Despite the importance of doing so, people do not always correctly estimate the distribution of opinions within their group. One important mechanism underlying such misjudgments is people’s tendency to infer that a familiar opinion is a prevalent one, even when its familiarity derives solely from the repeated expression of one group member. Six experiments demonstrate this effect and show that it holds even when perceivers are consciously aware that the opinions come from one speaker. The results also indicate that the effect is due to opinion accessibility rather than a conscious inference about the meaning of opinion repetition in a group. Implications for social consensus estimation and social influence are discussed.