The problem of matched-pair studies with misclassified ordinal data is considered. Misclassification is assumed to occur only between the adjacent columns/rows. Bias-adjusted generalized odds ratio and a test for marginal homogeneity are presented to account for misclassification bias. Data from lambing records of 227 Merino ewes are used to illustrate how to calculate these bias-adjusted estimators and – because validation data are not available – a sensitivity analysis is conducted.
"Matched-Pair Studies with Misclassified Ordinal Data,"
Journal of Modern Applied Statistical Methods:
1, Article 8.
Available at: http://digitalcommons.wayne.edu/jmasm/vol10/iss1/8