If ever there was a clear case for usability testing, this is it. Voters in Palm Beach County have justifiably called into question the ballot used for the 2001 presidential election. Usability testing of the ballot ahead of time would have shown either that voters would make a mistake, such as choosing Mr. Buchanan when they meant to choose Mr. Gore, or would have at least indicated the form caused unnecessary confusion.

When you vote, you shouldn't even have to think about how to use the ballot. It should be drop-dead, fallin' off a log, anyone-that-is-not-brain-dead-can-do-it, easy, without any training and with very, very little instruction. Furthermore, voters should not have any doubts later about their selections.

Why is the ballot confusing? Because two columns share one column for action. Most voters were thinking "for Bush, I look top left, then select the top hole in the middle," "for Gore, I look down one, then select the next hole down… no, the arrow points to the third hole. Why? Oh, I see the second hole belongs to the first candidate on the right side, Buchanan." For most people, this train of thought happened quickly, and caused little sense of confusion. For some people, these thoughts created enough confusion to slow down their voting process significantly.

Did the design cause anyone to make a mistake? I suspect it did. I have no idea, though, if the number of mistakes would have made a difference in the election. In any case, usability testing would have identified at least the potential for mistakes, and prompted a redesign that undoubtedly would have been clearer. At the least, we would have one less cause for controversy in this election.

In reviewing the news stories, we see arguments against those complaining about the ballot such as "any competent person would figure it out," or "this is just common sense." The confounding thing about my profession is that great, usable design does not call attention to itself, and appears to be just "common sense." However, the preponderance of unusable design in web sites, software, and yes, forms, proves that usable design doesn't just happen automatically.

Furthermore, the ballot design has been defended because the public and the proper authorities were given an opportunity to review it ahead of time. Does this sound familiar to anyone who has reviewed and signed off specifications for a web site or software project? Still, somehow, problems occur in real use after the web site is launched or the software is implemented.

Again, the whole difference is context of use. You do not and cannot know if anything is "easy-to-use" without usability testing. Talking about, inspecting, and signing off on a design doesn't even come close to ensuring ease of use. Do you usability test everything and anything? No, of course there is a limit to the use of any method. But don't we all wish that Palm Beach had done some usability testing before the election?

Note: I in no way endorse usability testing the Palm Beach forms to help determine the outcome of this election. The voting is done, and we open a Pandora's box to start questioning all of the unusable voting mechanisms across the country. However, I heartily endorse usability testing for the means of future voting, including testing of the wording of propositions.

By the way, I have worked on usability of forms. We found problems via usability testing, made changes, re-tested, and undoubtedly saved a large company thousands of dollars in mistakes, services calls, etc.

You can refer to further discussion of voting usability at useit.com

who we are

what we do

success stories

the business value

how we make a difference

articles and links

Grant Consulting, Inc. located in metro St. Louis, MO, USA	1013 Bradington Court Columbia, IL 62236	info@grantconsulting.com (314) 581-0384

Copyright © 2003 by Grant Consulting, Inc. All Rights Reserved.