A key role of science-based non-governmental organizations (NGOs) is to communicate scientific knowledge to policymakers. However, recent evidence has emerged showing that industry-backed groups often attempt to undermine the credibility of such NGOs and weaken their ability to influence policy. To investigate the mechanisms by which a firm can profitably create doubt about scientific information, we use a signaling model of interest-group lobbying in which the policymaker has fixed costs of taking action. We explore two mechanisms for the creation of doubt. The first involves using Bayesian persuasion to imply that the NGO may be a radical extremist whose lobbying is not credible. The second involves the creation of a think tank which can offer its own testimony on scientific matters. We show the firm prefers that the think tank does not act as a credible moderate, but instead sometimes takes radical, non-credible, positions. We identify conditions under which each mechanism is preferred by the firm.