Power Analysis

To estimate the number of experimental nests I needed, I carried out a power analysis. The pilot study was too limited in number of replicates to generate an accurate estimate of the effect size of edge effects, so effect sizes were gathered from a variety of papers who had performed similar experiments \cite{KOTZE_2007,Cooper_1998,Vaaland_Burkey_1993,TEMPLE_1988}. An average was taken which indicated a low effect size (w = 0.256). Using the default values of  p = 0.05 and power = 0.8, this showed that N=119.8, meaning 120 nests would be used for the full experiment.