Hi! In the DWS there is one variable each for “hourly wage at current job” (dwwagec), “hours worked each week at current job” (dwhrswkc) and “weekly earnings at current job” (dwweekc). In some cases dwweekc does not correspond to the product of dwwagec and dwhrswkc, in others it does. My questions are the following: 1) Do all survey respondents answer all three questions? 2) If the value based on dwweekc and the product of dwwagec and dwhrswkc gives the same value, which underlying survey question is that value most likely based on. 3) Is it possible to make a statement about whether dwweekc or the product of dwwagec and dwhrswkc is a strictly better measure for actual weekly earnings at the current job? Thanks a lot!
Sorry for the delayed response. I was discussing some of these questions with the IPUMS CPS Team. I will aim to address each of your questions one at a time. The below answers are valid for 1994 and onward, in which all three of the variables in question appear. Prior to 1994 only DWWEEKC is present and it is self reported.
(1) Both DWWEEKC and DWWAGEC are recoded variables, sourced from multiple questions. You can find the corresponding questions in the Questionnaire Text tab (Note that relevant samples will need to be selected for the text to appear).
(2) The DWWAGEC and DWHRSWKC variables have the same universe. You can find that info in the Universe tab.
(3) Looking into the questionnaire we have a suspicion that the discrepancies you observe have something to do with overtime pay or tips being added into DWWEEKC and not necessarily being reflected in hourly wage or number of hours worked per week. Do note however, that since we do not have an overtime pay variable available for the displaced worker supplement, we are not able to verify this suspicion. Moreover, DWWEEKC is allocated for missing responses. Based on these details, DWWEEKC is likely the better measure of weekly earnings.
1 Like
Thank you, Jeff!