Are Pre-Election Polls More Helpful than Harmful? Evidence from the Canadian Case

Abstract

The growing importance of polls in news coverage raises questions about whether legislators should regulate polls in election campaigns. Although restrictions are on the rise, little is known about some basic but important facts regarding polls. Who reads polls? Are citizens who are aware of polls more likely to change their voting behaviour? And do polls help them to better anticipate electoral outcomes? We answer these questions using data from the 2015 Canadian election. We show that being exposed to polls is not associated with voter’s likelihood of changing their vote choice compared with their vote intention during the campaign, does not affect the propensity to turn out or abstain, but does improve their ability to forecast the winner. We conclude that polls are more helpful than harmful.