A sender commissions a study to persuade a receiver, but influences the report with some state-dependent probability. We show that increasing this probability can benefit the receiver and can lead to a discontinuous drop in the sender's payoffs. We also examine a public-persuasion setting, where...