Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Appendix: inconsistency in the Naive and Nadiri models, 330.