Voters often have difficulty making choices on the myriad state constitutional amendments they vote on each year. Without partisan cues, they turn to other sources for these low-salience, high complexity measures. One such source is newspaper endorsements. In this article, we look at newspaper endorsements of ballot measures in Florida over 20 years both on “no” votes and roll-off. We argue that endorsements' effect on “no” votes and roll-off differs in ways not previously appreciated. Newspaper endorsements have a positive impact on no votes, as expected from the information theory of voter participation. Endorsements have little impact on roll-off, which we posit is because roll-off voters are not likely to seek information from newspapers. Thus, newspaper endorsements serve to persuade, but not entice, voters to vote for ballot measures.