Within-Unit Coverage

Within-unit coverage refers to the accuracy of selection of the respondent to be interviewed after a household is contacted, generally by phone or face-to-face methods, and the informant in the household identifies the prospective respondent, according to the interviewer's instructions. Theoretically, each eligible person living in the unit (e.g., all adults) should have a known and nonzero chance of selection. A number of probability, quasi-probability, and nonprobability methods exist for choosing the respondent, and only the probability methods allow estimation of chances of selection. Humans, however, do not always behave according to the logic of probability methods, and this leads to errors of within-unit coverage. Missed persons within households contribute more to noncoverage of the Current Population Survey, for example, than do missed housing units.

In ...