Issues with January 2016 CPS Job Tenure Supplement

CPS
1

I’m attempting to measure job attrition from 2015 to 2016 using the variables JTOCC (occupation in 2015) and JTSAME (have you changed jobs since last year?). I get identical results for all JTOCC categories except “0000” (NA) when I switch JTSAME from “01” (not doing same job as in 2015) to “02” (doing same job). I’m unsure what to make of this.

2

I am not entirely certain of what you are explaining here. In any case, I can note the following. For the January 2016 sample, all observations with JTSAME==2 have JTOCC==0. Observations with JTSAME==1, JTOCC takes on all of the available occupation categories.

3

Thank you.

4

Hi Jeff, in 1996 (only that year) there are a lot of cases where jtsame==2 and JTOCC is valid. Any clarification on what that would be the case?

5

Thanks for pointing this out. I’ve looked into this issue, and it appears that for 1996 only, the JTOCC codes are filled using the OCC codes for most people with JTSAME=2. This isn’t wrong per se, but it does conflict with the stated universe and isn’t consistent with the later years. There are also a number of cases where JTSAME=2 and JTOCC=9997 (an undocumented code). The IPUMS CPS team is looking into this further and will update the documentation based on what they find.