Using Biological Insight and Pragmatism When Thinking about Pseudoreplication

Abstract

Pseudoreplication is controversial across experimental biology. Researchers in the same field can disagree on whether a given study suffers from pseudoreplication and on to what extent any pseudoreplication undermines the value of a study. A recent survey indicated that concerns about pseudoreplication can strongly impact peer review of manuscripts submitted for publication. Here we explore controversies around pseudoreplication, identify issues requiring resolution, and in each case offer a resolution. We emphasise that having non-independence in data points and pseudoreplicating are not the same thing. Researchers should be able to demonstrate that in a given experiment they have minimised and controlled the risk of non-independence weakening their study. If they do that to the satisfaction of others, they have avoided pseudoreplication.

Keywords: experimental design; independence of data points; model selection; multilevel analyses; replication; sampling.