Biotech Values

[rfj1862]

Poorly answered the question on data quality (lots of data points had a p value of less than 0.05). Specifically, when asked to provide color on how far away from 0.05 - the reply was "I don't worship 0.05" and in really bad form. Then the question was not answered which in reality told you it wasn't far away.

I assume by "less than 0.05" you mean that on their actual slides it said "P<0.05."

From a statistics standpoint, asking "how far away from 0.05" is an absolutely terrible question that shows zero understanding of how statistics work. There is no such as thing as more or less significant if the data has reached the threshold of significance. A paper that has actually been reviewed by a statistician in a reputable journal would never, ever be allowed to use qualifiers like "highly statistically significant" nor would they allow you to say "approaches statistical significance." It is either significant or not per the design.

In other words, you cannot make a value judgement on the quality of data if P<0.001 vs P<0.05.

A better question is whether the delta is *clinically significant*. In some fields, the minimal clinically important difference (MCID) can be defined prospectively.

I have no position in HEPA and sure as hell never will FYI