Replies to post #712 on Dendreon Corp fka DNDNQ

[ocyanblue]

I will check with IR to see whether or not a two-step analysis was done with 9902a. One issue with 9902a is that it's a small trial cut off long before its enrollment target. So imbalances are more or less expected especially with many covariates being considered. Nonetheless, the inclusion of the two (or three depending on where we want to set the p-value cut-off) non-significant covariates in the analysis says that no game was being played in trying to adjust the model to fit the data.