# Some issues with HRS (Health and Retirement Study) data

I am using the 1992 HRS (Health and Retirement Study) dataset. The codebook is available at http://hrsonline.isr.umich.edu/modules/meta/1992/core/codebook/14_h1.htm

I am trying to construct the share of HRS respondents that lost their job because of business closure (see the variable 3612 at the link reported above). I cannot understand what is the meaning of: 1. Checked 5. Not checked

What happened to these respondents? There is a note about checking just below the question but I still do not understand what "Checked" and "Not Checked" actually mean.

Finally, the total number of observations in 1992 are 12652. In fact, including all answers, we have that there are 12652 observations (i.e., the summation of the possible answers reported below).

        V3612 Code  Frequency
---------------------
0        6338
1        1175
2         222
5        4888
9          29


It seems that both respondents that were "checked" by the interviewer and respondents that were "not checked" by the interviewer DID NOT ANSWER the question. Am I correct?

• You're probably best off contacting the authors of the study directly. – user7935 Aug 19 '16 at 10:10
• So a check (1) here means `yes' - so the business closing was the reason they lost their job. If not (5). If the question was not applicable they would get a (0) and other responses would be 2 or 9. – VCG Aug 21 '16 at 12:37