Hacker News Reader

Comment by nomilk | Hacker News Reader

For every false paper, I wonder how many other researchers waste time pursuing research on false premises. And worse, how much do fake papers influence broader society, through policy and individual decision making.

I wonder if any economists have tried to measure the $ cost of fake research.

View on HN

Papers don't establish ground truths in science; they start a conversation. When bad papers are published due to misconduct, there's a cost --- a pointless conversation occurs. But it doesn't shake the foundations of science. People that actually do science understand the implications of a published paper.

> Papers don't establish ground truths in science > it doesn't shake the foundations of science

the comment doesn't say any of that, just that resources are wasted participating to these "conversations"

Papers are often rejected due to lack of novelty compared to previous work, especially if they contradict it. How many paper were similarly rejected because this fraudulent work was published?

It's definitely a source of wasted effort (I think almost every postgrad student has a story of trying and failing to get something described in a paper to work), but I would say that fraud is only a small fraction of failure to replicate cases. It's also quite common that there is some other factor making it difficult to get something working in your lab that worked somewhere else (which to be clear is still a problem, but a different one).

The cost of pursuing dead ends in science is NECESSARY and much greater than the cost of pursuing fraudulent results, albeit the latter being unnecessary. Probably 70%+ of time is wasted on paths that lead to nowhere, and that is part of the process. I would guess less than 1% of time is wasted because of manipulation. The process finds fraudsters (bad) well enough, and the process of being wrong (good) will always be central to science, even though it's so expensive.

Dont forget people who spend resources just replicating studies.

And doping. Now you think what? But. It is hard to compete with people doing drugs to be able crunch 16h seven days a week to get a result.

Reply on news.ycombinator.com