[E-voting] more on proof-reading printouts from DREs
cansbro at eircom.net
Thu Jun 14 12:19:38 IST 2007
Bev Harris had this interesting response to my BlackBoxVoting.org
posting about the research showing most voter's don't catch errors on
screen or on a printout. [See the original for easier-to-read formatting]
Proofreading from a computer screen is generally considered more
error-prone (even!) than proofreading from paper.
*I just published this information in a comment on the BradBlog story:*
I'm glad to see this study. Just had a very passionate debate in New
Hampshire with Dr. Ron Rivest about this very topic.
Studies on cognitive processes have been very clear on this point, and
go back 50 years. There is a vast difference between ACTING on a
document, such as writing on it or marking on it, and PROOFREADING a
document. The brain is, not surprisingly, set up to take shortcuts.
Especially when one assumes that there will be no errors, as most people
do when computers print out something they have seen on a screen,
people's brain is programmed literally to skip over those errors.
The cognitive process of taking "shortcuts" in visual perceptive tasks
should not surprise us. If our brains didn't do this, we'd have to
ponderously analyze each thing we look at. That ... is ... red ...
somewhat round ... has a stem ... and ... therefore ... must ... be ...
an ... apple. No. The brain just tells you APPLE.
Especially with print proofreading tasks, our brain loves to take
shortcuts. In fact, in Psychology 101 textbooks there are sample
paragraphs where an amazingly high percent of people ALWAYS misread
words put together in certain ways, because of the brain's tendency to
This is such a strong tendency that professional proofreaders --- and
let's not kid ourselves, VVPATs and ballot marking devices involve
PROOFREADING skills, not just reading skills --- many professional
proofreaders read paragraphs BACKWARDS to block the brain's annoying
(for a proofreader) tendency to skip over errors.
*Academics are presumed to be speaking (1) in their area of expertise
and (2) after applying rigor*
Here's another point. I mentioned that I was debating this with Dr. Ron
Rivest, a cryptography and mathematics expert. One of the things I took
him to task for --- and I take every computer scientist to task for ---
is the failure to use a disclaimer when issuing opinions about voters
using VVPATS vs. marking a ballot themselves. Rivest told me he was
unaware of studies showing that proofreading a ballot was inferior to
marking it yourself. I told him:
1) Since cryptography and mathematics are an entirely different field of
study than cognitive process research, one wouldn't expect you to carry
around a compilation of the research in your head --- and you should be
clear that you have no expertise whatsoever in these areas when you
2) People assume that academics, even if a subject is not their
specialization, apply a certain rigor to evaluating information. Such
rigor would involve doing a review of the literature and examining
studies for replication and further clarification. When an academic
pronounces an opinion, there is an assumption that they have followed
scientific procedure and applied appropriate rigor. That's simply not
the case with the current crop of computer guys who presume to advise
and write legislation on the very multi-disciplinary field of elections.
To which Rivest, who is a very nice guy by the way, replied that in
academia everyone knows to discount opinions when they are not the
person's area of expertise. To which I replied that such is NOT the case
when "experts" provide input to congress, or sit on panels, and start
blabbing about things not in their area of expertise. To which he
replied that he will certainly take that point to heart and start being
more careful about stating disclaimers on opinions not in his area.
*Elections are a multidisciplinary endeavor*
This lengthy post is to illustrate the importance of bringing in people
who actually understand cognitive processes, rather than propping up
mathematicians and computer programmers to pronounce their OPINION about
what voting system is most appropriate.
And by the way, as one with a degree in applied psychology and decades
of experience in writing and publishing fields which involve
PROOFREADING, I can tell you that the number of studies that show the
high number of errors for proofreading tasks and the studies on the
brain's tendency for closure (shortcutting) are huge.
You cannot teach people to proofread by sticking a sign up that says
"check your ballot" and many people will not be able to proofread
accurately even when they take extra time and try hard.
*How many people have reading disabilities?*
And while we're bending over backward for disabilities, has anyone
thought about the impact of forcing DYSLEXICS to proofread their ballot?
More information about the E-voting