Accounting For Non-Independent Observations In 2×2 Tables, With Application To Correcting For Family Clustering In Exposure-Risk Relationship Studies

Authors

Leslie A. Kalsih, New England Research Institutes, Watertown, MA
Katherine A. Riester, Biogen, Inc., Cambridge, MA
Stuart J. Pocock, London School of Hygiene and Tropical Medicine, London, UK

Abstract

Participants in epidemiologic studies may not represent statistically independent observations. We consider modifications to conventional analyses of 2×2 tables, including Fisher’s exact test and confidence intervals, to account for correlated observations in this setting. An example is provided, assessing the robustness of conclusions from a published analysis.

DOI

10.22237/jmasm/1036109100