Similar to all other types of information, public opinion polls can influence public opinion. We present two hypotheses to understand how polls affect public opinion: the bandwagon and the underdog effect. The bandwagon effect claims that voters "jump on the bandwagon," which means that if a party is gaining in the polls, the party will gain additional support from the voters, and vice versa if the party is losing in the polls. The underdog effect suggests that if a party is losing in the polls, the party will gain some sympathy votes to offset this loss. We use a survey experiment to test the two hypotheses. We find evidence of the bandwagon effect, and the effect is strongest in the positive direction. When voters learn that a party is gaining in the polls, voters will be more likely to vote for it. There is also some evidence for the negative bandwagon effect. We find no evidence for the underdog effect. The effects head in the same direction regardless of the size of the party. We conclude by discussing the implications of our findings with regards to a potential ban on publishing opinion polls.