In the 1901 and 1911 Ireland files, the occstr variable is often blanked with a code such as " BBBBBBBBBBBBBBBBBBBBBBBBBB". Why is this? Given that blanked entries correspond to a variety of occhisco values, there must have been some useful information in the original strings. Also, is some field used to code occhisco in addition to the occupation string? For example, among the 1901 entries with the string “Scholar”, 64 are coded to occhisco values other than the usual (and seemingly correct) “99999”.
Any cases of blank occupation strings (represented by the string “BBBBBBBBBBBBBBBBBBBBBBBBBB”) are present in the original data. IPUMS has not suppressed or censored these variables.
Historians at IPUMS first used the original occupation strings to code the OCCHISCO variable for these samples. In some cases where the occupation string was blank in the original data, these IPUMS researchers assigned occupation based on other information about an individual and/or their household and household members. For example, if someone was part of a farming household, their occupation may have been imputed as farmer or agricultural worker. In general, the OCCHISCO codes for those with missing occupation strings are the result of both case-by-case decisions and programmatic assignments based on what was learned in hand editing. In the case of Ireland specifically, the assignments would have been largely based on a coding dictionary developed through a more manual process of assigning these for the UK data files and a number of other historical censuses.
OCC was coded later, and relied heavily on OCCHISCO codes.
Unfortunately, I do not have detailed information on the exact procedures used to impute OCCHISCO or OCC codes in cases where the original occupation string is blank, and there certainly may be imputations that were made programmatically or individually based on human discretion that are puzzling.