Abstract: Just as individuals believe things, so do groups. We routinely ascribe beliefs to groups such as corporations and governments claiming, say, that the government knows that there is a cost of living crisis or that a corporation knows that inflation is rising. We use such attributions to both predict the behaviour of groups ahead of time, as well as to evaluate them after the fact. Of particular interest are cases in which groups persist in dogmatically believing against the evidence. Thus, a government might dogmatically believe against the evidence that its climate change provisions are adequate, and a detective team might dogmatically believe against the evidence that a local gangster is responsible for a recent murder. My project here is to think about what it is for a group to dogmatically believe against the evidence, which in turn raises questions about the nature of group evidence. I will tentatively argue that considering group dogmatic belief provides reasons to reject “summative” accounts of group evidence on which a group’s evidence is a function of the evidence of its members in favour of “non-summative” accounts of group evidence which allow a group’s evidence to float freely from the evidence of its members.