Hi, I have been trying to run an analysis on SDA using the Study: CPS BMS: 1976-Present. I have not been able to get it to run. Just checking if the system is down. Thank you.
Rogelio Saenz
Hi Rogelio! I was able to run a quick tab of YEAR by EMPSTAT in the CPS SDA tool for the 1976-Present BMS just now. Could you please try again or share details about the specific analysis you are trying to run? Thanks!
Thank you. I still am having trouble. I turned off the computer and tried again with the same problem.
Here is what I am trying to run:
Thank you for your patience. I was able to create a table when I dropped raceethn from the selection filters, so inspected that “user-defined” variable further.
When I try to tab just raceethn, I am only able to produce unweighted results. When I try to produce weighted results, I receive an error message that the number of cases in raceethn doesn’t match the number of cases in WTFINL. This discrepancy is logical because it looks like you assigned any unspecified combinations of your input values for raceethn to missing.
After toying around with SDA for a bit, it seems that SDA will still allow you to use raceethn as a selection filter even though the number of cases are different, if that is the only variable specified in the selection filter. As soon as I add any other variables to that category, I encounter the same issue you describe (even if I run the analysis without weights). You can move your restriction on the year variables to the control field (i.e., CONTROL: year(2024-2026)), but there isn’t enough flexibility without selection filters to address your EMPSTAT criteria.
I think that assigning the missing cases in your raceethn variable to a specific code (which you can then filter out from your analytical sample in SDA) should allow you to move forward. I will consult with colleagues to determine if the selection criteria issue you are encountering is intended (but lacking an appropriate error message) or a bug.
Thank you very much, Kari. This is very helpful. I went back to use the HISPAN variable instead of the created RACEETHN variable and it worked fine. Thank you again.
Rogelio
1 Like