People love to talk about the pitfalls of statistics and how we can “lie” with it. Those people don’t do statistics. In science, statistics is often quite difficult. It’s there to help us with very complicated problems, not to hurt others with simple craven power leveragings.
The Dawdlers discuss the positive side of this dichotomy through a paper by the mid-to-late 20th Century statistician George E.P. Box. Here, he forays into the big vistas of Science and Statistics with his mantra in hand–“All models are wrong, but some are useful.” Through discussions of an exemplar, the infamous early 20th Century “scientist” (who was also his father-in-law), R.A. Fisher, Box unpacks the statistical insights of his century’s deepest dives into epistemology.
Bring an extra pair of socks, muck rakers, things are about to get saturated.
00:02:26 – Science and Statistics by G.E.P. Box / Box’s bumper sticker / Ryan has no plan / More “what is science?” shit
00:19:50 – What is statistics? / Qualitative data
00:33:56 – Who is R.A. Fisher? / Experimental design / Basic concepts in statistics
00:58:32 – Using English to express math / Mathematistry and cookbookery / Trying to control complicated systems
01:10:32 – Does one need advanced math to address complicated problems / Ryan wants to know how to speak math equations well in English / Thinking in words vs thinking in pictures
01:26:10 – Harland has two things to say still / Harland is paranoid Box is talking about him / Mathematical modeling
