This quite astounding article in the New Scientist posits a fascinating theory. An analysis of 30 years worth of election results from all round the globe, conducted by Santo Fortunato of the Institute for Scientific Interchange in Turin working with Claudio Castellano of the University of Rome, has shown that voting follows the same pattern, regardless of culture, nationality, economic situation or history. And what’s more, social networking can help explain it. Says Fortunato:
"When it comes to voting, 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."
It’s already known that in countries that have elections with many candidates, voting follows a pattern where most candidates receive a small number of votes whilst a few do much better. It’s a regular pattern with twice as many candidates receiving 20% of the vote as did 40%, twice as many again receiving 10% as did 20% and so on. Studying countries with different electoral systems however, the pattern soon breaks down. Until, that is, you control for the differing influence of the political parties in each country. Remove that, says Fortunato, and how votes get divided between candidates really does follow a universal pattern that is apparently unaffected by political systems, culture or economic conditions.
It seems that voters are attracted to candidates for two main reasons: the intrinsic appeal of the candidate (personality, character, their positioning, ability to connect with the voters) and the influence of political parties (the attraction of a broad philosophy or set of policies). Studying data from elections in which parties present a range of candidates and voters choose multiple representatives from the same party (such as in Poland, Finland and Italy pre-1992), enabled the academics to study situations in which party influence could be ruled out. In these cases variations in voter’s behaviour between nations and political systems disappears almost entirely. Regardless of economic and political conditions, the likelihood for a particular candidate to get a particular fraction of all the votes cast for his or her party boils down to a simple graph, reflecting "some more basic feature of how opinions get formed."
(image courtesy New Scientist)
So it’s the actual process of opinion formation that drives the pattern. And using simple computer models that replicate a simple social networking process (including accounting for the fact that some individuals have more contacts than others), the model matches the real data almost exactly. In this way, the diffusion of information is linked to the pattern of votes – social networking, and network effects, potentially play a key role in deciding elections.
The conclusion, says Fortunato, is that candidates should concentrate on not only their image and what they say, but who they contact. "Influential people and organisations may easily convince others, and the resulting cascade effect can be much more effective than directly targeting voters."
The parallels with advertising are inescapable.