The cost of failing to usability test designs before deployment has unfortunately been shown again in Scotland's controversial election results.
We see it all the time in usability testing of websites - the fact that users don't read instructions on forms. Instead, they tend to to start filling it in straight away, particularly when the form starts with easy questions about their name etc. The same thing happens on offline paper forms; a subconscious voice says 'I'll read the instructions if I need them' and potentially useful information is missed.
But you would think that users would take a bit more care when filling out a voting form, wouldn't you? Apparently that is not the case, hence the controversy over last week's important local elections in Scotland, where 1 in 20 votes have had to be thrown out in Britain's worst ever voting debacle.
The reason for the mistake? As shown in this example on the BBC site, voters had to put two crosses on their Holyrood voting papers - one for their constituency and one for the regional list - but it appears many wrongly put two crosses in one section. Simultaneously staging the council elections, in which voters had to rank candidates, also caused confusion.
Having completed the ballot papers myself through a postal vote, I agree that the system did seem confusing. On one ballot paper, the instruction was to use a single X to mark your choice - the traditional method. On another, I was asked to rank the candidates by order of preference, as part of the Single Transferable Vote (STV) system for local authority candidates.
I had the luxury of doing all of this at home and although I found it odd, I was put right by the instructions and hopefully did it right (but will never know for sure I suppose). The bigger problem I found was the rather complex origami needed to put the completed ballots in one envelope, then fold back another piece of paper to ensure that a return address was shown in the lower left, and put all of that into another envelope, aligned with the window of the envelope.
Voters in polling booths however often have other factors that make the likelihood of errors even greater. They may be in a rush, getting their vote in on their way to work, and it is certainly not a familiar task or environment, which will increase the likelihood of errors. As shown in the US presidential election in 2000, with the controversy over the 'hanging chads', users' ability to fully and completely follow voting instructions should not be overestimated. That event showed the importance of conducting usability testing of ballots, rather than relying on the view of the local Electoral Commission that it seems clear enough.
What does all of this have to do with the online world? We are faced with forms all of the time, and some basic rules and conventions have fallen into place to help minimise the chances of errors, such as :
- Keep the instructions clear and concise.
- Locate the instructions close to the form fields themselves.
- Apply ways to visually emphasise instructions, such as shading, bolding, larger fonts and borders.
- Provide examples on the correct completion of the forms.
- Provide clear and helpful error messages if the form does not validate (admittedly something that can't happen on offline ballot papers).
Hopefully the Electoral Commission has learned an important lesson and when major changes to the standard voting method are made, they will be reviewed carefully and tested with a few end-users to refine the designs.
Chris Rourke is the managing director of User Vision.