E-voting favours no party

Photo: Kristjan Teedema / Postimees

Fears of e-elections being biased by nature not scientifically proven.

As revealed by report underway at University of Tartu, there is no link strong enough, between e-voting and party popularity, to claim e-elections would give some political parties an advantage. The report analyses e-election data since its beginnings in Estonia, in 2005.

According to the report, voters opt for e-election due to demographics, attitudes and behaviour, not political preferences.

«In other words, Centre Party will not get less votes as Reform Party backers vote via Internet,» said Kristjan Vassil, author of report and research fellow for comparative politics at University of Tartu, a long-time researcher of e-elections.

The report is being prepared upon order of a scientific centre in Estonia called Software Technology and Applications Competence Centre (STACC); full content thereof will be published after local elections.

According to Kristjan Vassil, their international research team has conducted queries on e-elections, in Estonia, starting the 2005 elections and by now it has evolved into a world-leading unique  e-elections study, the longest one available – complete with time series. «The main goal has been determining the typical e-voter; how they differ from voters at large; the effect of e-elections on political participation etc.»

The report requested by STACC is dedicated to two issues, insufficiently studied so far: are e-elections politically biased; and, how has trust towards e-election system evolved over time.

To find out whether e-elections are politically neutral or biased, the question needs to be clearly defined. «I have arrived at the wording that a certain means of election is politically biased only when the election results, at the end of the day, would be different to what they would have been without use of the said means,» explained Kristjan Vassil. «Empirically, it is not possible to say if e-elections are useful to a certain party or not, as we may only survey elections containing e-voting or not; we cannot view both [kinds] at the same time.»

To proceed, we need to determine the logical condition which would need to be filled for some party to have an advantage due to e-elections. Should such conditions be empirically possible, the probability or their occurrence would be empirically analysable.

For instance: one way for e-elections to affect election results would be if e-elections attracted new voters – all supporting the same party.

Studies since 2005 show that e-elections indeed motivate a small amount of voters to participate.

The next step is finding out if the ones motivated by e-elections will be inclined to favour a definite party – or would their preferences be random, without a clear pattern.

«This we have analysed by descriptive methods and more sophisticated econometrical methods. Currently, we have no clear evidence of the voters mobilised favouring a definite political party. Rather the opposite: by their preferences, the voters are very heterogeneous,» assured Mr Vassil.

May e-elections bring about a situation where an e-voter will vote for a party he/she would not have voted for at the ballot box? «E-voters do not change votes more often. Rather the opposite is revealed: changing preferences is much more likely amongst ordinary voters,» said the author of the report.

Will participation at e-elections serve as means to forecast voter preference? «Analysis reveals that participation in e-voting is not linked to favouring any party on a level considered statistically vital,» noted Mr Vassil.

Why, then, do e-voting and ballot bow results differ, during the same elections? According to Mr Vassil, a large amount of people have shifted into e-voting, without changing political preference. However, supporters of certain parties are more prone to changing the channel they vote through. Even so, the overall elections results remain unchanged by that.