Cybergeo (Jan 2011)
L’impact du voisinage géographique des pays dans l’attribution des votes au Concours Eurovision de la Chanson
Abstract
The Eurovision Song Contest has been held every year since 1956. The originality of this contest lies in the points awarding system. There is no single jury but, on the contrary, each country is asked to award a given number of points to the countries which performed its favourite songs. Beyond the interest of such a system to collectively rank countries, it provides a long story of the swapped points between countries. Such information makes it possible to detect couples of countries {voter, performer} for whom votes are not exclusively guided by song quality. Currently the media covering the event regularly mention a bias in the votes distribution: according to them, this bias would be caused by geographical proximity and would lead to blocs of nearby countries which overwhelmingly vote for each other. In this article, we try to discuss this assumption. This latter requires to answer the following question: “how can we assess the influence of spatial proximity on the social ties formation?”. Besides this issue, several methodological challenges appear. First, we examine the votes of the 1993-2008 period and we identify the social ties of interest, that is the couples of countries {voter, performer} whose votes significantly diverge from the reference situation (i.e. a competition on song quality). Then, we compare the resulting social network with the spatial countries network by a well-suited statistical method to prove that “over-votes” concern nearby countries. Finally, we highlight clusters of countries tending to over-vote for each other and, in that respect, we define clustering criteria which make sense. We show that these blocs strongly structure the abnormally high votes of the 2009 event. This analysis method combines geographical and social networks and can be extended to the study of phenomena concerning relations between spatialised entities.
Keywords
