Misclassification of Categorical Data

Proceedings of the American Statistical Association, Section on Survey Research Methods, 1985.

Project details

There are several situations where misclassification of categorical data can be observed.  Dual system estimation studies (used for evaluation of the coverage of a census) match records from a survey to a census, where both studies have collected basic demographic variables like age, race, sex, and ethnicity.  Studies are also conducted matching survey results to administrative records, such as health studies which match survey data to patient records.  Reinterview studies are conducted to evaluate the accuracy of censuses and surveys, or to monitor the work of interviewers.  Finally, panel studies designed to measure change over time of certain variables will often collect the same demographic information if there is a sufficiently long time between interviews.

-Cowan, Charles D.

Back to Publications

Related Projects