Finally, small samples in a field make replication attempts
particularly difficult. If an effect exists, the probability of finding it
in two unrelated studies (i.e. the probability of replicating an
existing effect) is the product of the power of both studies. When
samples are small, a failure to replicate (i.e. to find the same effect
in both studies) is ambiguous: It may either mean that the effect
really does not exist or that the study did not have enough power
to detect it. Additionally, the smaller samples in a field are, the
more effect sizes vary, which makes it likely that effects from
different studies differ.2 Therefore, a field with underpowered