Co-written with Noah van Dongen.
Our field is meta-research: the study of the research process itself. When we first got into this field, we both remember reading Munafo et al.’s “A manifesto for reproducible science” and Chambers’ “The seven deadly sins of psychology: a manifesto for reforming the culture of scientific practice” and feeling inspired. “Yes, goddamnit — these people get it!”
A few years later, while browsing Twitter, one of us saw the preprint, “A Manifesto for Team Science”. This time though, the feeling was different. “Really? Another manifesto?” A month later, a colleague sends a link to a paper, along with a winky-emoji.
Soon enough, self control lost out to curiosity. Click.
“Five ways to ensure that models serve society: a manifesto.” It was a comment published in Nature. 3 pages. 22 authors.
It turns out that it’s not just our field of meta-research that’s into manifestos. Do a quick google scholar search, and you’ll find more than 20,000 items with “manifesto” in their title and more than 1,500 items that have titles including “a manifesto for”.
There are manifestos for everything: growth econometrics, earth justice, ethnography, processual philosophy of biology, postmodern feminist legal proceedings, romance, cyberpunk, cyborgs, doomed youth, and people not being gadgets.
Given how prominent some manifestos become, you might start to think “hey, maybe I should write a manifesto too?”
Here, we present a set of graphical tools to help you answer this question. These tools produce robust inferences across a range of parameters and can be straightforwardly used by any scientist, at any career stage, in any field.
Posterior probability that you should write a manifesto
Should you write a manifesto? The approach developed here suggests a clear answer:
Some plots generated with code modified from McElreath’s Statistical Rethinking.