Multiplicity Control in Large Data Sets Presents New Challenges and Opportunities
In recent years, “multiple testing” or “multiplicity” has taken central stage as an issue of interest and urgency in the statistical community. This is related to, but not in my opinion the same as, the classical question of “multiple comparisons.” Although the terms are often used interchangeably, I think multiple testing and multiple comparisons have different connotations and represent two different inference problems. The former is of particular relevance in the analysis of large data sets.Some content is only viewable by ASA Members who subscribe to CHANCE. If you are an ASA member, log in to Members Only and look for CHANCE under your publications.