RTI uses cookies to offer you the best experience online. By clicking “accept” on this website, you opt in and you agree to the use of cookies. If you would like to know more about how RTI uses cookies and how to manage them please view our Privacy Policy here. You can “opt out” or change your mind by visiting: http://optout.aboutads.info/. Click “accept” to agree.
This article begins by reviewing the measurement error model proposed by , with particular emphasis on their concept of simple response variance. More recent developments in the modeling of measurement error are linked to their model and are shown to be extensions and generalizations of their essential concepts. The index of inconsistency, which was first formally described in their paper, is shown to have at least three interpretations, depending upon the statistical framework adopted for describing the gross differences in an interview-reinterview study. Several examples illustrate and compare their classical methods with more modern approaches that employ latent class analysis to estimate the error parameters. It is shown that use of estimates of simple response variance for survey evaluation may obscure important error structures that are more visible using estimates of classification error probabilities to assess data quality.