To learn about the human world, we should accept uncertainty and embrace variation. We illustrate this concept with various examples from our recent research (the above examples are with Yair Ghitza and Aki Vehtari) and discuss more generally how statistical methods can help or hinder the scientific process.
I like something of the form “calculate uncertainty & estimate variation” better than “accept uncertainty and embrace variation”
Not all uncertainty deserves embracing. Uncertainty due to crappy measurement or a bad model or a weak statistical procedure is different from uncertainty inherent to the physical phenomenon itself.
To the extent uncertainty reveals our ignorance about the world we want to reduce it not embrace it. Fuzzy estimates are not always better than point estimates.
Rahul:
I said “embrace variation” not “embrace uncertainty.”
True. But neither does all variation deserve embracing, IMO.
Agreed. As Niebuhr might have said, let us have the serenity to embrace the variation that we cannot reduce, the courage to reduce the variation we cannot embrace, and the wisdom to distinguish one from the other.
In the case of some of the psychology experiments we have been discussing, it indeed is possible to reduce variation by using within-subject designs with careful measurements.
+1
When I saw this talk announcement, my first thought was, well: if it’s not further than across the street at Club Monaco, then I am not ruling out attending. And it turns out not to be further than across the st. at club Monaco. But Gelman might prefer that I contrapose the first conditional so that it’s in the form of a “falsification”.