4 ms·

One lab is definitely not adequate. Replication attempts don't give you an unambiguous signal, many things can go wrong. If one lab hasn't succeeded in replic

One lab is definitely not adequate. Replication attempts don't give you an unambiguous signal, many things can go wrong. If one lab hasn't succeeded in replicating the paper, does that mean the paper is wrong, or just that a necessary step wasn't documented clearly or followed correctly? More labs trying to replicate give you more independent signals.

faraggi 3y ago
But if one lab *does* succeed, it means there probably is something of substance.
- scarmig 3y ago
  Cold fusion also replicated across multiple labs, before they all retracted shortly after.
  - GolfPopper 3y ago
    My memory is fuzzy, but my recollection is that Pons & Fleischmann's non-discovery involved unclear "waste heat" measurements, and a setup that wasn't well described, making it difficult for other researchers to duplicate. That would seem to contrast with Sukbae Lee, Ji-Hoon Kim, and Young-Wan Kwon's work, where everything suggests they've provided a readily replicable description of the material and process, and the evidence (magnetic levitation) is going to be pretty obvious.
paulddraper 3y ago
Depends on the finding. Getting a false negative from a lab is plausible. Getting a false positive seems very unlikely.
- btilly 3y ago
  Cold fusion got several false positives very quickly. And if this is a diamagnetic that seems to do some of the right behavior, a false positive is quite possible here as well. Hopefully all corrected in time.
  - paulddraper 3y ago
    Although the experimental protocol had not been published, physicists in several countries attempted, and failed, to replicate the excess heat phenomenon. The first paper submitted to Nature reproducing excess heat, although it passed peer review, was rejected because most similar experiments were negative and there were no theories that could explain a positive result;[notes 2][42] this paper was later accepted for publication by the journal Fusion Technology. Nathan Lewis, professor of chemistry at the California Institute of Technology, led one of the most ambitious validation efforts, trying many variations on the experiment without success,[43] while CERN physicist Douglas R. O. Morrison said that "essentially all" attempts in Western Europe had failed. https://en.m.wikipedia.org/wiki/Cold_fusion https://en.m.wikipedia.org/wiki/Cold_fusion
hatsunearu 3y ago
when an "adversarial" lab (as in, competing with the original lab) tries to replicate and succeeds, you can be pretty damn sure it's legit.
jncfhnb 3y ago
If one lab fails they may have fucked up. If one lab succeeds then it’s very noteworthy. This is a fairly straight forward claim. It’s not like “most science” with dubiously small effect sizes that may not replicate.

faraggi 3y ago

But if one lab *does* succeed, it means there probably is something of substance.

scarmig 3y ago
Cold fusion also replicated across multiple labs, before they all retracted shortly after.
- GolfPopper 3y ago
  My memory is fuzzy, but my recollection is that Pons & Fleischmann's non-discovery involved unclear "waste heat" measurements, and a setup that wasn't well described, making it difficult for other researchers to duplicate. That would seem to contrast with Sukbae Lee, Ji-Hoon Kim, and Young-Wan Kwon's work, where everything suggests they've provided a readily replicable description of the material and process, and the evidence (magnetic levitation) is going to be pretty obvious.

paulddraper 3y ago

Depends on the finding. Getting a false negative from a lab is plausible. Getting a false positive seems very unlikely.

btilly 3y ago
Cold fusion got several false positives very quickly. And if this is a diamagnetic that seems to do some of the right behavior, a false positive is quite possible here as well. Hopefully all corrected in time.
- paulddraper 3y ago
  Although the experimental protocol had not been published, physicists in several countries attempted, and failed, to replicate the excess heat phenomenon. The first paper submitted to Nature reproducing excess heat, although it passed peer review, was rejected because most similar experiments were negative and there were no theories that could explain a positive result;[notes 2][42] this paper was later accepted for publication by the journal Fusion Technology. Nathan Lewis, professor of chemistry at the California Institute of Technology, led one of the most ambitious validation efforts, trying many variations on the experiment without success,[43] while CERN physicist Douglas R. O. Morrison said that "essentially all" attempts in Western Europe had failed. https://en.m.wikipedia.org/wiki/Cold_fusion https://en.m.wikipedia.org/wiki/Cold_fusion

hatsunearu 3y ago

when an "adversarial" lab (as in, competing with the original lab) tries to replicate and succeeds, you can be pretty damn sure it's legit.

jncfhnb 3y ago

If one lab fails they may have fucked up. If one lab succeeds then it’s very noteworthy. This is a fairly straight forward claim. It’s not like “most science” with dubiously small effect sizes that may not replicate.