in the alphabetical pulldown menu. And the quality flag group links ared still broken! Please ask whoever is responsible for maintaining this stuff to go to the page for any quality variable. At the top, they will see a group link like all the variables have, e.g. “Group: Data Quality Flags — PERSON”. But if you click that link, instead of taking you to the group, it just takes you back to the menu page. This really is a broken link, not just an inartful presentation. Could you convey my strong request that we get at least one of these ways of finding quality values fixed sometime soon?

The data qualoity flags have become more important in recent years, because we now know that the imputed values are downwardly biased from the true values. In addition, to the best of my knowledge, the entire tree of income totals, subtotals and ultimate values is nowhere presented. I am trying to create a function that provides that tree for every year, and checks whether they add up, and another that tells you if the income total includes top-coded, bottom-coded, or imputed values. Since only ultimate values, and not all of these, get quality flags, a list of quality flags would be a valuable way of checking whether values are ultimate or intermediate, and of identifying missing ultimate values.

I’ve forwarded your suggestions to the IPUMS CPS Team. I agree that something should be done to make the search for data quality flags more simple for users. One of the ways the the extract system tries to make this easy is by including the “Select Data Quality Flags” option to the “Extract Request” page. This option allows users to add the data quality flags associated with any of the variables included in their extract. That being said, I think that this functionality can be improved particularly in the case of the income variables.

<strike forehead with heel of hand> Right! Of course! I already asked for all the data quality flags in my three extracts that contain household and personal income. That means I have already downloaded the data I’ve been complaining I couldn’t get.

You know, I envision a plausible pathway to a unified global human consciousness, as a security glitch in Google Cortex, an app to let you search your own brain. When such an app is released, I’d install it in a second.