Policy proposals often contain complex legal, technical, or scientific jargon, making it difficult for people to evaluate their favorability toward the policy. We proposed one experiment testing the effect of language complexity on people’s evaluation of a policy proposal as moderated by their initial policy beliefs. We hypothesized that when a policy was consistent with one’s beliefs or if participants had no policy preference, they would evaluate it more favorably when it was simple than when it was complex; when a policy was inconsistent with one’s beliefs, they would evaluate it less unfavorably when it was complex than when it was simple. Results confirmed our hypotheses. This demonstrates that complex information does not always make people judge policies more negatively but rather cause people to weigh complex information less heavily in their judgments.