Read on Twitter
I find it interesting that some in animal behavior/behavioral ecology are discussing the field's "replication crisis". This is interesting to me because we don't actually know that such a crisis exists. Instead, what we have are alleged instances of misconduct. 1/n
This has led to discussions of best practices, HARKing, reproducibility, open science practices, etc. This is GREAT and we should support these discussions and work to improve our own approaches accordingly. However, none of this gets at the potential for an actual crisis. 2/n
A replication crisis, as the name says, is the broad failure to replicate major findings of a field. This simply has not been examined for AB/BE. While meta-analyses suggest publication bias, a consistent failure to replicate major findings hasn't been demonstrated (yet). 3/n
That said, I've no doubt that if there ever is a reckoning by the field, we'll see the same sort of crisis but **That Work Hasn't Been Done.**
One instance overturning a major textbook example I'm aware of is: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0037785
but there's not a lot more. 4/n
One instance overturning a major textbook example I'm aware of is: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0037785
but there's not a lot more. 4/n
We don't have a good idea of how solid our field's foundation is. Much of what we assume in AB/BE is likely based on research we'd now consider HARKing, p-hacking, etc. Even in the late 90's and early 2000's those practices were what was taught. Probably still are. 5/n
This has consequences as many of our premises may be flawed or even just incomplete. Re-examination of Bateman's principle by P. Gowaty & reviewed by Z. Tang-Martínez & others gets at this but it's likely a lot of other foundational work is similarly shaky. 6/n
Even if a real crisis is demonstrated will people care? Honestly, based on responses to meta-analyses, I'm skeptical. It's very easy to decide that replication failures are unique one-offs, even if the one-offs are more common than confirmations. How can we address that? 7/n
There's also a MAJOR problem in that much of what's now assumed is based on work in only a tiny sliver of the animal diversity that's out there. @WidowWeb 's group had a fantastic paper in 2017 on this topic. 8/n https://www.sciencedirect.com/science/article/abs/pii/S0003347217300647
So, as a field, do we need replication consortia like those in other fields? If so, how do we fund them? Given the focus on "transformative" work how do we convince @NSF that this is not "just" iterative? How do we ensure our findings are taxonomically generalizable? End rant/9
