Most perspectives on public participation share the notion that dialogues should be open, allowing participants to articulate and evaluate different views and knowledge claims. We hypothesize that participants' evaluation of claims may be biased because participants have a preference for a particular type or source of a claim. This would hamper an open dialogue. We tested the effect of three variables on scientists' evaluation of claims of the general public about GM food: the claim's favorability towards GM food, the phrasing, and the source of the claim. Results are based on a survey-experiment among 73 biotechnology-scientists. Biased processing occurred when scientists evaluated claims. Claims that were corresponding with the attitude of the scientists and that were phrased in a cognitive way were evaluated more positively than claims that were contrasting the attitude of the scientists and that were phrased in an affective way. Contrary to our expectation, scientists evaluated claims of the public more positively than claims of experts. © 2009 SAGE Publications.