How do ballot boxes affect elections?
Strategic voice plays a key role in this. In previous elections, more than a quarter of the Dutch voted strategically to a greater or lesser extent. For example, these voters find it important to try to increase the chance that a particular party will participate in government. Strategic voters often base their votes on opinion polls, because they roughly show how many seats parties will get.
What about the “bandwagon effect”?
This term, brought from the United States, refers to a brass band that ran through villages and towns. Audiences followed the music in ever-increasing numbers. In politics, the bandwagon effect refers to people’s tendency to ally themselves with winners or politicians who do well in opinion polls.
This is not just due to people blindly following the herd, says Tom van der Meer, professor of political science at the University of Amsterdam (UvA). “It’s also partly a cognitive process.” I look around and think: Looks like important things are happening at that party, maybe I should join them.
Scientists conduct research on the bandwagon effect by asking test subjects their opinions on a particular political topic in experiments. Part of the volunteers are first shown a poll about what others think about the topic, while the other part answers without seeing that poll first.
Are there other ways polls affect elections?
Yes, through indirect effects. Think about journalists who, based on polls, make different choices about which candidate to interview or what kind of questions to ask. The pitfall of some coincidental loss in percentages compared to the previous poll, followed by journalists asking questions like “Your campaign isn’t working, what’s going wrong?”, is notorious.
Polling agencies report margins of uncertainty, but this nuance is sometimes lost in reports. Van der Meer: “Or journalists, in their desire to explain, say, ‘The transitions don’t matter, but…’ Then a whole story follows that can color the picture.”
The professor also points to a self-reinforcing effect: opinion polls drive media reporting, which in turn influences the public, and then the polls change further, and so on. Van der Meer: “This will not continue indefinitely, of course. Parties have a potential number of voters that they can theoretically attract, but they cannot pull all voters in their direction.
How many seats do all these types of poll effects matter?
This can only be said with certainty if there are two of the Netherlands: one version with polls and one without polls. Since such an experiment is impossible, scientists are trying to get a signal through studies such as the one conducted by Sjoerd Stolwijk (UvA). He measured the voting preferences of more than a thousand Germans before and after the campaign, and analyzed the media use of all these people. Media reports about opinion polls were independently determined whether they were positive or negative in tone.
For example, Stolvik concluded that poll reporting could explain about 5% of the vote. But in other elections, this percentage might look very different. For example, polls in which a single candidate leads by a larger margin are quickly viewed as “boring” and receive less media attention than if a new poll threatens to create a close race.
All those polls right before the election, couldn’t they be a little lower?
In many countries, there is a ban on holding opinion polls in the final days before elections, for example because that would discourage voters from choosing the party in which their hearts truly lie. Van der Meer opposes this surveillance ban. “A poll is one source of information that people can base their votes on. Plus, there’s the election data, how leaders emerge, what they look like, you name it. I don’t know why you should ban those polls.”
In countries that prohibit conducting surveys, organizations and media are always finding creative tricks to conduct surveys. For example, by taking the survey from another country and presenting the results creatively. For example, in the midst of the ballot ban, Italian newspapers sometimes published horse race times with names suspiciously similar to those of political candidates.
In the Netherlands, there is a “honorary agreement” between research agencies not to publish new polls during the hours when polling stations are open; Otherwise there would be a strategic interest in voting later in the day so that you can still factor the latest poll into your considerations.
