"Accept that your data gathering techniques will never be 100% accurate. Work to improve them whenever possible. Use the best information available to make the most informed decisions possible."

Who in science & engineering would criticize this approach at first blush?
However, sometimes the COST of implementing a decision based on imperfect data is very high in economic terms. The risk of being wrong then makes it imperative to get more independent data before puting expensive measures in place.

Reasonable people on the other side argue similarly that the cost of not doing X could be even higher. Therefore it is best that we should act now, even with imperfect data.

That's the rub.

Add in that there are people, humans supplying the data. And that economic interests (business on one side, govt study grants on the other) have on occasion been known to influence or bias results and you see the inherent opportunity for reality distortion.

-p