One of the difficulties in attempting to assess changes in SOC is that the magnitude of any change tends to be very small in relation to the variation (Hungate et al., 1995 and Kravchenko and Robertson, 2011). This high coefficient of variation (CV) in turn results in low experimental power, i.e. a poor ability the separate any signal from the noise. Shi et al. (2013), Upson and Burgess (2013) and Harper and Tibbett (2013) have all highlighted the importance of sampling sufficiently deeply for SOC. However increasing the depth of sampling also lowers the experimental power as the CV tends to be greater with increased sampling depth. For this reason, post-hoc power analyses are useful for determining if the experiment can be expected to result in type II errors, i.e. the incorrect conclusion that there are no differences in SOC between treatments.