“…If teachers improve skills not captured by test scores, then excellent teachers who improve long-run outcomes may not raise test scores, and the ability to raise test scores may not be the best predictor of effects on long-run outcomes. Indeed, Chetty, Friedman, and Rockoff (2011) note that teachers may have important effects on longer-run outcomes that are not reflected in their test score value-added and, using their same data, Chamberlain (2013) finds that test score effects may account for less than one quarter of the overall effect of teachers on college entry. This paper speaks to this second critique by being the first to investigate (a) whether teachers affect skills not captured by test scores, and (b) whether, and to what extent, teacher effects on a proxy for non-cognitive skills predict effects on long-run outcomes (that are missed by effects on test scores).…”