Insmed Inc (INSM)

[Guaranteed Sceptic]

Dear Digforgold196 and schwabd

Let me point you to message #2320 (yes, one of mine). TRCA isn't saying that p=0.48 is signficant. They are saying "Hey, it's NOT AT ALL significant", i.e. their product isn't causing hypoglycemia. I am not a fan of TRCA but let's not credit them with such a level of stupidity as to think that p=0.48 is significant.

And, at the considerable risk of starting an MD vs PhD flame war, most of the physicians I've seen do statistics are happy applying sequential t-tests to a whole slew of comparisons in one study. A corollary to this is the assumption that p<0.05 is a magic border between nonsense and science. To further alienate my fellow bench workers, there was an interesting study (sorry I've lost the reference) by some professional statisticians who looked at several major research journals for the past few years and concluded that most of the articles didn't use the appropriate form of analysis and a surprising number wound up with the wrong conclusion by so doing.

So a lot of us (myself included) don't do statistics properly and we can conclude almost nothing from a press release such as that which TRCA threw out today. The one thing that the prof for my advanced stats course drummed into us was to talk to a statistician before starting the experiments and not after. Smart advice that I've ignored for many years.

GS