Previous studies have focused on either media coverage of polls or on their effects. This study investigated the visibility and quality of news reporting of opinion polls and the public evaluations of polls in the context of the Danish referendum on the introduction of the euro. A content analysis of the news coverage showed that more than one-third of all news in the campaign referred to polls and that newspapers were significantly better than television news in providing methodological information about the polls. Focus group interviews revealed that voters were dissatisfied with too many polls in the news and concerned about the possible influences of polls. Panel survey data substantiated this finding suggesting a broadly shared perception that the referendum campaign was dominated by polls. Voters who were frequently exposed to news about the referendum showed an increase in dissatisfaction with the dominance of polls during the campaign. Those who supported a more restrictive policy on the publication of polls in future campaigns included elderly voters, people dissatisfied with the campaign, and people who had cast their vote for the losing side. Results are discussed in the light of policy recommendations for the reporting of polls in campaigns. The centrality of public opinion polls in politics today is well established (e.g., Herbst , Lavrakas and Holley , Lavrakas and Traugott , Mancini ). Previous studies have either focused on the quantity and quality of news reports about opinion polls during national elections (e.g., Andersen , Brettschneider , Smith and Verall ) or the effects of opinion polls and exit polls during national and state level elections (e.g., Atkin , Ceci and The authors wish to thank the Danish Research Academy, the Nordic Film Foundation, the Danish Broadcasting Association and the Dutch Science Foundation (NWO) for providing funding for the study. The authors gratefully acknowledge the thorough and helpful comments provided by Peter Neijens, the Journal’s editor and three anonymous reviewers. A previous version of this article was presented to the WAPOR conference Democracy and the Millennium: The Role of Media and Public Opinion in Elections, Pamplona, Spain, November –, . The article was first submitted to IJPOR May , . The final version was received April , . World Association for Public Opinion Research Kain , Schmitt-Beck , Sudman ). Little attention, however, has been paid to the role of polls in referendum campaigns (LeDuc ). As a political and democratic instrument, referendums play an increasingly important role as a supplement to the representative democracy, and they have been of special importance in the process of European integration (Butler and Ranney ). A number of countries have held referendums with close races on the issues of joining the European Community/Union or on ratifying key EU treaties, and several member states have announced referendums to determine national policy positions on European issues in the future (Hug and Sciarini ). Denmark is one country that has utilized national referendums to determine policies of integration in Europe. The referendum outcomes have been both supportive and opposing of further European integration, often following heated public debates and close electoral races. Denmark remains one of the countries where there are currently no legal restrictions on the publication of polls at election time. There is a longstanding gentleman’s agreement, however, to end campaigning the day before an election and not to publish exit polls before voting has ended (Siune ). This corresponds with a number of countries where pollsters, networks, and others engaged in exit polling accept a ‘selfdenying ordinance not to release their figures until the last vote has been cast’ (Butler , p. ). In this study we investigate both the media coverage of polls and the effects on public perceptions of polls in a national referendum campaign: the Danish referendum on the introduction of the euro, the common European currency. Contrary to the advice of most political parties, financial elites, and major newspapers, a majority of Danes ( percent) voted ‘no’ to the euro. We also explore public support for policies regarding the publication of opinion polls. We begin with a review of the literature on the visibility and assessment of the quality of poll coverage in the news. We then turn to an investigation of the public perception of polls before discussing public support for policy recommendations on publications of polls. 1 While pre-election polls and exit polls are mentioned together, the distinction between the two is acknowledged and is also applied in the investigation of public support for policies concerning the publication of polls. Pre-election polls are typically based on (representative sample) surveys prior to election day and they typically measure turnout intention, vote preference, party leader preferences, policy issue support, etc. Exit polls are (representative) surveys that measure actual voting behavior of voters after they have cast their votes. The organization and methodology of exit polling is different from pre-election polls and are discussed elsewhere (Hofrichter ).