[EAS]Visual perception of docume
peter.kindlmann at yale.edu
Thu Nov 9 00:28:06 EST 2000
Mail*Link® SMTP Visual perception of document images
Forwarded from the RRE list. Sensitized by this high-impact
election event, it is worth reflecting for a moment on the
cumulative detrimental effect of interacting with computer screen
interfaces often just as poorly designed, and daily events. --PJK
Date: 11/8/00 9:15 PM
From: Phil Agre
[Eric Saund is a research scientist at Xerox PARC. I have known
him for many years. He is an authority on the architecture of human
visual perception and its consequences for the design of documents.
Here he analyzes the claims that the disputed presidential election
ballot in Palm Beach County, Florida is misleading and ambiguous.
You can see images of the disputed ballot in many places on the net,
Here is another analysis by Dan Bricklin:
I have reformatted Eric's message to 70 columns.]
This message was forwarded through the Red Rock Eater News Service (RRE).
You are welcome to send the message along to others but please do not use
the "redirect" option. For information about RRE, including instructions
for (un)subscribing, see http://dlis.gseis.ucla.edu/people/pagre/rre.html
Date: Wed, 8 Nov 2000 20:00:26 PST
From: Eric Saund <saund at parc.xerox.com>
Subject: Visual perception of document images
Controversy has arisen over a ballot in Palm Beach County, Florida
that many people have claimed is misleading and ambiguous in its
design. By at least one news report, the local election commissioner
claims "There is nothing wrong with this ballot". Clay Roberts,
director of Florida Department of Elections, is quoted to have said,
"The ballot is very straightforward. You follow the arrow, you punch
I am a visual perception scientist at the Xerox Palo Alto Research
Center. I study visual perception of document images: How do people's
brains connect the image on their retinas to the meaning formed in
Spatial layout is critical.
In my judgement, this ballot is visually ambiguous. There are two
valid ways of parsing this image. One way is by following the arrows
from candidates names to punch holes.
The arrows are rather small and not clearly shaped, and they have
little numbers next to them that add to the visual clutter. It would
be perfectly natural for your visual system to treat the arrows as
visual texture, and just filter it out.
A second valid way of parsing this image is by reading order.
When we open a book we don't take it in all at once. We direct our
attention first to the left page, then move our eyes left to right,
top to bottom. Then we look at the right page.
It would be perfectly natural for a person to read down the left
page, see the candidate they want to vote for, then stop reading.
Now they switch tasks, to finding which hole to punch. One way of
doing this is by noticing and following the arrow. Another way is
by keeping a mental count. If you want to vote for the second entry,
count down two holes. You probably couldn't vote for the sixth ballot
entry this way, but the second, sure.
Why didn't they catch this before?
If you're inspecting the ballot to proofread it, making sure no
one's name is spelled wrong, then you might not notice the layout
When you know the intent of the ballot layout, then your top-down
processing can influence your perception and resolve the ambiguity
automatically so it all looks like it makes sense. But to someone
seeing this image for the first time, in the polling booth, they have
to figure it out in the moment. It takes a different kind of looking
to notice the layout problem. It's something that good graphic
designers do intuitively.
Seeing is an automatic, unconscious process. We are not aware of
all the assumptions our minds make when we view a scene.
It is perfectly plausible that a visually ambiguous ballot could get
through the inspection process. I would not fault anyone for punching
the wrong hole on this ballot. This ballot is poorly designed.
Eric Saund, Ph.D.
3333 Coyote Hill Rd.
Palo Alto, CA 94304
(650) 812-4334 (fax)
saund at parc.xerox.com
More information about the EAS-INFO