Trivial Effect of Multiple Testing on a Popular Test for Overall Disease Clustering

• Activity Number: 85
• Type: Contributed
• Date/Time: Sunday, August 3, 2014 : 4:00 PM to 5:50 AM
• Sponsor: Section on Statistics and the Environment
• Author(s): Matthew Loop*+ and Leslie McClure
• Companies: University of Alabama at Birmingham and University of Alabama at Birmingham
• Keywords: spatial clustering ; prevalence ; Monte Carlo simulation ; multiple testing

Abstract

When studying the prevalence of a disease, geographic locations of cases and controls can be used to investigate potential spatial variation in disease prevalence. When locations and sizes of potential clusters of disease cases are unknown, general tests of overall clustering are used as exploratory tests. Because the potential ranges of clustering are also unknown, investigators often test for clustering at many ranges. However, corrections for multiple comparisons are rarely made, and it is unknown whether this omission affects the overall type 1 error rate of these tests. We demonstrate that in small datasets, testing at multiple ranges does not inflate the type 1 error rate of a test based upon the difference in Ripley's K functions, which is the most popular test of overall clustering in epidemiology studies.