While security fears always get a regular airing in debates about electronic voting, another question that has so far escaped attention is whether electronic voting itself can change who people vote for. We have known for decades that the structure of paper ballots has an impact on the way people vote. We know there is a small bias in favour higher placed candidates on the vertical lower house ballot paper, and a left to right bias on horizontal upper house ballot papers. This bias by position is as a result of the order in which people read the ballot paper. Some electors seem to stop and vote for the first candidate or party they recognise rather than look at all options. It can also lead to donkey voting, where people simply number candidates top to bottom or left to right. These factors get worse the larger the ballot paper. Some of the giant ballot papers in recent years have shown evidence of voter confusion as voters have struggled to find the parties they do know amongst a profusion of micro party offerings.
At the September 2013 Senate election the Liberal Democrats drew column A on the NSW Senate ballot paper and recorded an unprecedented 9.5% of the vote. At an election where the Coalition polled its highest House vote in NSW since 1975, it simultaneously recorded its lowest Senate vote since 1943 and the largest ever gap between % vote for elections in the two houses. The gap was considerably lower in electorates held by the National Party rather than the Liberal Party, suggesting confusion over the printed party name ‘Liberal Democrats’ in Column A of the Senate ballot paper played a part.
Last November’s Victorian Legislative Council election provides further evidence for the party name confusion hypothesis. As the table below shows, the Liberal Democrat vote was more than two percentage points higher in the three regions where the party appeared to the left of the Liberal or Liberal/National group on the ballot paper.