For the diffeye (after year 2009) variable denoting whether a person is blind or visually impaired, I have observed that for working age adults (age>=25 & age<=64), a small proportion of respondents are coded “NIU”. My question is what type of respondents would be coded “NIU”? Should I include them in the denominator if I want to compute an estimate of the prevalence rate of blind or visually impaired individuals? If so, how to incorporate the “NIU” individuals?
For these variables, the universe is civilians age 15+. You can find this on the Universe Page for DIFFEYE. The NIU you see are most likely in the military. I would suggest excluding them from any calculations.