The degree of replication between candidate-gene studies
is frequently poor. This can be due to various factors,
such as small sample sizes (and hence inadequate study
power), unidentified population stratification or a failure
to correct for multiple testing during statistical analysis.