You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
This is a very nice (albeit bleak) general argument: if science requires "failed" work, but individuals whose work fails are punished, then that's a strong push toward cargo cult science. https://www.psychologytoday.com/us/blog/adaptive-behavior/201804/is-science-showing-diminishing-returns …pic.twitter.com/z2zdfLbOC8
The reason I like this argument is that it unifies and explains a _lot_ of bad behaviour. You can play methodological whack-a-mole with things like p-hacking, but if the underlying reasons remain, other problems will crop up in their place.
Need to shift to a model of resolving hypotheses + strength of evidence. Eg someone who kills a commonly held falsehood with a conclusive experiment should be revered.
1) The pressure and competition to produce positive results and pub can be crippling for scientific curiousity and cooperation.. completely agreed - system needs reconfiguration in incentives and feedback loop..
2) That said, one issue is that - based on an assumption that there are fewer pos results than neg- what's the threshold for pubbing neg results? Peer-review of course but how to prevent the flood of neg findings submissions from overloading the system/ overwhelming researchers?
For the key word is "research administrators". Research is intrinsically bottom-up, not top-down. As long as there are administrators and managers in power, there will be pressure on scientists to reach predefined goals.
The other keyword is "careers". Hierarchies in academia have a similar effect as management, just milder because the goals are fuzzier and less short-term.
This is insightful. What if doing science worked like DARPA program management and you knew you were going to leave in <5 years? Is there any real evidence of tenure producing real results in the past 50 years?
I am not aware of anyone having done any kind of study on how best to run research. Nor have I seen any convincing measure for "good" science.
what this excerpt didn't cover was when the success of a trial is promoted beyond its true accomplishment. take #alphago. with only the basis of beating a go master as a measure of success, it ignored the constraints of processing power, learning time, and learning methods.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
