Unpopular Science 240121

Links and ideas likely to make you unpopular

Eric Daza is a statistician who specializes in n-of-1, personal science experiments. I highly recommend all of his podcasts and videos, including an interview Eric Daza | N-of-1 Science & Causal Inference | Philosophy of Data Science on the YouTube channel Data&Science with Glen Wright Colopy.

Three takeaways:

On his epitaph : The Law of Total Expectation. The expected outcome of two different probability spaces is equal to the proportional size of the two

If you want to know the average outcome of something, you can first look at how the outcome depends on some other factor, and then average over all the possible values of that factor. This will give you the same answer as if you just looked at the outcome directly.

Both speakers highly recommend Harvard statistician Susan Murphy and her work on wearables.

His third and final point is about Eric’s interest in DEI, and how he feels it’s important to get more under-represented groups into the field.