Odd values in displaced worker survey income data prior to 1994


I’m getting odd values in the displaced worker supplement survey for weekly income prior to 1994.

I get median values for variable DWWEEKL of around $7 until 1994 (but then it’s around 560). I’ve tried looking at distribution tables and can’t figure out why there are so many low numbers. There appears to be a lot of missing variables, coded at 18439, but there’s also quite a few between 5 and 17.

Am I missing something? Thank you for any help


I have verified your observation of very low median values for pre-1994 samples and larger medians for post-1994 samples. A couple things to note about the overall quality of the variable. First, the Description Tab notes that unusually low values, such as zeroes or values near and below one are found in the original data. Second, the universe statement for this variable and the Displaced Worker Supplement is not consistent over time. From 1984-1992, respondents who lost or left a job during the past five years are included in the supplement; from 1994 on, those who lost or left a job during the p ast three years are included. Therefore, your observation could be caused by a combination of these details. Unfortunately, there is not a whole lot that can be done to address these issues.