Ann: iPPS SIGNIFICANT OA PAIN REDUCTION AT 12 MONTHS, page-162

It depends on how close the sample mean is to the population mean. A larger trial is more representative of the population mean than a smaller trial, so the results of PARA 002 will be more representative of the population mean than say 008 and 005. The idea of running a smaller trial, particularly two of them is to provide a confidence estimate of the population mean so you can then determine the beta (or difference between placebo and treatment groups) to ensure the larger study is powered to achieve statistical significance.

In other words, if the results of 008 are representative of the population mean, then a larger study will lead to a tightening of the confidence intervals and therefore greater separation, improving the chances of achieving statistical significance.