Recommended tests and confidence intervals for paired binomial proportions

We describe, evaluate, and recommend statistical methods for the analysis of paired binomial proportions. A total of 24 methods are considered. The best tests for association include the asymptotic McNemar test and the McNemar mid‐ p test. For the difference between proportions, we recommend two simple confidence intervals with closed‐form expressions and the asymptotic score interval. The asymptotic score interval is also recommended for the ratio of proportions, as is an interval with closed‐form expression based on combining two Wilson score intervals for the single proportion. For the odds ratio, we recommend a transformation of the Wilson score interval and a transformation of the Clopper–Pearson mid‐ p interval. We illustrate the practical application of the methods using data from a recently published study of airway reactivity in children before and after stem cell transplantation and a matched case–control study of the association between floppy eyelid syndrome and obstructive sleep apnea‐hypopnea syndrome. Copyright © 2014 John Wiley & Sons, Ltd.