,
7 tweets,
4 min read
Read on Twitter
Excited to share this new paper "Productivity, prominence, and the effects of academic environment" with @samfway @alliecmorgan and @DanLarremore, out today in @PNASnews pnas.org/content/early/… Here’s a little summary: 1/7 
Previously, we showed how faculty at higher prestige departments tend to be much more productive. But, since faculty in prestigious departments also tend to hire prestigious PhDs, it's unclear which of these—PhD or faculty prestige—drives this higher productivity 2/7
Obviously, both PhD and faculty prestige could be important, so we needed a way to (1) control for PhD location to measure the effect of faculty prestige, and (2) control for faculty location to measure the effect of PhD prestige. This led us to use a matched-pair design 3/7
Using data on 2500 faculty in Computer Science, we first matched faculty working at a similarly ranked university, but with PhDs of different prestige. Further matching on various possible confounders, we find that early-career productivity is independent of PhD prestige 4/7
Next, we matched faculty with PhDs from similarly ranked places, but with jobs at departments of different prestige. Again accounting for confounders, we find that the scientist working at the more prestigious institution is clearly more productive 5/7
How does prestige drive productivity? We tested three obvious explanations: preferential selection, retention, or high productivity expectations. None is well supported by our data. Instead, using extensive @CRAtweets data, we show how departments facilitate productivity 6/7
Takeaways: contrary to popular belief, we find that once you have a faculty job, doctoral prestige has little future bearing on early-career scientific productivity and impact. Instead, scholarship is driven by where you work, though exactly how remains an open question. 7/7
