This paper reports the results from a laboratory experiment designed to study strategic communication by an expert who is concerned about appearing to be well informed. A reporter privately observes a signal about a state of the world and reports a message to an evaluator. The informativeness of the signal is uncertain and unknown to both the reporter and the evaluator. The evaluator assesses the informativeness of the signal on the basis of the reporter’s message and the realization of the state of the world. The reporter benefits from being perceived as having accurate information, as her payoff increases in the evaluator’s assessment. With a series of carefully designed experimental treatments, we anatomically dissect reporters’ incentive to misreport. We manipulate the degree of common prior uncertainty on the state of the world and, using computerized evaluators, the beliefs of, as well as learning by, the evaluators. In line with the theoretical predictions, we find that reporters are less likely to misreport when there is more uncertainty about the state of the world and when evaluators conjecture that reporters truthfully report the signal. On the other hand, evaluators find it difficult to learn reporters’ strategies and overreact to messages that inaccurately predict the state of the world. In turn, this reduces reporters’ incentive to misreport and can account for reporters’ empirical strategies.