Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
A general mixture model of contests is introduced, combining stochastic performance and multivariate incomplete information. Performance is determined by a mixture distribution with endogenous weights on a good and bad distribution, respectively. Incomplete information does not increase the complexity of the problem, and actions are unambiguously strategic substitutes when technologies are homogeneous. These properties enable comparative statics with respect to changes in the type distribution. Applications include examining the role of the dependence structure between private valuations on the one hand and private budget constraints, starting advantages, and productivity parameters on the other. Extensions to asymmetric contests are considered.