I am looking for data on school enrollment in 1985 (survey year). SCHLCOLL captures this information from 1986 onwards. I am replicating a paper that uses CPS data from 1985-1991 but does not mention that SCHLCOLL 1985 is missing (i.e. the authors makes it look like the data exists: “Data are from survey years 1985-1987 and 1989-1991 of the March CPS. The sample contains unmarried women between the ages of 16 and 44. We exclude women who where separated during the previous year, ill or disabled, in school.”) Unfortunately, the actual data set used in the study does not exist anymore. It appeares to be impossible to replicate the results with enrollment data missing for 1985. I am looking for clues on whether it is possible to obtain or construct the type of information SCHLCOLL provides for the survey year 1985. Any input is very welcome – Thank you!
Is there a way to obtain school enrollment data for young children prior to year 1986?
The IPUMS-CPS variable SCHLCOLL is actually composed of three CPS variables (in March 1986 these variables were called battend, bhscolge, and bshlftpt). Comparable variables are not available prior to 1986. There are, however, a number of variables that have “At school” as a possible response category. Most of these variables are actually about work so IPUMS-CPS did not include them as Education indicators. For instance, EMPSTAT has the response option, “Not in Labor Force, School”. MAJORACT gives the respondents major activity the previous week, with “At School” being one of the answer options. These variables are not perfect measures of school attendance, but may be what the authors used in 1985. Of these two variables, I would recommend MAJORACT as EMPSTAT only identifies school attendants who are not in the labor force. You should also note that MAJORACT and EMPSTAT have a different universe than SCHLCOLL so you will want to limit respondents to Civilians 16-24 to match the universe of SCHLCOLL over time.
I hope this helps.
Thank you Joe_Grover! You answer was - again - quite enlightening. I was not aware of MAJORACT. I will see what I can do with that while I am still hoping to hear back from the authors. That EMPSTAT and MAJORACT are limited to 16-24 would indicate that they were not used in this study, but I will definately see what I can do with MAJORACT.
Ah - that’s a good clarification. Indeed, I had confused the universe of the two when writing this.