Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
I. Introduction, 366. — II. The optimal rate of learning, 368. — III. Modifications and extensions, 376. — IV. Discussion and application of the model, 377.