Are you swayed by TV and internet voting campaigns? Political parties all over the world certainly think you are, and spend millions on advertising their candidates. Now an analysis of election results over 30 years in different countries shows that, for each political party, voting follows the same pattern, regardless of nationality, culture, history or economics.
The most important factor determining a candidate’s success compared with rivals in the same party turns out to be his or her personal ability to connect with the public. In other words, the key factor could be how many friends you’ve got on Facebook.
“When it comes to voting,” says Santo Fortunato of the Institute for Scientific Interchange in Turin, Italy, “people act in the same way regardless of national identity and the economic or political context. Even modern campaign tools like television and the internet have no great effect.”
Advertisement
The influence of the candidate’s political party and the prevailing economic conditions usually confound attempts to uncover a voting pattern. Fortunato’s approach has revealed that there is indeed a pattern, and that social networking can explain it.
“When it comes to voting, people act in the same way regardless of national identity and the economic or political context”
Over the past decade, several independent teams of researchers studying the mathematics of voting behaviour noted an intriguing pattern in election results in countries where elections had many candidates. They found that most candidates received a small number of votes while a few did much better, winning a large fraction of votes.
Looking in detail at elections in Brazil and India, the actual numbers seemed to vary in a regular way: twice as many candidates received 20 per cent as did 40 per cent, twice as many again received 10 per cent as did 20 per cent, and so on. Mathematically, in other words, the number of candidates receiving x per cent of the votes seemed to be simply inversely proportional to x.
However, this didn’t seem to be the case in other countries with different the election systems, says Fortunato. Working with physicist Claudio Castellano of the University of Rome, Fortunato looked at data from elections in Germany, France, Italy and Poland, and found that there appeared to be no pattern across nations or political systems to the numbers of votes received by each candidate, as one might well expect. On closer examination, though, they found that the differences between nations do not seem to be down to political or cultural differences, but instead reflect the differing influence of political parties within each nation. Controlling for this, they have discovered that how votes get divided between candidates really does follow a universal pattern that seems to be unaffected by political systems, culture or economic conditions.
Voters, they say, can be drawn to vote for a candidate for two main reasons. First, there’s the intrinsic appeal of a candidate, based on their personality and character, their articulated positions and ability to connect with the voters. Then there’s the influence of political parties, which attract voters to a broad philosophy or set of policies. The distribution of votes between candidates depends on both factors, and so doesn’t reflect a candidate’s personal attraction alone.
Fortunato and Castellano suspected that there might be a voting pattern hidden by this mingling of influences that held across nations. To test the idea they looked for data from elections in which party influence could be ruled out, as in elections in which the parties present a range of candidates on “open lists”, and voters choose multiple representatives from those lists.
In elections of this kind, which are held in Poland and Finland, and were used in Italy up to 1992, each voter first chooses a party and then chooses a number of candidates from that party’s list. This means the election data can be used to study the distribution of votes between candidates who all belong to the same party.
Mining 30 years’ worth of this information, Fortunato and Castellano found that the variation in voters’ behaviour between nations and political systems disappears almost entirely, everything collapsing down to one simple curve giving the likelihood for a candidate to get a particular fraction of all the votes cast for his or her party (see Graph).
Economic and political conditions changed markedly over this time. Hence, whatever process gives rise to the pattern seems to have little to do with the details of a culture or political system, and probably reflects some more basic feature of how opinions get formed (Physical Review Letters, ).
As Fortunato points out, despite how well this mathematical pattern fits the data, it isn’t likely to be helpful in predicting the outcome of an election. The pattern emerges in the statistics of elections, rather than in any specific result. It might be useful in monitoring elections for fairness, Fortunato says, as a large deviation from the expected pattern might betray a fixing of the vote.
However, a deeper understanding of why elections follow this pattern might well clarify the process by which candidates attract voters, and indicate ways to do it more effectively. In particular, Fortunato and Castellano suggest that the most important factor driving the pattern is the process of opinion formation, as voters come to form views of different candidates through news media and person-to-person contact. Using a simple computer model, they have shown that the person-to-person process is enough to generate the universal pattern observed in the data.
In their model, they supposed that each candidate starts out trying to convince others to vote in their favour. Those he or she convinces, whether they are individuals, groups or organisations, then try to convince others, and so on. These influences percolate through the social network until everyone has made a decision. Simulating the outcome of this opinion-spreading process, and using realistic assumptions on the nature of social networks – for example, accounting for the fact that some individuals and organisations have far more contacts than others – the physicists’ results fit the real data almost exactly.
“This work neatly links diffusion of information to the pattern of votes,” says political scientist Stephen Ansolabehere of the Massachusetts Institute of Technology.
Social networking may play a central role in elections, Fortunato says, and candidates might do better if they concentrate not only on their image and what they say, but on who they contact and how. “A candidate should pay more attention to the kind of people they contact,” he suggests. “Influential people and organisations may easily convince others, and the resulting cascade effect can be much more effective that directly targeting voters.”
Of course, persuasion only works on those willing to be persuaded, and those who vote along strict party lines usually won’t change their minds. Even so, some recent evidence suggests that partisan affiliations may be losing some of their power in developed democracies, perhaps making the social process more important than in the past (see “All partied out”). If so, campaigns may need to pay a little less attention to spin, and more to the details of social networking.
All partied out
Voters in the US seem to be less tied to political parties than they once were, and more likely to make up their minds about candidates using information they gather themselves, according to a study by Russell Dalton of the University of California at Irvine.
While only 25 per cent of Americans identified themselves as political independents in the 1950s, by 2000 that number was over 40 per cent. Dalton used survey data from the American National Election Studies to measure changes in two particular groups: those identified as strongly partisan and those he describes as “apartisan” (Electoral Studies, ).
“[Both these groups] are political independents,” he says, “who possess the skills and resources necessary to orient themselves to politics without depending on party labels.” In the last 30 years, Dalton found, the number of strongly partisan voters fell by nearly half, so that far fewer people now use habitual party cues to guide their behaviour. In contrast, the number of apartisan voters has doubled. Dalton suggests the rise is largely down to increasing levels of education.
