a thoughtful web.
Share good ideas and conversation.   Login or Take a Tour!

Because you won't be able to make a point. It just goes against the whole point of scientific peer review and the way we do science.

You would basically increase the irreproducibility from its current 50-60% to 90% because everyone was playing wishful thinking just to not be able to get the results they wanted at the end. And then you would have to retract everything. Except, no one would do that. So the thing they will do is twist that story until it shows what they want to show.

Leading to very bad quality data. This starts a chain that could fuck science for another decade because people will build hypothesis on them, apply to grants etc. and since everything takes so long nowadays, you will lose 5-10 years of "for nothing" data.

A recent example from my field of research. We got a paper to review from Science. We check the paper, the quality is garbage, the analysis is horrible. Why that fuck does Science send us the paper for review? Turns out it is from a famous lab in the states (ivy league). We write a very long review giving suggestions for improvement and data analysis.

Paper is rejected in Science and turns up 5 months later in Neuron. A respectable journal in my field. Unaltered. None of the suggestions implemented. Zero.

The results and conclusions of the paper are worth nothing. However, we just got a grant proposal to review from another lab basing their whole hypothesis on the results from that one paper.

Luckily, we were able to tell the poor guy that the paper is problematic. But many others won't have our experience or would just believe the paper. A few millions in research investments later, all of it was a fluke. OR they would also twist the data to the point that they agreed with the paper from Neuron. Because it is Neuron and so on and so forth.

Usually, trends like that die out because there will be enough reports going against them.

However, the game changes if you have much much more of those cases with data that has "gaps" or "filled gaps" where squares are pushed into circles just to get the same conclusion because the authors fear retractions. The damage would be insane.