Scott Alexander reviews research on teacher effectiveness, examining Value-Added Modeling (VAM) and its criticisms, while expressing uncertainty about claimed long-term impacts on adult earnings.
Longer summary
Scott Alexander reviews research on teacher effectiveness, focusing on Value-Added Modeling (VAM) and its criticisms. He explores studies showing teachers account for about 10% of variance in student test scores, with effects decaying quickly over time. However, some controversial studies claim long-term impacts on adult earnings. Scott examines methodological issues, potential biases, and alternative explanations, ultimately expressing uncertainty about the validity of these long-term effects. He suggests non-cognitive factors like behavior might explain persistent impacts if they exist, but remains skeptical given the contrast with research showing minimal parental influence on outcomes.
Shorter summary