Stats duded wanted, page-8

cafa,

Not sure what you are trying to ask. Are you saying if the 600 trial produces a p-value of 0.01, what would the 1000 trial p-value be if successful? If this is what you are asking, then provided the additional 400 people produced a result as good as the original 600 people (ie 64% survive with treatment whereas 33% without treatment for both the original 600 and the additional 400, or something like that) then the p-value will be smaller than 0.01 for the 1000 trial. If the OS rate was higher for the additional 600, then the p-value will be even smaller again.

Now this is where it gets tricky. If the additional 400 perform slightly worse than the original 600, then there is a chance that the p-value will be 0.01, smaller or larger. This is because the sample size has increased and the sampling distribution has become 'slimmer' (ie the standard deviation has decreased due to the increased sample size) but the data is better.

If the additional 400 is much much worse than the original 600, then the p-value will be bigger than 0.01 for the 1000 trial.

If you are asking what the p-value of the 600 trial will be if the 1000 p-value is 0.01, well it depends on how the additional 400 compared to the original 600. If they are the same, the p-value of the 600 will be more than 0.01 because the sample size has decreased as the standard deviation of the sampling distribution has increased and made the distribution 'fatter'.