Trouble with local elections: wrongly voted?
“In a system designed for use by everyone, even the smallest errors in user experience can bring the whole system down. I would think it has now become clear to everyone that this just cannot happen.”
“E-voting can accelerate voting procedures, improve the reliability of elections, decrease the workload of authorities and save costs. E-voting poses no threat to the integrity of the voting process.”
This is what was written before the first Finnish experiment in e-voting. The same text can be found in the Ministry of Justice-administered website vaalit.fi. It was apparently written in earnest.
Karkkila, Kauniainen and Vihti are the first towns to enjoy councils with council members with luck on their side. Some were even luckier than others. Hopefully, that good luck will be reflected through their council work for the citizens.
What went wrong?
Thus far, we know that the voting machines’ program was slow. When planning usability, it was assumed that the program would not be sluggish. Therefore, it was assumed that there was no need to include messages telling voters what the machine is doing.
Did the users receive information on whether vote has been counted or not? No. Did the users receive information on whether their vote has been saved and forwarded in the system? No. After all, the assumption was that the system would be sluggish, so there was no need for feedback.
The e-voting success story was a close call. In the testing phase, a corresponding situation, or freezing of the system, only happened once. The story doesn’t tell whether the situation was equivocally remarked on in a meeting with a mouth full of Danish pastry or if the matter was shrugged off because the pastry was so dry that it made speaking impossible. In any case, the faulty system was delivered for use in a real election.
The voting system must be foolproof. Because there is no wish to identify the voter on the basis of their votes, no one can tell what the voter was trying to do. Did he/she attempt to sabotage his/her voice by using the machine in the wrong way on purpose, i.e. giving a Mickey Mouse vote, or did he/she leave the machine satisfied in the belief that his/her voted had been recorded.
Reorganising the election is a lousy substitute because new elections will not provide the same results as the initial voting. After seeing the results, many people see opportunities for tactical voting and organising new elections would be only slightly less contemptuous of democracy than leaving the results of the first e-voting in force.
The e-voting experiment provided a valuable lesson for the whole Finnish debate on information society. This is due to the fact that problems pertaining to the efficiency of public administration will be solved electronically, even in the future. Voting is one of the few systems that must be suitable for all users, but it is certainly not the only one.
In a system designed for use by everyone, even the smallest errors in user experience can bring the whole system down. I would think that it has now become clear to everyone that this just cannot happen.
I also hope that an old wisdom is kept in mind when a program progresses from the production stage to production use: the road to hell is paved with good intentions. Despite this initial setback, it is worth keeping on trying.
Juha Lamminkari
Juha Lamminkari’s column on e-voting was published on Digitoday website 16 December 2008 and on It-viikko website on 19 December 2008.