3 ms·

> This is an extreme example of that, and should stand as a cautionary tale that data should be believed especially when it is in conflict with theory. Rather t

> This is an extreme example of that, and should stand as a cautionary tale that data should be believed especially when it is in conflict with theory. Rather than being rejected and ignored. This is a terrible sentiment and it needs to stop spreading with these articles about Nobel rejections. Noticing that your data is in conflict with theory is the best way of realizing that you have made a mistake. Go back to your analysis and work through it until you realize what you did wrong. If you can't find the mistake, then share it with your colleagues/PI, and get them to tell you what you did wrong. If all of their ideas don't explain your result, then you may have found an interesting finding, but you should probably triple check your work and then try to reproduce the finding in new data.

btilly 7y ago
It is true that most impossible results are experimental error of one sort or another. But it is a systemic failure that biologists broadly rejected a result out of hand that had been successfully replicated many times, at multiple labs, over decades. And the attitude that you describe is part of why it happened. Data that does not fit in existing paradigms should be scrutinized with extra care, not rejected out of hand.
- lonelappde 7y ago
  All data should be scrutinized. If you are going to blindly accept data that matches prediction, than don't waste resources collecting the data you aren't going to use -- just use the theory and don't lie about verifyit it.
lonelappde 7y ago
The charge of electron was incorrectly reported in many studies for many years because scientists refused to publish results that contradicted Millikan's incorrect result, and happily published incorrect results because they matched the accepted theori of the time. This is one of Feynman's favorite stories to illustrate Cargo Cult science. http://calteches.library.caltech.edu/51/2/CargoCult.htm http://calteches.library.caltech.edu/51/2/CargoCult.htm

btilly 7y ago

It is true that most impossible results are experimental error of one sort or another. But it is a systemic failure that biologists broadly rejected a result out of hand that had been successfully replicated many times, at multiple labs, over decades. And the attitude that you describe is part of why it happened. Data that does not fit in existing paradigms should be scrutinized with extra care, not rejected out of hand.

lonelappde 7y ago
All data should be scrutinized. If you are going to blindly accept data that matches prediction, than don't waste resources collecting the data you aren't going to use -- just use the theory and don't lie about verifyit it.

lonelappde 7y ago

The charge of electron was incorrectly reported in many studies for many years because scientists refused to publish results that contradicted Millikan's incorrect result, and happily published incorrect results because they matched the accepted theori of the time. This is one of Feynman's favorite stories to illustrate Cargo Cult science. http://calteches.library.caltech.edu/51/2/CargoCult.htm http://calteches.library.caltech.edu/51/2/CargoCult.htm