I think identifying scientific fraud, or identifying experts on the TED talk circuit who are doing substantially dishonest or misleading work, is valuable
Is this specific to dishonesty/fraud? I’ve done some replications showing that findings are not reliable because of data errors or lack of robustness (here, here, here).
I think checking whether results replicate is also important and valuable work which is undervalued/underrewarded, and I’m glad you do it.
One dynamic that seems unique to fraud investigations specifically is that while most scientists have some research that has data errors or isn’t robust, most aren’t outright fabricating. Clear evidence of fake data more or less indicts all that scientists’s other research (at least to my mind) and is a massive change to how much they’ll tend to be respected and taken seriously. It can also get papers redacted, while (infuriatingly) papers are rarely redacted for errors or lack of robustness.
But in general I think of fraud as similar in some important ways to other bad research, like the lack of incentives for anyone to investigate it or call it out and the frequency with which ‘everyone knows’ that research is shady or doesn’t hold up and yet no one wants to be the one to actually point it out.
Is this specific to dishonesty/fraud? I’ve done some replications showing that findings are not reliable because of data errors or lack of robustness (here, here, here).
I think checking whether results replicate is also important and valuable work which is undervalued/underrewarded, and I’m glad you do it.
One dynamic that seems unique to fraud investigations specifically is that while most scientists have some research that has data errors or isn’t robust, most aren’t outright fabricating. Clear evidence of fake data more or less indicts all that scientists’s other research (at least to my mind) and is a massive change to how much they’ll tend to be respected and taken seriously. It can also get papers redacted, while (infuriatingly) papers are rarely redacted for errors or lack of robustness.
But in general I think of fraud as similar in some important ways to other bad research, like the lack of incentives for anyone to investigate it or call it out and the frequency with which ‘everyone knows’ that research is shady or doesn’t hold up and yet no one wants to be the one to actually point it out.