The Anthropic Principle in Statistics and Science
The anthropic principle in physics states that our existence implies certain constraints on the natural conditions under which we evolved. In statistics, a corresponding anthropic principle can be used to infer properties of the models we should fit to data. For example, experiments are typically aimed to have a precision sufficient to estimate effects of interest but without overkill; it is rare to have an estimate that is 10 standard errors from zero. We demonstrate through several examples in social and medical sciences how the anthropic principle, combined with Bayesian inference, can be used to improve statistical practice.
Here are a couple of applications of the idea:
• [2000] Should we take measurements at an intermediate design point?
• [2022] A proposal for informative default priors scaled by the standard error of estimates (with Erik van Zwet)
In my talk I’ll discuss these and other examples. I think this anthropic principle is really important, arguably more important in statistics than in physics, which is the field where it originated.
Here’s the zoom information for the talk on Mon 29 June, 4:20pm London time:
https://imperial-ac-uk.zoom.us/j/97341955036?pwd=1kKNbPAwJthKtG55ynXMVF3TLSvIbl.1
Meeting ID: 973 4195 5036
Passcode: J3Ue$f
I’ll be speaking (remotely) at this conference celebrating the 60th birthday of physicist Andrew Jaffe. This seems to be the season for 60th birthday conferences.
I know AJ from when he was visiting the Flatiron Institute last year. We worked together on The Squealer: Sensification of model exploration and model misfit. There’s no connection between the Squealer and the anthropic principle; I decided to speak on the latter topic because I thought it would be of general interest to an audience of physicists.
When I asked my chatbot for examples of the Anthropic Principle applied to statistics, I got examples “in medicine” and “in social science,” suggesting I had circled back around to Andrew’s work. But the examples just seemed like observer and observation bias relabeled as the Anthropic Principle, so I am left wondering what the application of this principle adds in the field of statistics.
Matt:
To answer the question at the end of your comment, I suggest you read the two papers linked in the above post, as they show two different uses of the anthropic principle in applied statistics.
In machine learning, there’s the dogma that you need to balance/weight unbalanced datasets.
Peel back a layer and doing this just assigns a utility function to the confusion matrix such that identifying the rare case is more important than false positives.
Peel back one more layer, and it’s the anthropic principle (I think)—we wouldn’t be running a binary classifier on a 99:1 unbalanced dataset if we weren’t *really* interested in that 1%.