I am using the ACS waves from 2000 to 2016. In this data I am particularly interested in tracking individuals' ancestry and/or their mothers' place of birth.The raw sample (before any manipulation) counts circa 85 million observations, among these only 25 millions have informations about ancestry and/or mother's place of birth. In this count, 2.5 millions observations for which these informations are coded as "not reported" are also included. I would thus like to make sense of the remaining missing values ("true missing") that do not show any specific coding (in Stata format this corresponds to "."). Because the number of true missings is constant over different years, and also for different variables (e.g., mother citizenship status) I am wondering if mothers' information have been collected only for random subsample.