Disclosure of research results: the cost of proving your honesty.

In situations where a biased sender provides verifiable information to a receiver, I study how strategic reporting affects the incentives to search for information. Research provides series of signals that can be used selectively in reporting. I show that the sender is strictly worse off when his research effort is not observed by the receiver: he has to conduct more research than in the observable case and in equilibrium, discloses all the information he obtained. However this extra research can be socially beneficial and mandatory disclosure of results can thus be welfare reducing. Finally I identify cases where the sender withholds evidence and for which mandatory disclosure rules become more attractive.