Twitter consists of micro-communities, which should be factored into any attempts at statistical analysis. Some of those communities have migrated to Mastodon, where the presence of multiple servers provides additional signal for micro-community and cluster identification. In addition, Twitter(s) circa <year> are different datasets due to both platform policy and evolving membership.

My wife is a plant pathologist with a strong interest in mycology. She uses twitter specifically for fungus-focused posts. Whether she actually finds it useful for research I don't know, but it seems to help with networking and finding out about events.

Compared with many things we believe we know, p<0.05 is still pretty weak.

For instance, I've seen quite a few medical correlations that were P<0.001.

I try now to avoid the mindset of "significant" vs "insignificant" as having a discrete black or white cutoff.

As people move to other platforms it feels like the signal could change and people may have more control over their feed but less serendipity?

If it's true Twitter doesn't matter that would mean academia is truly broken.

It's certainly broken, but I doubt to that extent.

There's a conversation to be had about the costs (time and the risk of being cancelled) of running a Twitter account as a researcher and if you don't, how do you get your work on Twitter?

Anyway checkout Altmetric if the overlap between social media and academia interests you -

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