Cross-national patterns in early voting and time-of-voting-decision

By Simon Labbé St-Vincent, University of Montreal

What is the story?

A perfectly rational voter waits until the very last moment of the campaign to make a decision regarding his or her vote. In doing so, he or she has the possibility to gather a maximum of information about all the parties and all the candidates and increase his or her probability to pick the good one. However, in reality, many voters make a decision early in the campaign, sometimes they even cast their vote before Election Day. In this entry, I use the MEDW survey data to highlight the variations between country regarding time-of-voting-decision and early voting. I then see whether these behaviours are associated to the level of attention given to the electoral campaign.1

Early voting

In all but two countries surveyed (i.e., France and Belgium), voters could decide to cast a vote before Election Day.2 Among all respondents, 25% had already voted at time we circulated our pre-election MEDW survey (typically 5 to 10 days before the election). However, this figure varies quite a lot between countries. As shown in Table 1, about 50% the electorate (and an overwhelming majority of voters) cast an early vote in Switzerland, about 20% in Canada and Germany, and only 6% in Spain.

Country % of early votes N
Belgium . .
Canada 17% 2,294
France . .
Germany 24% 8,832
Spain 6% 4,848
Switzerland 50% 4,819
Total 25% 20,783
Table 1: Early voting by country

This variation is probably due to the electoral law concerning early voting. For example, in Spain, the rule changed just before the elections we surveyed. The legislator made it more difficult for voters to be eligible for early postal voting.


In the post-election MEDW survey (conducted within a week after Election Day), we asked voters when during the campaign they decided to vote the way they did. They could choose between a few months, a few weeks, a few days before Election Day, or on Election Day. Figure 1 reports the proportion of respondents having picked each of these categories by country. Overall, 61% of respondents made their decision a few months before the election.

We also observe some variations between countries. About 75% of respondents in Belgium said they decided how to vote months before Election Day, about 70% in France, 65% in Canada, 60% in Germany and Spain, and only 30% in Switzerland. The low level of ‘very’ early deciders in Switzerland is compensated by a larger number of respondents reporting having decided weeks before Election Day (45%), which is more than twice as much as the overall average in the other 5 countries (20%). There are fewer variations in the proportion of Election Day’s deciders between countries: from a minimum of 3% (Switzerland) to a maximum of 10% (Spain).

Figure 1: Time-of-voting-decision by country

Campaign attention, early voting and time-of-voting-decision

From these figures, an interesting question is ‘were those early voters and early deciders less attentive to the electoral campaign?’ In the MEDW post-election survey, we asked people to report, on a 0 (no attention at all) to 10 (a lot of attention) scale, their level of attention to the electoral campaign on five different medias: television (mean=5.6, n=22k), newspapers (mean=4.6, n=22k), radio (mean=3.9, n=22k), internet (mean=4.0, n=22k), and Twitter (mean=1.2, n=12k). From these data, I create an index, ranging from 0 to 10, capturing the maximum (self-reported) attention given to any of these four medias (mean 6.8, n=22k).

I compare the maximum level of attention of early and late deciders/voters. Voters who have decided how to vote ‘months before’ the election score higher on the attention index (7.4 compared to 6.7 for the rest of the electorate, n= 17k). The same observation can be made for early voters: their attention is on average of 7.3 (compared to 6.6 for other voters, n=16k, excluding France and Belgium). Results show that early deciders/voters are, contrary to our expectation, more attentive to the electoral campaign. One explanation is that voters do not pay attention to the electoral campaign to become more informed about the various parties and candidates. They might simply keep themselves informed, as they are genuinely interested in politics in general. Obviously, further analyses need to be performed to really understand the cross-national patterns of early voting and time-of-voting-decisions.

  1. The survey samples used to calculate the frequencies of this blog entry are those listed in this other blog post

  2. Belgians living abroad could cast an early vote if they asked their embassy to participate in the election through ‘postal voting’. However, this is only one way among others for Belgian expatriates to participate in a national election. 

This entry was posted in Blog. Bookmark the permalink. Both comments and trackbacks are currently closed.