Tuesday, March 27, 2012

Is online voting all its cracked up to be?

The Canadian New Democratic Party elected their leader at the weekend. As it happens they chose a sitting MP, Thomas Mulcair. It's a curiosity that just one year ago Mulcair was the sole NDP MP in Quebec in a generation (and only the second one ever) but now the NDP have most of the seats in Quebec (and Quebec has most of the NDP MPs) and Mulcair's victory is in part because the party needs to retain its position in Quebec. There's a further irony that what began as a western protest party is now led by a Quebecer (although this is not the first time that's happened - see the Social Credit Party of Canada).

But ironies and Canadian specifics aside, the leadership election experienced severe problems with voting. Votes could be cast in three ways:
  • Alternative Vote postal ballots, to be sent in advance with preferences chosen without knowing who would get knocked out early & who would endorse who.
  • Votes cast at computer terminals at the party convention, with delegates voting for a single candidate in each round with about an hour to vote each time.
  • Votes cast online from home, again for single candidates in each round with an hour.

And far from being a smooth process the voting descended into chaos with the website struggling to cope and being allegedly hit by a denial-of-service attack. (The Globe and Mail: Hackers attack NDP, delaying electronic leadership vote) The result was a mess as many people took ages to access the site and cast their vote, others either ran out of time or gave up, voting had to be extended multiple times and even split into separate periods for the convention and rest of the world, and eventually the leader was declared much later than expected.

Exactly what went wrong and why expected safeguards didn't work is no doubt already the subject of an investigation. But as an example of high profile online voting it's a worrying sign of how vulnerable it can be and how there's strong potential for disruption that deters many people from casting their votes.

Now sure online voting is used for a lot of private organisations. But most of their votes are much lower key with much less risk. The last online vote I participated in was for an alumni rep on a college council - who sets out to sabotage that sort of election? Rather fewer people than a high profile party or public election and so less conscious or financial consideration needed to be given to protecting the site.

I am sure there are all manner of procedures and safeguards that can be used for online voting. But either they're standard and the hackers still got through, or else the decision was made without enough consideration of the problem. And remember this is a quite technical and modern area it's unsurprising if those making the ultimate decisions and paying the cheques do not have the strongest grasp of every aspect needed. Prospective contractors can make their pitches based on the aspects, but they're ultimately selling to decision makers who have not grown up with computers all around them and often the necessary budgets are not forthcoming. A similar thing could be seen back during the run on Northern Rock when the bank's website proved unable to cope with the demand.

Could we safely use online voting for public elections? Leaving aside the wider issues I think the security factor is a big one that needs to be demonstrably handled first before jumping on the online bandwagon. A rerun of the NDP's problems on a public scale would bring a result into dispute, potentially ending up in the courts if the result was close, and would be far too tempting a target for the most disruptive. And with so much of our electoral administration split across many, many different bodies, including a lot of councils facing strong financial squeezes, I am not persuaded that no expense would be spared to ensure problems did not happen.


David Morris said...

For many years, I have been a firm believer that e-voting can be a success.

You correctly state that those people wanting these systems don't always have the greatest understanding of the technology. However, I definitely think the security issues can be addressed in time.

Having an online method of voting has been proven to increase turnout, so I wonder what the voting numbers were for that particular leadership election in the past.

Tim Roll-Pickering said...

It's tricky to compare because the 1995 election was done by representative delegates at a convention, then the 2003 election combined mail in advance votes and on the day internet voting (and had just one round) albeit with 25% of votes cast by affiliated organisations (dropped for 2012). So there's no direct point of comparison, but most comment suggested the turnout was lower.

One general factor to bear in mind is that the NDP has integrated federal & provincial wings in all but one province (rare amongst the major Canadian parties) so there may be many people entitled to vote for the federal leader who only really joined primarily for the provincial side, particularly in British Columbia.

I'd like to see the proof that online voting in itself automatically increases turnout. I've seen all manner of figures from students' unions that show the figures in all directions, but with other factors such as how well the organisation and candidates adapt to the newer methods, and how much voter education & direction there is. And of course the nature of the organisation and electorate is a big factor - private membership organisations where people have consciously chosen to partake usually have a more involved electorate than when it's a side feature or when people are automatically added to an electoral roll without consciously realising it.


Related Posts Plugin for WordPress, Blogger...