Why so many NIU in CLASSWRK or in JTCLASS?

[Pilar_Cuevas_Ruiz]

I am trying to analyze the number of self-employees of the full sample. I can get them from the variables CLASSWRK or JTCLASS, however I got NIU in almost 90% of the observations. How is it possible? Aren´t the 90% of the population addressed by the survey working? Thanks in advance

[Matthew_Bombyk]

If you look in the Universe Tab for JTCLASS you will see that this variable from the Job Tenure supplement has a very restricted universe:

2000+: Civilians age 15 or older who were employed during the reference week, not doing the same work last year, were working for the current employer less than a year, and have a valid entry for the occupation variable specific to the Job Tenure supplement.

On the other hand, CLASSWKR is asked (since 1989) for all people over 15 who ever worked. You should have many fewer NIU cases for this variable.

Note that in most cases, CLASSWKR refers to the respondent’s current employment. JTCLASS is only asked for people who left their job in the past year, and is in reference to their previous employment.