Article dans une revue: The literature on special interest groups emphasizes two main influence channels: campaign contributions and informational lobbying. We introduce a third channel: providing information about the credibility of political rivals. In particular, nongovernmental organizations (NGOs) often aim to communicate scientific knowledge to policymakers, but industry‐backed groups often attempt to undermine their credibility. We extend a standard signaling model of interest‐group lobbying to include fixed costs of policymaker action and show that these costs make possible two mechanisms for creating doubt about the value of policy action. The first uses Bayesian persuasion to suggest the NGO may be a noncredible radical. The second involves creating an opposition think tank (TT) that acts as a possible radical, not a credible moderate. We show that the TT cannot always implement the Bayesian persuasion benchmark, and we characterize how optimal TT design varies with exogenous parameters.