The last few years, possibly since the introduction of Europe-wide televoting, it is being suggested that the voting results of the Eurovision Songcontest are (getting) more related to geographical location or common history than to the songs performed by the artists on the event. This page is dedicated to a modest analysis at Leiden Institute of Advanced Computer Science (LIACS) about the extend to which these suspicions are genuine.

First of all, it might be interesting to check if there are groups of countries awarding each other more than average points. To the right we present a top ten of countries that comply with this. The list is arranged into groups of countries starting at the group that awards the most points to each other to the groups that keep less (but still many) points to itself. The score next to the countries represents the average number of points the countries gave to each other, everytime they participated together. If we take Greece and Cyprus as our example, the score of 10.47 means that during 18 events of co-participation both Greece awarded an average of 10.5 points to Cyprus and Cyprus dedicated its 10.5 points to Greece in every event! Taking into account that a country is allowed to assign a maximum of 12 points per event this is a highly remarkable situation. Greece and Cyprus are therefore the unchallenged leaders in this ranking. Groups containing countries from the Balkan and the old Sowjet Republics are rising fast, but due to a very limited amount of participations (often only once) they are left out of these statistics.

It should be noted that a value of 5.5 or less, falls within statistical expectation and therefore, altough possibly interesting, cannot be presented as extraordinary.

Furthermore, next to the examination of countries that award lots of points to each other, it might be interesting to review a list of countries that award many points to other countries (in one direction). Of course their is a strong relation between the both, as Greece and Cyprus teach us, but the inclusion of 'one way traffic' makes it possible to include effects that things like immigration etc. might have on voting. Just like in the list above the score depicts the average amount of points that the leftmost country awarded the rightmost country in every match they co-participated.

Despite the fact that there is strong correlation there are some remarkable differences between the two lists. In this one, for example we may find Monaco, awarding its big brother France with lots of points while France doesn't repay the favour. Rising stars are the Netherlands and Germany awarding points to Turkey, but this phenomenon only appears in the last five years. In some of the above cases Russia (participating under the name of Russian Federation) also contains some countries that participated independently for the first time in the 2004 contest.

Completely different from looking at countries awarding each other the bulk of points is looking at groups of countries giving their preference to the same of group of different other countries. You might, for example, expect neighbouring countries to share the same musical heritage or tradition and therefore award points to the same countries. In the top ten list to the right all averages were taken into account except the points arwarded to each other. The score assigned to the two countries depicts the 'difference' between the respective voting behaviour. In concreto, this means that a lower score means more resemblance in voting behaviour. Remarkably enough we note a large number of Northern- and Western-European countries in this list. This could indicate the sharing of a very specific taste or could mean they all behave equally average. No conclusions can therefore be drawn from this list about the extend to which 'cooperation' takes place. We also note an occurence of the Netherlands for the first time (even 3 times). We can conclude that the typical Dutch person votes alike its German, Norwegian or French Counterpart.

The image below provides a more complete image about the resemblance of voting behaviour. The picture not only depicts the countries mentioned before in the top 10 list, but all the contestants in the songcontest throughout the years. The physical distance between two flags represents the extend to which the voting behaviour of the two corresponding countries deviate; therefore, the further two flags are placed from each other the less their voting behaviour looks alike and vice versa. Furthermore the same mapping is depicted in the second image where all the flags are replaced by the European region the countries respectively belong to. This will provide a more clear image to investigate whether or not the voting behaviour of regions looks alike.

Below are a number of documents aggregated during the analysis:
- Total voting data Eurovision Songcontest
- Distance matrix Eurovision Songcontest

Beforementioned analysis took place at the Algorithms group of LIACS of Leiden University of prof. dr. J. N. Kok. Questions can be directed through email to drs. T.K. Cocx

During the analysis a number of (re)sources was used:
- Voting data 1957-2003 van Songcontest.nl
- Voting data 2004 van The Eurovision Database
- Flags from Eurosong.nl