Statement Concerning Usability Issues of the Palm Beach Ballot
The undersigned include prominent researchers in various technical
fields including interface usability, human-computer interaction
(HCI) and the psychology of notations, which address the usability
of various visual representations of information. We have reviewed the
design of the Palm Beach County ballot used in the November 7, 2000 U.S.
Presidential election and the related circumstances. The following
represents our consensus opinion on inherent usability problems in this
ballot:
1) The layout and "notational design" of this ballot can be
demonstrated to have a strong probability of creating recurring
patterns of error for voters, using a variety of different
methodological measures and models which suggest various "routes" for
these errors. Some of the work describing these patterns of error can
be seen at:
2) There is no evidence from the ballot design itself that there was
deliberate intent to mislead voters in the design of the ballot.
Research has shown that designers find it difficult to imagine
themselves in the position of the user of a device or a notation. As
a result designers produce systems that are very clear to themselves,
and they are sometimes genuinely surprised when other
people find those systems difficult to use.
3) The patterns of error include various possible "routes" for error,
which may have resulted in votes being lost or gained by various parties
in the election. However, though the ambiguity of the ballot could in
principle have affected any party, due to particular aspects of the design
it lends itself to having disproportionate affect on certain candidates.
Thus the probable errors introduced by the ballot design flaws cannot be
dismissed as being evenly distributed.
4) There is strong evidence, based on research literature correlated
with reported reactions from voters, suggesting that the ballot
layout could have created patterns of error which would not have been
dependent on particular voter characteristics such as age, visual
acuity or impairment, English language skills, etc.
5) There is strong evidence, based on research literature correlated
with reported reactions from voters, suggesting that many voters
would have made errors without realizing the error was being made at
the time. Thus asking for help or clarification from voting officials
or requesting replacement ballots would not have been an option. In
effect, through no deliberate attempt to mislead, these voters would
have been misled into casting votes for other candidates.
6) The patterns of error could help explain dual-voting anomalies as
well as votes going to candidates other than that intended by the
voter. Thus it is likely that many of the dual-vote invalid votes
were directly influenced by the layout and poor supporting
documentation of the ballot.
7) The question of the legality of the ballot under Florida law
involves questions such as user perception of ordering and
sequencing, and clear association of candidate names with voting
buttons, which should be determined making use of relevant research
in usability and related fields.
8) Regardless of the ballot's conformance to Florida state law,
regardless of the best of intentions of the designers of the ballot
and those who approved it prior to the election, and regardless of
attempts to counteract the poor layout with advisory memos to voting
staff, a highly probable result of the use of this ballot in the
election was the disenfranchisement of the rights of many of the
voters in Palm Beach County. Any reasonable remedies that can be
applied to this miscarriage of the voters' intent should be explored.
Thank you for your attention.
*************
Signers:
(1) T. R. G. Green also at:
preferred postal address: Computer-Based Learning Unit
Oriel House, 27 Allerton Park, University of Leeds
Leeds LS7 4ND, U.K. Leeds LS2 9JT, U.K.
Dr. Thomas Green is an Applied Psychologist who has been researching
the usability of visual notations since the 1970s. Currently a
Research Fellow at the Computer-Based Learning Unit at the University
of Leeds, he is generally recognized as one of the "founding fathers"
of the field of psychology of notations, which addresses the
usability of various visual representations of information.
Alan Blackwell is a University Lecturer at Cambridge University,
and Research Fellow of Darwin College Cambridge. He is
responsible for teaching the main human-computer interaction
course at Cambridge University. He is also a leading researcher
in visual notations who, among other activities, convened the
first interdisciplinary conference on Thinking with Diagrams.
(3) Bob Spence bob_spence@escore.com
Senior software Engineer: eScore.com
Bob Spence has been a software engineer since the late 1960s. He has been a
Macintosh software developer since the Macintosh was introduced in 1984, with
a strong interest in issues of software usability.
(4) Kevin Fox kfox@cs.berkeley.edu
Student Researcher
Group for User Interface Research
University of California at Berkeley
http://fury.com/ (510) 845-9502
Kevin Fox is a cognitive scientist specializing in visual perception and
cognition. Before his return to UC Berkeley, he was employed as an
information architect at marchFIRST, Ikonic Interactive, and Eleven, Inc.,
where he designed user interfaces for several prominent e-commerce
clients.
(5) David Adler DAdlersan@aol.com
David Adler is a user interface and Computer-Human Interaction (CHI)
professional who has been a senior product and interface designer at
Interleaf, Lotus Development and iCast.
(6) Paul Resnick presnick@umich.edu
Paul Resnick is an Associate Professor at the University of Michigan School
of Information, where he teaches master's students about both usability and
information policy issues. He received a Ph.D. in computer science from
MIT.
(7) Mark A. Simos (simos@mediaone.net)
Co-founder of the software company Synquiry Technologies, Ltd. Not a usability specialist, he initiated and
circulated this position paper among colleagues in the field.
