Bob Luddy and the Thales Academy

Scott Alexander reports:

Bloom’s Two Sigma Problem: children given private tutoring will do two sigmas better than average (ie the average tutored student will be in the 98th percentile of nontutored students). But see here for some argument that the real value is lower, maybe more like 0.4 sigma. Some further discussion on the subreddit asks the right question – can we simulate this with some kind of clever computer-guided learning? – and gives the right answer – apparently no. TracingWoodgrains has a great comment. Especially interested in their discussion of Direct Instruction: “One of the few schools to use it as the basis of their program for math and English, a libertarian private school in North Carolina called Thales Academy, is reporting results exactly in line with the two-sigma bar: 98-99th percentile average accomplishment on the IOWA test. Their admissions process requires an interview at the elementary level, but no sorting other than that, so it’s not a case of only selecting the highest-level students.” (though note that IOWA is nationally normed, and Thales is in the well-off Research Triangle area). On the other hand, it costs half of what public schools do, so file this under “cost disease” too.

By the way, I have been enjoying my read of Robert L. Luddy’s Entrepreneurial Life: The Path from Startup to Market Leader.  Luddy is founder of Thales Academy, and the final chapter of his memoir covers his thoughts and praxis on education.