People use analogies for many purposes such as building mental models, making inspired guesses, and extracting relational structure. Here we examine whether and how analogies have more direct influence on knowledge: Do people treat analogies as probabilistically true explanations for uncertain propositions? We report an experiment that explores how a suggested analogy influences people’s confidence in inferences. Participants made predictions while evaluating a suggested analogy and observed evidence. In two conditions, the evidence is either consistent or in conflict with propositions based on the suggested analogy. We analyze the responses statistically and in a psychologically plausible Bayesian network model. We find that analogies are used for more than just generating candidate inferences. They act as probabilistic truths that affect the integration of evidence and confidence in the target and source domains. People readily treat analogies not as a one-way projection from source to target, but as a mutually informative connection.