ABSTRACT

Pseudo-replication is a phenomenon whereby the observations which are taken as replicates do not necessarily reflect the variation in the population of experimental units about which information is being sought. As has been said, it occurs when the replicated observations are not independent of each other. It can occur in experiments involving field crops or other field plants such as weeds by applying a particular treatment to a continuous part of the field in one sweep and taking measurements from the treated area and considering them as if they were from properly replicated plots. Such a practice ignores the variation induced in observations by inaccuracies in independently applying the treatment to the individual plots and the effect of spatial variation over all the field. Consequently the uncontrolled variation amongst the pseudo-replicates is likely to be smaller than the variation which would have been observed with independent replicates. It is therefore unwise to rely on standard errors which are estimated from pseudo-replicates instead of independent replicates as they will be biased downwards.