Teachers: Much More Than You Wanted To Know

An exemplary MMTYWTK on teacher effectiveness, opening from a genuine puzzle: how can a single good 4th-grade teacher reportedly add ~$80K to lifetime earnings when behavioral genetics says parents have minimal non-genetic effect on earnings? Scott works through the literature with characteristic rigor and honesty. (1) Magnitude: teachers explain ~5-20% (most studies ~10%) of same-year test-score variance; student/individual factors dominate; the most robust teacher predictor is verbal/cognitive ability (not certifications or degrees). (2) VAM (value-added modeling): a +1 SD teacher raises same-year scores ~0.1 SD, but 50-75% decays within two years; the Sanders-Rivers 'three great teachers in a row' super-additivity claim is contested. (3) The VAM-bias drama (VAMboozled!, Ravitch's 'junk science'): Rothstein's devastating retro-causal test (5th-grade teachers 'predict' their students' 4th-grade scores -> bias), Briggs & Domingue finding that effect as large as the real one (suggesting ~100% bias), and Plomin's twin study finding VAM itself 40-50% heritable (Spotted Toad's elegant explanation via IQ becoming more genetic with age). (4) The Chetty $39-80K earnings claim and the central paradox: test-score effects fade fast, yet earnings effects persist 20+ years ('fade-out and re-emergence'); the transfer-teacher quasi-experiment; the inconclusive Rothstein-Chetty exchange. (5) Project STAR -- the randomized experiment where kindergarten class quality predicts age-25 earnings, which even skeptic Caplan calls 'one of the most impressive empirical papers ever written' and which resists the usual anti-VAM critiques. (6) Resolution via NON-COGNITIVE skills / behavior (Chetty's 8th-grade-behavior data; the disruptive-peer studies; Scott's own 'behavioral-problem' and 'early-school-as-neurological-insult' theories). Lands honestly at 50-50, 'mostly based on low priors rather than weakness of the studies.' Comprehensive, fair, and a model of grappling with a confusing literature.

Quality 78 · Excellent. High-Excellent (78): a standout MMTYWTK -- comprehensive, rigorous, scrupulously fair, and intellectually honest, genuinely illuminating a confusing literature (the test-score-decay-vs-earnings-persistence paradox) by surfacing and weighing the key studies, debates, and a strong genetic-confound analysis. Co-tier with the strong reviews (Bauhaus/Origins-of-Woke 78), a notch below Missing-Heritability (81) because it conveys and synthesizes others' research rather than producing an original investigative breakthrough, on a somewhat narrower topic, and lands explicitly unresolved (50-50).

Claude’s paradigm shift 46 · Moderate. Moderate (46): the value is the comprehensive, fair synthesis and the crisp framing of the paradox and its genetic-confound resolution, rather than a novel thesis -- the findings, debates, and even the non-cognitive-skills explanation are drawn from the literature.

Real-world impact 2 · Minor. A standout MMTYWTK that illuminates a confusing teacher-effectiveness literature (the test-score-decay-vs-earnings-persistence paradox; value-added and the genetic confound) on a materially-consequential education-policy topic. Real relevance, but a review informing understanding with no change to practice — low RWI.