[exe34]

> Much data may not be representative or usefully representative of reality due to complexity of the situation

I've personally been on the receiving end of "the data we collected shows...x" (in a non-medical setting), but when I asked to have a look at it, it turned out that while this was true for a large part of the population sampled, there was a material difference between that population and a smaller population that can be clearly identified and for the latter, the data showed the exact opposite conclusion.

(think 100 men and 30 women, kind of scenario, except the difference wasn't gender, but job role).