Replies to post #81559 on Amarin Corp Plc (AMRN)

[jfmcrr]

I discount "statistical significance" in subgroup analysis.

Why would you do that, when AMRN's instructions to DMC require it?

Some data is worth more than other data....

Yer power goes away with the data mining of portions of the data. Some of the ability to detect significant differences goes away in small groups as a coupla subjects breaking one way rather than the other can flip the decision.

I don't reject it. I discount a subgroup result from , say 130 out of an "n" of 1000.