Pseudoreplication occurs when treatments
are not replicated but a t-test or an ANOVA (analysis of variance) is
carried out anyway by assuming sub-samples (or in some cases individual
trees) are the same as replication. This results in use of an incorrect
number of error degrees of freedom. A major factor that leads to
pseudoreplication is a failure to define the correct experimental unit.
In addition, some journal editors do not require ANOVA tables be
included in manuscripts. As a result, a reviewer might not know if the
error term involved 13 or 2841 degrees of freedom. Reviewers who would
normally reject nonreplicated trials can sometimes be fooled into
thinking that a statistical analysis (that involves pseudoreplication)
is valid
