State of the science

[firecat]

This is one of the best critiques I’ve ever read of the current state of scientific research and communication. It makes me want to figure out how to help communicate science to non-scientists in ways that don’t distort the science or its strengths and limitations.

https://undark.org/2024/03/08/interview-paul-sutter-science-trust/

Comments

*

[minoanmiss]

Omg. saves

[silveradept]

Ah, that's a lovely piece, and I think all of the criticisms are spot-on. The hardest part about communicating science accurately is that the communication requires all the caveats, and humans don't like having to juggle four caveats with a conclusion. So somehow there has to be a way of explaining what the actual conclusion is, along with all the limitations that come with it.

And, I think, there has to be a greater willingness to accept and publish papers that say "we set out to see if this thing was correlated, it turns out it isn't." and to have that be applauded and counted as Real Science.

[firecat]

And, I think, there has to be a greater willingness to accept and publish papers that say "we set out to see if this thing was correlated, it turns out it isn't." and to have that be applauded and counted as Real Science.

Absolutely, which is another reason publish [a big-deal finding in a major journal] or perish is a poor way to do science.

[silveradept]

It really is. What we need is the Journal of the Null Hypothesis, which only accepts papers where whatever was hypothesized, it turned out not to be, but the science was good and so it should be an accepted and published paper. And that should be a way for someone to build their metrics, if they have to, because J. Null Hyp. is an excellent journal to publish in.

[firecat]

I love the idea of the Journal of Null Hypothesis! Why isn’t there one?!

[silveradept]

There are, apparently! The closest one in title is the Journal of Articles in Support of the Null Hypothesis, but other journals apparently do offer channels to publish null results in as well, or they have a null hypothesis collection for materials that did not reach sufficient statistical significance, but were otherwise excellent and rigorous studies that deserve publication. PLOS One has two, a collection called "The Missing Pieces" published in 2020 that had an older sibling from 2015 called "Positively Negative."

[firecat]

I’m glad to know that!