Based upon our simulated data, the below table lists how often across simulations we would proceed to a full trial when applying the below decision rules to our simulated pilot study results.

We now do a power analysis to determine the sample size required to detect effects of a certain size in a full trial. This is done in two ways: once using the pre-determined practically significant effect size, and once using the effect size observed in the simulated pilot study. The table below shows, across all simulations, the average sample size needed to detect effects in a full trial.

PNG PDF EPS

PNG PDF EPS

We now make conclusions about the effectiveness of the intervention based upon our full trial.