An error in the survey instrument resulted in information about layoff and looking for work not being collected for people who did not work in the reference period (EPDJBTHN=2). Changes were made to both the edit and imputation specifications in an attempt to compensate for the error, but such methods, however good, are unlikely to create a perfect substitute for reported data. This means that the quality of the data for the variables that directly or indirectly use the missing information is likely not as high as it would have been had the instrument problem not occurred.

The variables affected the most are:

ELKWRK (looking for work)
ELAYOFF (on layoff)
EWKLKG01-EWKLKG18 (weeks looking for work)
RM1WKLKG- RM4WKLKG (monthly recode for number of weeks looking for work)
RM1ESR- RM4ESR (monthly recode for employment status)

The extent of the actual damage is probably small, at least at the macro level, because the large majority of non-workers likely were neither on layoff nor looking for work in the reference period, and because the damage-control measures were probably at least moderately successful.

The user should be especially wary in using the RTAKJOB and RNOTAKE variables, since they were not edited to account for the instrument error.