Hacker News new | ask | show | jobs
by gerbal 4508 days ago
Good, any incredible result should be doubted until it can be replicated.
1 comments

What about credible results- should things require less evidence if they support the status quo?
The status quo by definition has a lot of support in scientific literature. Sure, you could right a botched paper supporting it, but people would basically ignore it because it would be a bad paper (though a peer reviewer would likely tear it to shreds anyway).
> The status quo by definition has a lot of support in scientific literature.

Not really. Science is by its very nature subversive, and as many Nobels have been awarded for work that falsified the status quo as confirmed it.

When the Michelson-Morley experiment falsified the luminiferous ether, much of physics fell into limbo (without a credible "status quo"), until Einstein's relativity theory offered an alternative explanation for electromagnetic propagation. But Einstein's proposal was met with incredulity until very good evidence confirmed it. That's science at its best -- perpetual doubt and skepticism.

I think we're using two different definitions of "status quo". Science most definitely has a status quo - that's the established theoretical basis for various fields.

That of course can be very different to the cultural status quo.

> Science most definitely has a status quo - that's the established theoretical basis for various fields.

Yes, that's true, but to call a falsifiable theory a status quo is to imply that it represents an anchor, a stabilizing influence, rather than a tempting moving target for skeptical scientists, its true role.

> What about credible results- should things require less evidence if they support the status quo?

Not in science. The problem with accepting credible results is that they may create an atmosphere of credibility where none really exists, and over time become the status quo without any real basis.

There are any number of stories in science where an expected value was reliably detected in a replication only because it was expected, and later, after the expected value changed, so did the replication results.

Remember that, in science, the null hypothesis is the default position -- in other words, there's no effect until persuasive evidence contradicts that assumption.