3 ms·
>why are scientists falsifying data like in ClimateGate? You've repeated this accusation several times in this thread. Can you back this up? I've asked people
by epistasis 15y ago
>why are scientists falsifying data like in ClimateGate?
You've repeated this accusation several times in this thread. Can you back this up? I've asked people many times, nobody ever produces anything. They just assume that it was proved, without ever actually checking their hypotheses.
Here's an example of what scientific misconduct looks like, from my field:
http://www.economist.com/node/21528593 http://www.economist.com/node/21528593
The researchers screwed up through basic sloppiness, then refused to correct their errors after they were pointed out. Has there been anything this bad in the the leaked climate emails? BTW, as a scientist I know that "trick" has more than one meaning, particularly when it comes to problem solving, so if you think that the use of the word "trick" is rock-solid proof, please explain how deception was perpetrated rather than clever problem solving.
- gavanwoolery 15y agoReally, ClimateGate is the least of my concerns as far as "bad science" goes - most of the bad science is not being hidden behind conspiracies, it is in our face. To be honest - I do not know if the researchers really hid or falsified data, and I do not know what attempts were made (if any) to further cover this up. But I do know that it did come into public light, and that raises my suspicions a little bit. But regardless if they did hide anything or not, I am more concerned with the fact that scientists are so eager to prove that AGW has a major impact and seemingly hesitant to try and prove the opposite.
- epistasis 15y agoSmoke and mirrors, smoke and mirrors...
- sesqu 15y agoThe hypothesis testing approach to statistical inference, which is what most of the scientific community uses, explicitly demands researchers try to prove that, say, AGW has no major impact, before reporting on their failure to do so. The hypothesis of a major impact is suggested because negligible impact cannot be, in the face of data. The popular Neyman-Pearson test doesn't even account for failing to detect an effect - it concerns itself strictly with not seeing ghosts.