Hacker News new | ask | show | jobs
by traverseda 2512 days ago
I can do both of those things separately.

https://en.wikipedia.org/wiki/Replication_crisis offers the best evidence I can offer for bad science being done.

https://en.wikipedia.org/wiki/Political_views_of_American_ac...

Goes over the issues related to politics, conservatives aren't as under-represented when you include STEM fields, but pay special attention to the disciplinary variance.

>Focusing specifically on social psychology academics, a 2014 study found that "[b]y 2006, however, the ratio of Democrats to Republicans had climbed to more than 11:1.

Also note what fields are facing the most obvious replication crisis.

I expect you'll want to do your own research, but that should get you started.
1 comments
euroPoor 2512 days ago
I was not talking about bad science in general. I could say that in general, science seems to be in crisis [1] and pinning it down to people whose political affiliation does not suit your needs is reductive. Also, because there’s some different distributions does not mean leftwing = bad.

Another interpretation could be: the GOP platform is highly anti-science and most scientists wouldn’t want to be associated to such a horrendous organization which doubts scientific consensus regarding an issue that’s threatening the whole of civilization.

[1] https://blogs.scientificamerican.com/cross-check/is-science-...

link
traverseda 2512 days ago
Right, like I say I can show both of those things separately, but I can't demonstrate that they're correlated.

How would you go about finding that evidence? What would be sufficient evidence that they're correlated for you?

link
euroPoor 2512 days ago
You put it in your question, right? I’d like to see a study that EXPLICITLY correlates political views and activism with measures of ‘bad science’ like no citations, impossible reproducibility etc
link
traverseda 2512 days ago
I found a preprint of a study examining exactly that. I haven't had a chance to really read through it in depth though, so I can't speak to whether or not it's actually a good paper.

https://psyarxiv.com/6k3j5/

My naive first skim seems to demonstrate that political slant does have an effect on reproducibility, although it doesn't matter if that ideological slant is conservative or liberal.

It's not much, but it was tricky to find and it is what you're looking for. More research is needed.

link