A certain number of housing units are selected to be included in the survey, but the Census Bureau is not able to collect data on all of them. In some cases, addresses cannot be found or do not meet the Census definition of a housing unit. In other cases, respondents refuse to be interviewed. Thus, the completed sample size is different from the planned size. In addition, because the AHS is a longitudinal survey, some cases may be refusals in one year but not the next because the previous occupant may have moved.
For sample size information, see our Definitions, Accuracy, Historical Changes, and Questionnaires page.. Click on the survey year of interest and see the “Sample Design, Weighting, and Error Estimation” documents listed under the Accuracy of the Data section of the webpage.