We describe the goals of replication and statistical rules of thumb for distinguishing chance from true associations in the first section of this article. We then discuss the importance of exact replication—seeing a consistent association with the same risk allele using the same analytic methods across multiple studies—and describe analytic methods for combining evidence across multiple studies, along with their relative advantages and disadvantages. We close by discussing why an association may fail to replicate and place replication efforts in the wider picture of contemporary genetic epidemiology, with its focus on large-scale collaborations and data sharing.
