ArticlePDF Available

Emotion & Design: Attractive Things Work Better

Don Norman Revised February 24, 2003
Emotional Design
1. ATTRACTIVE THINGS WORK BETTER
Donald A. Norman
1. Attractive Things Work Better
Three Levels of Processing: Visceral, Behavioral and Reflective
Focus and Creativity
The Prepared Brain
References
Endnotes
Noam Tractinsky, an Israeli scientist, was puzzled. Attractive things certainly
should be preferred over ugly ones, but why would they work better? Yet two
Japanese researchers, Masaaki Kurosu and Kaori Kashimura1, claimed just that.
They developed two forms of automated teller machines, the ATM machines that
allow us to get money and do simple banking tasks any time of the day or night.
Both forms were identical in function, the number of buttons, and how they
worked, but one had the buttons and screens arranged attractively, the other
unattractively. Surprise! The Japanese found that the attractive ones were easier
to use.
Tractinsky was suspicious. Maybe the experiment had flaws. Or perhaps the
result would be true of Japanese, but certainly not of Israelis. “Clearly,” said
Tractinsky, “aesthetic preferences are culturally dependent.” Moreover, he
continued, “Japanese culture is known for its aesthetic tradition,” but Israelis?
Nah, Israelis are action oriented—they don’t care about beauty.2 So Tractinsky
redid the experiment3. He got the ATM layouts from Kurosu and Kashimura,
translated them from Japanese into Hebrew, and designed a new experiment,
with rigorous methodological controls. Not only did he replicate the Japanese
findings, but the results were stronger in Israel than in Japan, contrary to his
belief that beauty and function “were not expected to correlate” -- Tractinsky was
so surprised that he put that phrase “were not expected” in italics, an unusual
thing to do in a scientific paper.
This is a surprising conclusion. In the early 1900s, Herbert Read, who wrote
numerous books on art and aesthetics stated that "it requires a somewhat
mystical theory of aesthetics to find any necessary connection between beauty
and function,” 4 and that belief is still common today. How could aesthetics affect
how easy something is to use? I had just started a research product examining
the interaction of affect, behavior, and cognition, but these results bothered me –
I couldn’t explain them. Still, they were intriguing, and they supported my own
personal experiences, some of which I described in the prolog. As I pondered the
experimental results, I realized they fit with the new framework that my research
collaborators and I were constructing as well as with new findings in the study of
affect and emotion. Emotions, we now know, change the way the human mind
solves problems – the emotional system changes how the cognitive system
operates. So, if aesthetics would change our emotional state, that would explain
the mystery. Let me explain.
Until recently, emotion was an ill-explored part of human psychology. Some
people thought it an evolutionary left-over from our animal origins. Most thought
of emotions as a problem to be overcome by rational, logical thinking. And most
of the research focused upon negative emotions such as fear, anxiety, and anger.
Modern work has completely reversed this view. Science now knows that
Draft manuscript: Norman, D. A. (2004). Emotional Design: Why We Love (or Hate) Everyday Things. New York: Basic
Books. Copyright © 2002, 2003 Donald A. Norman. All rights reserved. http://www.jnd.org don@jnd.org
Don Norman: Chapter 1: Emotional Design 2 3/23/2003:
evolutionarily more advanced animals are more emotional than primitive ones,
the human being the most emotional of all. Moreover emotions play a critical role
in daily lives, helping assess situations as good or bad, safe or dangerous. As I
discussed in the prologue, emotions aid in decision making. Most of the research
on emotions has concentrated upon the negative: stress, fear, anxiety, anger. But
positive emotions are as important as negative ones -- positive emotions are
critical to learning, curiosity and creative thought and today, research is turning
toward this dimension. One finding particularly intrigued me: The psychologist
Alice Isen and her colleagues have shown that being happy broadens the
thought processes and facilitates creative thinking. Isen discovered that when
people were asked to solve difficult problems, ones that required unusual “out of
the box” thinking, they did much better when they had just been given a small gift
– not much of a gift, but enough to make them feel good. When you feel good,
Isen discovered, you are better at brainstorming, at examining multiple
alternatives. And it doesn’t take much to make people feel good: all Isen had to
do was ask people to watch a few minutes of a comedy film or receive a small
bag of candy. 5
We have long known that when people are anxious they tend to narrow their
thought processes, concentrating upon aspects directly relevant to a problem.
This is a useful strategy in escaping from danger, but not in thinking of
imaginative new approaches to a problem. Isen’s results show that when people
are relaxed and happy, their thought processes expand, becoming more creative,
more imaginative.
These – and related – findings suggest the role of aesthetics in product design:
attractive things make people feel good, which in turn makes them think more
creatively. How does that make something easier to use? Simple, by making it
easier for people to find solutions to the problems they encounter. With most
products, if the first thing you try fails to produce the desired result, the most
natural response is to try again, only with more effort. In today’s world of
computer-controlled products, doing the same operation over again is very
unlikely to yield better results. The correct response is to look around and see
what alternatives exist. This tendency to repeat the same operation over again is
especially likely for those who are anxious or tense. This state of negative affect
leads people to focus upon the details that are giving trouble, and if this fails to
provide a solution, they get even more tense, more anxious, and increase their
concentration upon those details. Contrast this behavior to that of people who are
in a positive emotional state, but encountering the same problem. These people
are apt to look around for alternative approaches, which is very likely to lead to
the appropriate response. Afterwards, the tense and anxious people will complain
about the difficulties whereas the relaxed, happy ones will probably not even
remember them. In other words, happy people are more effective in finding
alternative solutions and, as a result, are tolerant of minor difficulties. Herbert
Read thought we would need a mystical theory to connect beauty and function.
Well, it took one hundred years, but today we have that theory, one based in
biology, neuroscience, and psychology, not mysticism.
Human beings have evolved over millions of years to function effectively in the
rich and complex environment of the world. Our perceptual systems, our limbs,
the motor system – which means the control of all our muscles – everything has
evolved to make us more function more effectively in the world. Affect, emotion,
and cognition have also evolved to interact with and complement one another.
Cognition interprets the world, leading to increased understanding and
knowledge. Affect, which includes emotion, is a system of judgment: good or bad,
safe or dangerous. It makes value judgments, the better to survive.
The affective system also controls the muscles of the body and, through chemical
neurotransmitters, changes how the brain functions. The muscle actions get us
Don Norman: Chapter 1: Emotional Design 3 3/23/2003:
ready to respond, but they also serve as signals to others, which provides yet
another powerful role of emotion – as communication: our body posture and
facial expression tells others our emotional state. Cognition and affect –
understanding and evaluation. Together they form a powerful team.
THREE LEVELS OF PROCESSING: VISCERAL, BEHAVIORAL AND REFLECTIVE
Human beings are, of course, the most complex of all animals with brain
structures that are accordingly complex. A lot of preferences are wired in at birth,
designed to be part of the body’s basic protective mechanisms. But we also have
powerful brain mechanisms for accomplishing things, for building, constructing,
creating, and acting. We can be skilled artists, musicians, sports players, writers,
or carpenters. All this requires a much more complex brain structure than is
involved in automatic responses to the world. And finally, unique among animals,
we have language and art, humor and music. We are conscious of our role in the
world and we can reflect upon past experiences, the better to learn and reflect
forward to the future, the better to be prepared and reflect inward upon current
activities, the better to supervise them.
My studies of emotion, conducted with my colleagues Andrew Ortony and William
Revelle, Professors in the Psychology Department at Northwestern University –
suggests that these human attributes result from three different levels of brain
mechanism: the automatic, prewired layer, the visceral level; the part that
contains the brain processes that control everyday behavior, the behavioral level;
and the contemplative part of the brain, the reflective level6. Each level plays a
different role in the total functioning of people. And, as I discuss in detail in
chapter 3, each level requires a different style of design.
Figure 1.1. Three levels of processing: Visceral, Behavioral, and Reflective.
The visceral level is fast: it makes rapid judgments of what is good or bad, safe or
dangerous, and sends appropriate signals to the muscles (the motor system) and
alerts the rest of the brain. This is the start of affective processing. These are
biologically determined and can be inhibited or enhanced through control signals
from above. The behavioral level is the site of most human behavior. Its actions can
be enhanced or inhibited by the reflective layer and, in turn, it can enhance or inhibit
the visceral layer. The highest layer is that of reflective thought. Note that it does not
have direct access either to sensory input or to the control of behavior. Instead it
watches over, reflects upon, and tries to bias the behavioral level. (Modified from
Norman, Ortony, & Russell, 2003)
Don Norman: Chapter 1: Emotional Design 4 3/23/2003:
The three levels in part reflect the biological origins of the brain, starting with
primitive one-celled organisms and slowly evolving to more complex animals, to
the vertebrates, the mammals and finally, the primates, of which we are a
member. For simple animals, life is a continuing set of threats and opportunities,
and an animal must learn how to respond appropriately to each. The basic brain
circuits, then, are really response mechanisms: analyze a situation and respond.
This system is tightly coupled to the animal’s muscles. If something is bad or
dangerous, the muscles tense in preparation for running, attacking, or freezing. If
something is good or desirable, the animal can relax while also approaching and
taking advantage of the situation. As evolution continued, the circuits for
analyzing and responding improved and became more sophisticated. Put a
section of wire mesh fence between an animal and some desirable food: a
chicken is likely to be stuck forever, straining at the fence, but unable to get to the
food; a dog simply runs around it. Human beings have an even more developed
set of brain structures. They can reflect upon their experiences and communicate
them to others. Thus, not only do we walk around fences to get to our goals, but
we can then think back about the experience – reflect upon it – and decide to
move the fence or the food, so we don’t have to walk around the next time. We
can also tell other people about the problem, so they will know what to do even
before they get there.
Animals such as lizards operate primarily at the visceral level. This is the level of
fixed routines, where the brain analyzes the world and responds. Dogs and other
mammals, however, have a higher level of analysis, the behavioral level, with a
complex and powerful brain that can analyze a situation and alter behavior
accordingly. The behavioral level in human beings is especially valuable for well-
learned, routine operations. This is where the skilled performer excels.
At the highest evolutionary level of development, the human brain can think
about its own operations. This is the home of reflection, of conscious thought, of
the learning of new concepts and generalizations of the world. Sure, dogs can
learn to do lots of actions, but they can’t think about them and come up with
general knowledge in the way a person can.
The behavioral level is not conscious, which is why you can successfully drive
your automobile subconsciously at the behavioral level while consciously thinking
of something else at the reflective level. Skilled performers make use of this
facility. Thus, skilled piano players can let their fingers play automatically while
they reflect upon the higher-order structure of the music. This is why they can
hold conversations while playing and why performers report instances of losing
their place in the music and having to listen to their playing until they recognized
the part: it was the reflective level that was lost, but the behavioral level did just
fine.
Don Norman: Chapter 1: Emotional Design 5 3/23/2003:
Figure 1.2 People pay money to get scared. The roller coaster pits one level of
affect – the visceral sense of fear – against another level – the reflective pride of
accomplishment.
Now let’s look at some examples of these three levels in action; riding a roller
coaster; cutting food for cooking with a sharp, well balanced knife, a good cutting
board, and the act of dicing; and contemplating a serious work of literature or art.
These three activities impact us in different ways. The first is the most primitive,
the visceral reaction to falling, excessive speed, and heights. The second, the
pleasure of using a good tool effectively, refers to the feelings accompanying
skilled accomplishment, and derives from the behavioral level. This is the
pleasure any expert feels when doing something well, such as driving a difficult
course, playing a piece of music, or reciting a poem or joke to an appreciative
audience. This behavioral pleasure, in turn, is different from that provided by
serious literature or art, whose enjoyment derives from the reflective level, and
requires study and interpretation.
Most interesting of all is when one level plays off of another, as in the roller
coaster. If the roller coaster is so frightening, why is it so popular? There are at
least two reasons. First, some people seem to love fear itself: they enjoy the
high arousal and increased adrenaline rush that accompanies danger. The
second reason comes from the feelings that follow the ride: the pride in
conquering fear and of being able to brag about it to others. In both cases, the
visceral angst competes with the reflective pleasure – not always successfully, for
many people refuse to go on those rides or, having done it once, refuse to do it
again. But this adds to the pleasure of those who do go on the ride: their self
image is enhanced because they have dared do an action that others fear.
Don Norman: Chapter 1: Emotional Design 6 3/23/2003:
FOCUS AND CREATIVITY
The three levels do more than simply determine what we find attractive or not,
they also affect the very way the brain works. This works in both a bottom-up and
a top-down manner. The terms “bottom-up” and “top-down” come from the
standard way of showing the processing structures of the brain, with the bottom
layers associated with interpreting sensory inputs to the body and the top layers
associated with higher thought processes, much as I did in Figure 1.1. Bottom-up
processes are those driven by perception whereas top-down are driven by
thought. The brain changes its manner of operation when bathed in the liquid
chemicals called neurotransmitters. A neurotransmitter does what its name
implies: it changes how neurons transmit neural impulses from one nerve cell to
another (that is, across synapses). Some neurotransmitters enhance
transmission, some inhibit it. See, hear, feel or otherwise sense the environment,
and the affective system passes judgment, alerting other centers in the brain, and
releasing neurotransmitters appropriate to the affective state. That’s bottom-up
activation. Think something at the reflective level and the thoughts are
transmitted to the affective system which, in turn, triggers neurotransmitters.
The result is that everything you do has both a cognitive and an affective
component – cognitive to assign meaning, affective to assign value. You cannot
escape affect: it is always there. More important, the affective state, whether
positive or negative affect, changes how we think.
When you are in a state of negative affect, feeling anxious or endangered, the
neurotransmitters focus the brain processing. Focus refers to the ability to
concentrate upon a topic, without distraction, and then to go deeper and deeper
into the topic until some resolution is reached. Focus also implies concentration
upon the details. It is very important for survival, which is where negative affect
plays a major role. Whenever your brain detects something that might be
dangerous, whether through visceral or reflective processing, your affective
system acts to tense muscles in preparation for action and to alert behavioral and
reflective level to stop and concentrate upon the problem. The neurotransmitters
bias the brain to focus upon the problem and avoid distractions. This is just what
you need to do in order to deal with danger.
When you are in a state of positive affect, the very opposite actions take place.
Now, neurotransmitters broaden the brain processing, the muscles can relax, and
the brain attends to the opportunities offered by the positive affect. The
broadening means that you are now far less focused, far more likely to be
receptive to interruptions, and to attending to any novel idea or event. Positive
affect arouses curiosity, engages creativity, and makes the brain into an effective
learning organism. With positive affect, you are more likely to see the forest than
the trees, to prefer the big picture and not to concentrate upon details. On the
other hand, when you are sad or anxious, feeling negative affect, you are more
likely to see the trees before the forest, the details before the big picture.
What role do these states have in design? First, someone who is relaxed, happy,
in a pleasant mood, is more creative, more able to overlook and cope with minor
problems with a device – especially if it’s fun to work with. Recall the reviewer of
the Mini Cooper automobile, quoted in the prologue, who recommended that the
car’s faults be ignored because it was so much fun. Second, when people are
anxious, they are more focused, so where this is likely to be the case, the
designer must pay special attention to ensure that all the information required to
do the task is continually at hand, readily visible, with clear and unambiguous
feedback about the operations that the device is performing. Designers can get
away with more if the product is fun and enjoyable. Things intended to be used
Don Norman: Chapter 1: Emotional Design 7 3/23/2003:
under stressful situations require a lot more care, with much more attention to
detail.
One interesting effect of the differences in thought processes by the two states is
its impact upon the design process itself. Design – and for that matter, most
problem solving – requires creative thinking followed by a considerable period of
concentrated, focused effort. In the first case, creativity, it is good for the designer
to be relaxed, in a good mood. Thus, in brainstorming sessions, it is common to
warm up by telling jokes and playing games. No criticism is allowed because it
would raise the level of anxiety among the participants. Good brainstorming and
unusual, creative thinking require the relaxed state induced by positive affect.
Once the creative stage is completed, the ideas that have been generated have
to be transformed into real products. Now the design team must exert
considerable attention to detail. Here, focus is essential. One way to do this is
through deadlines just slightly shorter than feel comfortable. Here is the time for
the concentrated focus that negative affect produces. This is one reason people
often impose artificial deadlines on themselves, and then announce those
deadlines to others so as to make them real. Their anxiety helps them get the
work done.
It is tricky to design things that must accommodate both creative thinking and
focus. Suppose the design task is to build a control room for operators of a plant
-- think of a nuclear power plant or a large chemical-processing plant, but the
same lessons apply to many manufacturing and production facilities. The design
is meant to enhance some critical procedure or function -- say to enable control
room operators to watch over a plant and solve problems as they arise -- so it is
probably best to have a neutral or a slightly negative affect to keep people
aroused and focused. This calls for an attractive, pleasant environment so that in
normal monitoring, the operators are creative and open to explore new situations.
Once some plant parameter approaches a dangerous level, however, then the
design should change its stance, yielding a negative affect that will keep the
operators focused upon the task at hand.
How do you design something so that it can change from invoking a positive
affect to invoking a negative one? There are several ways. One is through the
use of sound. The visual appearance of the plant can be positive and enjoyable.
During normal operation, it is even possible to play light background music,
unless the control room is located where the sounds of the plant operating can
be used to indicate its state. But as soon as any problem exists, the music should
go away and alarms should start to sound. Buzzing, ringing alarms are negative
and anxiety producing, so their presence alone might do the trick. Indeed, the
problem is not to overdo it: too much anxiety produces a phenomenon known as
“tunnel vision”: the people become so focused that may fail to see otherwise
obvious alternatives.
The dangers of too much focus are well known to people who study accidents.
Thus, special design and training is required of people if we want them to
perform well under high stress. Basically, because of the extreme focus and
tunnel vision induced by high anxiety, the situation has to be designed to
minimize the need for creative thought. That’s why professionals are trained over
and over again in accident scenarios, through training exercises and simulators,
so that if a real incident occurs, they will have experienced it so many times in
training that their responses follow automatically. But this training works only if
the training is repeated frequently and performance is tested. In commercial
aviation, the pilots and crew are well trained, but the passengers are not. Even
though frequent fliers continually hear and see the instructions on how to escape
the airplane in case of fire or crash, they sit passively, only partially-attentive.
They are not apt to remember them in an emergency.
Don Norman: Chapter 1: Emotional Design 8 3/23/2003:
“Fire,” yells someone in a theater. Immediately everyone stampedes toward the
exits. What do they do at the exit door? Push. If the door doesn’t open, they push
harder. But what if the door opens inward and must be pulled, not pushed?
Highly anxious, highly focused people are very unlikely to think of pulling.
When under high anxiety – high negative affect – people focus upon escape.
When they reach the door, they push. And when this fails, the natural response is
to push even harder. Countless people have died as a result. Now, fire laws
require what is called “panic hardware.” The doors of auditoriums have to open
outward, and they must open whenever a body is pushed against it.
Similarly, designers of exit stairways have to block any direct path from the
ground floor to those below. Otherwise, people escaping a fire head for the stairs,
go to the next floor down, the next, and the next, keeping on until the stairway
ends. Unless forced out at the ground floor, they are likely to continue all the way
into the basement – and some buildings have several levels of basements – to
end up trapped.
THE PREPARED BRAIN
Although the visceral level is the simplest and most primitive part of the brain, it is
sensitive to a very wide range of conditions. These are genetically determined,
with the conditions evolving slowly over the time course of evolution. They all
share one property, however: the condition can be recognized simply by the
sensory information: the visceral level is incapable of reasoning, of comparing a
situation with past history. It works by what cognitive scientists call “pattern
matching.” What are people genetically programmed for? Those situations and
objects that, throughout evolutionary history, offer food, warmth, or protection
give rise to positive affect. These conditions include:
warm, comfortably lit places,
temperate climate,
sweet tastes and smells,
bright, highly saturated hues,
“soothing” sounds and simple melodies and rhythms,
harmonious music and sounds,
caresses,
smiling faces,
rhythmic beats,
“attractive” people,
symmetrical objects,
rounded, smooth objects
“sensuous” feelings, sounds, and shapes.
Similarly, here are some of the conditions that appear to produce automatic
negative affect:
heights,
sudden, unexpected loud sounds or bright lights,
“looming” objects (objects that appear to be about to hit the observer),
extreme hot or cold,
darkness,
extremely bright lights or loud sounds,
empty, flat terrain (deserts),
crowded dense terrain (jungles or forests),
crowds of people,
Don Norman: Chapter 1: Emotional Design 9 3/23/2003:
rotting smells, decaying foods
bitter tastes,
sharp objects,
harsh, abrupt sounds,
grating and discordant sounds,
misshapen human bodies,
snakes and spiders,
human feces (and its smell),
other people’s body fluids,
vomit.
These lists are my best guess about what might be automatically programmed
into the human system. Some of the items are still under dispute, others will
probably have to be added. Some are politically incorrect in that they appear to
produce value judgments on dimensions society has deemed to be irrelevant.
The advantage of the human being over other animals is our powerful reflective
level that enables us to overcome the dictates of the visceral, pure biological
level. We can overcome our biological heritage.
Note that some biological mechanisms are only predispositions rather than full-
fledged systems. Thus, although we are predisposed to be afraid of snakes and
spiders, the actual fear is not presenting all people: it needs to be triggered
through experience. Although human language comes from the behavioral and
reflective levels, it provides a good example of how biological predispositions mix
with experience. The human brain comes ready for language: the architecture of
the brain, the way the different components are structured and interact,
constrains the very nature of language. Children do not come into the world with
language, but they do come predisposed and ready. That is the biological part.
But the particular language you learn, and the accent with which you speak it, are
determined through experience. Because the brain is prepared to learn language,
everyone does so unless they have severe neurological or physical deficits.
Moreover, the learning is automatic: we may have to go to school to learn to read
and write, but not to listen and speak: spoken language – or signing, for those
who are deaf – is natural. Although languages differ, they all follow certain
universal regularities. But once the first language has been learned, it highly
influences later language acquisition. If you have ever tried to learn a second
language beyond your teenage years, you know how different it is from learning
the first, how much harder, how reflective and conscious it seems compared to
the subconscious, relatively effortless experience of learning the first language.
Accents are the hardest thing to learn for the older language-learner, so that
people who learn a language later in life may be completely fluent in their speech,
understanding, and writing, but maintain the accent of their first language.
Tinko and losse are two words in the mythical language Elvish, invented by the
British philologist J. R. Tolkien for his trilogy, Lord of the Rings. Which means
“metal,” which “snow”?7 How could you possibly know? The surprise is that when
forced to guess, most people can get the choices right, even if they have never
read the books, never experienced the words. “Tinko” has two, hard, “plosive”
sounds – the “t” and the “k.” “Losse” has soft, liquid sounds, starting with the “l”
and continuing through the vowels and the sibilant “ss.” Note the similar pattern
in the English words where the hard “t” in “metal” contrasted with the soft sounds
of “snow.” Yes, in Elfish, “tinko” is metal and “losse” is snow.
The Elfish demonstration points out the relationship between the sounds of a
language and the meaning of words. At first glance, this sounds nonsensical –
after all, words are arbitrary – just look how difficult it is to learn the vocabulary of
a foreign language. But more and more evidence piles up linking sounds to
particular general meanings: vowels are warm and soft: feminine is the term
frequently used. Harsh sounds are, well, harsh – like the word “harsh” itself – the
Don Norman: Chapter 1: Emotional Design 10 3/23/2003:
sound of “sh” in particular. Snakes hiss and slither: and note the sibilants, the
hissing of the “s” sounds. Plosives, sounds caused when the air is stopped
briefly, then released -- explosively -- are hard, metallic – the word “masculine” is
often applied to them. The “k” of “mosquito” and the “p” in “happy” are plosive.
And, yes, there is evidence that word choices are not arbitrary: a sound
symbolism governs the development of a language.8 This is another instance
where artists, poets in this case, have long known the power of sounds to evoke
affect and emotions within the readers of – or, more accurately, listeners to –
poetry.
All these prewired mechanisms are vital to daily life and our interactions with
people and things. Accordingly, they are important for design: While designers
can use this knowledge of the brain to make designs more effective, there is no
simple set of rules. The human mind is incredibly complex, and although all
people have basically the same form of body and brain, they also have huge
individual differences.
Emotions, moods, traits, and personality are all aspects of the different ways in
which people’s minds work, especially along the affective, emotional domain.
Emotions change behavior over a relatively short term, for they are responsive to
the immediate events. Emotions last for relatively short periods – minutes or
hours. Moods are longer lasting, measured perhaps in hours or days. Traits are
very long-lasting, years or even a lifetime. And personality is the particular
collection of traits of a person that last a lifetime. But all of these are changeable
as well. We all have multiple personalities, emphasizing some traits when with
families, a different set when with friends. We all change our operating
parameters to be appropriate for the situation we are in.
Ever watch a movie with great enjoyment, then watch it a second time and
wonder what on earth you saw in it the first time? The same phenomenon occurs
in almost all aspects of life, whether in interactions with people, in a sport, a book,
or even a walk in the woods. This phenomenon can bedevil the designer who
wants to know how to design something that will appeal: one person’s appeal is
another one’s rejection. Worse, what is appealing at one moment may not be at
another.
The source of this complexity can be found in the three levels of processing. At
the visceral level, people are pretty much the same all over the world. Yes,
individuals vary, so although almost everyone is born with a fear of heights, this
fear is so extreme some people that they cannot function normally – they have
acrophobia. Yet others have only mild fear, and they can overcome it sufficiently
to do rock climbing, circus acts, or other jobs that have them working high in the
air.
The behavioral and reflective levels, however, are very sensitive to experiences,
training, and education. Cultural views have huge impact here: what one culture
finds appealing, another may not. Indeed, teenage culture seems to dislike things
solely because adult culture likes them.
So what is the designer to do? In part, that is the theme of the rest of the book.
But the challenges should be thought of as opportunities: designers will never
lack for things to do, for new approaches to learn.
Don Norman: Chapter 1: Emotional Design 11 3/23/2003:
REFERENCES
Ashby, F. G., Isen, A. M., & Turken, A. U. (1999). A neuropsychological theory of
positive affect and its influence on cognition. Psychological Review, 106, 529-
550.
Hinton, L., Nichols, J., & Ohala, J. J. (1994). Sound symbolism. Cambridge (UK):
Cambridge University Press.
Isen, A. M. (1993). Positive affect and decision making. In M. Lewis & J. M.
Haviland (Eds.), Handbook of emotions (pp. 261-277). New York: Guilford.
Kurosu, M., & Kashimura, K. (1995, May 7-11). Apparent usability vs. inherent
usability: experimental analysis on the determinants of the apparent usability.
Denver, Colorado. Conference companion on Human factors in computing
systems. 292-293.
Norman, D. A., Ortony, A., & Russell, D. M. (2003). Affect and machine design:
Lessons for the development of autonomous machines. IBM Systems Journal,
42 (1), 38-44. http://www.research.ibm.com/journal/sj/421/norman.pdf
Ortony, A., Norman, D. A., & Revelle, W. (In progress). Effective functioning: A
three level model of affect, behavior, and cognition. In J.-M. Fellous & M. A.
Arbib (Eds.), Who Needs Emotions? The Brain Meets the Machine. New York:
Oxford University Press.
Read, H. E. (1953). Art and industry, the principles of industrial design (3rd. ed.).
London: Faber and Faber.
Tolkien, J. R. R. (1954a). The fellowship of the ring : being the first part of The
lord of the rings (Vol. pt. 1). London: George Allen & Unwin.
Tolkien, J. R. R. (1954b). The lord of the rings. London: Allen & Unwin.
Tolkien, J. R. R. (1954c). The two towers : being the second part of The Lord of
the rings (Vol. pt. 2). London: G. Allen & Unwin.
Tolkien, J. R. R. (1956). The return of the king : being the third part of The lord of
the rings (Vol. v. 3). Boston: Houghton Mifflin.
Tractinsky, N. (1997). Aesthetics and Apparent Usability: Empirically Assessing
Cultural and Methodological Issues. CHI 97 Electronic Publications: Papers
http://www.acm.org/sigchi/chi97/proceedings/paper/nt.htm
Tractinsky, N., Adi, S.-K., & Ikar, D. (2000). What is Beautiful is Usable.
Interacting with Computers, 13 (2), 127-145.
E
NDNOTES
1 “two Japanese researchers, Masaaki Kurosu and Kaori Kashimura” (Kurosu &
Kashimura, 1995)
2 “Japanese culture is known for its aesthetic tradition,” (Tractinsky, 1997)
Don Norman: Chapter 1: Emotional Design 12 3/23/2003:
3 “So Tractinsky redid the experiment.” (Tractinsky, 1997; Tractinsky, Adi, & Ikar,
2000)
4 It requires a somewhat mystical theory.” (Read, 1953, p. 61.)
5 “The psychologist Alice Isen and her colleagues” (Ashby, Isen, & Turken, 1999;
Isen, 1993)
6 “My studies of emotion, conducted with my colleagues.” (Ortony, Norman, &
Revelle, In progress)
7 “two words in the mythical language Elvish” Tolkien’s books are, of course, well
known .The demonstration, that given ten novel words in Elfish, (Tolkien, 1954a,
b, c, 1956). This particular experiment was done in my classroom by Dan
Halstead and Gitte Waldman (in 2002) for introduced me to the sound symbolism
of Tolkien and, in a class demonstration, showed that people who had never
heard Elfish could still reliably determine the meaning of its words.
8 “there is a sound symbolism.” (Hinton, Nichols, & Ohala, 1994)
CHI 97: Aesthetics and Apparent Usability: Empirically Assessing Cultural and Methodological Issues
CHI 97 Electronic Publications: Papers
Aesthetics and Apparent Usability: Empirically
Assessing Cultural and Methodological Issues
Noam Tractinsky
Industrial Engineering and Management
Ben Gurion University of the Negev
P.O. Box 653, Beer Sheva, 84105, Israel
+972 (7) 647 2226
noamt@bgumail.bgu.ac.il
"... it requires a somewhat mystical theory of aesthetics to find any necessary connection
between beauty and function."
Herbert Read, Art and Industry, p.61
ABSTRACT
Three experiments were conducted to validate and replicate, in a different cultural setting, the results of a
study by Kurosu and Kashimura [12] concerning the relationships between users' perceptions of interface
aesthetics and usability. The results support the basic findings by Kurosu and Kashimura. Very high
correlations were found between perceived aesthetics of the interface and a priori perceived ease of use of the
system. Differences of magnitude between correlations obtained in Japan and in Israel suggest the existence
of cross-cultural differences, but these were not in the hypothesized direction.
Keywords
Aesthetics, apparent usability, system acceptability, cross-cultural HCI, human-computer interface.
© Copyright ACM 1997
ABSTRACT
Keywords
INTRODUCTION
AESTHETICS AND USABILITY
http://www.sigchi.org/chi97/proceedings/paper/nt.htm (1 sur 12)27/11/2007 17:28:40
CHI 97: Aesthetics and Apparent Usability: Empirically Assessing Cultural and Methodological Issues
PURPOSE OF THE STUDY
METHODOLOGY
Study Materials
Variables
THE EXPERIMENTS
Experiment 1: Replicating KK's study in Israel
Results
Experiment 2: Testing for potential response dependency
Results
Experiment 3: Testing for medium bias
Procedure
Results
DISCUSSION
CONCLUSION
ACKNOWLEDGMENTS
REFERENCES
INTRODUCTION
The quest for usable computer technology is one of the major goals in the field of Human-Computer
Interaction (HCI), and critiquing the poor usability of systems designed for human use has become the bon
ton in the field (e.g. [14]). Shackel [27] and Nielsen [21] place the concept of usability within the framework
of system acceptability: Together with other system attributes, such as cost, utility (functionality) and
likeability, system usability determines whether people will accept the use (or purchase) of a computerized
system. With issues of cost determined by technological and economic factors, and utility predominated by
specific tasks, HCI researchers have concentrated mainly on the study of usability. Here, knowledge of
human cognitive and perceptual capabilities and limitations has provided fertile ground for formulating
principles (e.g., [22 ] ) and guidelines (e.g., [28 , 21]) of usable systems. Nielsen [21]defines the usability of a
computer system in terms of five attributes: Learnability, efficiency, memorability, errors, and satisfaction. In
general, the evaluation of system usability requires that these attributes be measured during or after people
have actually used the system. Thus, while designers might rely on principles and guidelines to design usable
systems, the use of certain inspection methods, if not extensive testing, is required to establish a certain
degree of usability. On the other hand, evaluating other determinants of system acceptability may not require
such a lengthy process. For example, system cost or likeability can be evaluated relatively simply and
quickly. Thus, much effort might be invested by HCI designers in their attempts to evaluate and improve
usability whereas, at the same time, other acceptability attributes may have an overriding effect on purchase
or usage decisions. Whether these attributes are not considered to be an integral part of the HCI field, or
because of an implicit assumption of orthogonality between these attributes and usability, the relationships
between system usability and other determinants of system acceptability have not received extensive
treatment in the HCI literature. This study calls attention to these relationships and to their potential
contribution to our evaluation of system acceptability and to our understanding of its antecedents.
AESTHETICS AND USABILITY
http://www.sigchi.org/chi97/proceedings/paper/nt.htm (2 sur 12)27/11/2007 17:28:40
CHI 97: Aesthetics and Apparent Usability: Empirically Assessing Cultural and Methodological Issues
The role of aesthetics in human affairs has been widely documented (e.g., [17]). Conventional wisdom relates
it to our appreciation of, and attitudes towards computer systems as well. However, aesthetics may not always
coincide with usability. In fact, the opposite might occur. In one of HCI's most influential books, "The
Psychology of Everyday Things", Norman vividly ridicules the tendency of designers to neglect usability in
favor of aesthetics [21] (pp. 151-155). Similarly, others (e.g. [7, 18]), while acknowledging the role of
aesthetics in HCI, warn against a tendency among designers to emphasize the aesthetic elements of the user
interface, because these might degrade usability. The contribution of aesthetics to HCI, they argue, should be
measured in terms of facilitating information processing, not in terms of engaging the user in a pleasing
experience. Perhaps, because aesthetics mainly reflect on the latter, HCI literature in general, and on usability
in particular, mostly seem to neglect the aesthetics issue completely. For example, the indices of 4 prominent
HCI textbooks and reading collections [1, 4, 24, 28] do not contain a single entry for "aesthetics" (or
synonyms and related concepts such as “appearance”, “attractiveness”, “beauty” or “form”). Thus, it would
appear that mainstream HCI (but, of course, see Laurel [15,16] for a notable exception) either belittles the
importance of aesthetics or ignores it altogether.
A recent study by Kurosu and Kashimura [12] hints that interface aesthetics may play a greater role in
people's attitudes towards computerized systems than we might be willing to admit. In their study, Kurosu
and Kashimura (KK) explored the relationships between a priori perceptions of the ease of use of an
automatic teller machine (ATM)—which they termed "apparent usability"—and other variables. These
included factors believed by HCI professionals to enhance usability (termed "inherent usability"), and the
appearance (beauty) of the interface. Surprisingly, high relationships were found between the interface
judgments of aesthetics and apparent usability (r = 0.59). The correlations between the apparent usability and
inherent usability factors were mostly negligible, with the exception of one variable (familiarity with the
numeric keypad). It can be argued that KK found close relationships between aesthetics and perceived
usability before the actual use, whereas usability should actually be measured during or after system use.
While this argument is valid, it should be noted that first impressions often influence attitude formation to a
large extent (e.g., [3]). There is no reason to assume that this process of attitude formation does not pertain to
the HCI domain. In fact, in a study of information systems use, researchers found that "if computers were
perceived initially as difficult to use, users were more likely to express dissatisfaction with the interface of
the system after four months of use." [10] (p. 752, italics added). Thus, it is possible that among the various
factors that affect system usability in particular and system acceptability in general, interface aesthetics play a
major role. Aesthetics affect people's perceptions of apparent usability—which, in turn, may influence longer
term attitudes towards the system.
PURPOSE OF THE STUDY
The main motivation for this study is twofold. First, Kurosu and Kashimura's findings are somewhat
unexpected. If these results are robust, then the importance of aesthetics in HCI should rise considerably,
given the relationships between interface aesthetics, initial perception of usability, and later attitudes towards
computers. However, KK's study was not void of a potential method bias that might have shifted the results
in favor of stronger relations between apparent usability and aesthetics. Some modifications to their design
are needed to assess the robustness of their findings.
Second, living in a culture that does not seem to value aesthetics as much as do the Japanese, the author of
this study was particularly surprised by the high correlations between apparent usability and aesthetics.
Clearly, aesthetic perceptions are culturally dependent [6, 17]. Thus, one can reasonably expect the
relationships between aesthetics and apparent usability to vary across cultures. For example, whereas
Japanese culture is known for its aesthetic tradition (e.g., [6]), Israeli culture is probably better known for its
http://www.sigchi.org/chi97/proceedings/paper/nt.htm (3 sur 12)27/11/2007 17:28:40
CHI 97: Aesthetics and Apparent Usability: Empirically Assessing Cultural and Methodological Issues
action orientation [8, 11]. Unfortunately, there is no scientific literature that assesses Israeli aestheticism, so
mere intuition and shared feelings among Israeli colleagues were used in proposing that: Japan and Israel
potentially represent two different attitudes towards the importance of aesthetics in computerized systems and
its relationships to usability and overall acceptability.
METHODOLOGY
Three experiments were designed and conducted in Israel to test the robustness of Kurosu and Kashimura's
findings to cultural and methodological bias. Experiment 1 tested the cross-cultural robustness of KK's
findings. Experiment 2 tested for possible response dependency bias in KK's experimental procedure, and
Experiment 3 tested for potential medium bias. The unique aspects of each experiment and its main results
are described in the next section.
Study Materials
The original study materials (26 ATM layouts in Japanese) were provided to the author by Kurosu and
Kashimura. The ATM layouts were solicited by KK from a group of 26 people: 9 GUI designers, 6 industrial
engineers, 8 engineers and 3 secretaries. All layouts included the same components and were differed only by
how these components were arranged [12]. Where necessary, the materials were translated into Hebrew and
the 26 layouts were replicated. A few problems of user interface translatability were encountered at this stage.
Trade-offs had sometimes to be made between ensuring the natural look of the interface to Israeli eyes on one
hand, and totally preserving the Japanese original version on the other hand. In most cases, the former
approach was preferred. For example: (a) Several controls in the Japanese interface are represented as one
character, whereas in Hebrew they require a whole word.
(b) The Japanese interface included an image of a lady who is presumed to bow repeatedly to indicate that the
system is processing. This concept was totally foreign to Israelis and potentially would have looked odd.
Therefore, the image was replaced with a an image of an hourglass which is a more familiar representation of
an active system in Israel.
(c) In Japan, some actions are represented by a symbol that can't be translated directly into Hebrew. For
example, the currency (Yen) symbol in the Japanese interface (one character) denotes an operation for which
Israeli ATMs use (the Hebrew word for) "confirmation". Thus, it was translated to the Hebrew equivalent of
"confirmation" rather than to the Israeli currency symbol.
(d) The Japanese material was produced using Claris Works for the Mac. To use the original Japanese
software one had to use a Kanji-aware Mac operating system. These are such rare birds in Israel, that
different software was eventually used on a different platform (Microsoft's Visual Basic).
The results of the translation process can be seen in Figures 1 and 2, which show two examples of original
Japanese layouts and their Israeli counterparts. Figure 1 presents a layout that was rated high on apparent
usability and aesthetics both in Japan and in Israel. Figure 2 presents a low rating layout. (Unfortunately, for
technical reasons the reproduction of the Japanese layouts in this paper is of lower quality.)
http://www.sigchi.org/chi97/proceedings/paper/nt.htm (4 sur 12)27/11/2007 17:28:40
CHI 97: Aesthetics and Apparent Usability: Empirically Assessing Cultural and Methodological Issues
Figure 1(a). An original Japanese interface, rated
high on apparent usability and aesthetics.
Figure 1(b). The equivalent Israeli interface,
rated high on apparent usability and aesthetics.
Figure 2(a). An original Japanese interface, rated
low on apparent usability and aesthetics.
Figure 2(b). The equivalent Israeli interface,
rated low on apparent usability and aesthetics.
Variables
In this study, seven independent, objective variables, which KK considered to be the determinants of inherent
usability of the ATM layout were used. These variables reflect design strategies that were actually used by
designers of the ATM interfaces to affect the cognitive and the operational efficiency of user interaction with
the ATM. These variables included: (1) location of the main display (DISTANCE), (2) type of numeric
keypad (KEYPAD), (3) grouping of keys according to their functions (GROUPING), (4) sequence of the
special numeric keys (SEQUENCE 1), (5) location of the numeric keypad (HAND-DOMIN), (6) location of
the "Confirm" key (SEQUENCE 2), (7) location of the "Cancel" key (SAFETY). A more detailed description
of the variables can be found in [12]. Two dependent variables, apparent usability and aesthetics, were
operationalized by asking participants to rate each design on two 1-10 scales: How easy it is to use (apparent
usability), and how beautiful it is (aesthetics).
THE EXPERIMENTS
http://www.sigchi.org/chi97/proceedings/paper/nt.htm (5 sur 12)27/11/2007 17:28:40
CHI 97: Aesthetics and Apparent Usability: Empirically Assessing Cultural and Methodological Issues
Experiment 1: Replicating KK's study in Israel
Experiment 1 was designed to test the robustness of KK's results to cultural variation. Participants were 104
first-year engineering students in an Israeli University. The procedure used was identical to the one used by
[12]: The twenty-six design layouts were displayed in a large classroom, using an overhead screen projector.
Each layout was displayed for about 20 seconds. During that time, participants rated each layout on a 1 to 10
scale regarding how usable it appeared to be, and how beautiful it was.
Results
Participant ratings were averaged to form an apparent usability and an aesthetics score for each of the 26
designs. Mean scores for all 26 designs were similar for this and KK’s study ( 5.9 vs. 6.0 for aesthetics and
5.4 vs. 5.8 for apparent usability in Israel and Japan, respectively). Scores’ variability was higher in Israel
than in Japan. Aesthetics scores ranged between 3.5 and 8.5 in Israel, compared to a range between 5.2 and
6.8 in Japan. Similarly, apparent usability scores for the 26 designs in Israel ranged between 2.7 and 8.5,
compared to 4.4 to 6.5 in Japan. Relationships with apparent usability using the coefficient of correlation for
the interval scales and the coefficient of contingency for the categorical variables are presented in Table 1,
alongside the corresponding correlations from KK. In general, the relationships resemble those obtained by
KK. However, the magnitude of correlation between aesthetics and apparent usability (0.921) was notably
higher in this experiment. A test for differences between correlations [9] found a significant difference
between this correlation and the one obtained by KK (Z = 3.09, p = 0.001). This suggests that, even more
than their Japanese counterparts, Israelis perceived ease of use and design aesthetics to be closely related.
These results are quite surprising, given the expectation that Israeli students would be less sensitive to the
aesthetic aspects of the interface.
Table 1. Correlations (bold: p<.01) and coefficients of
contingency (#) of aesthetics and seven inherent usability
variables with apparent usability for the experiment in
Japan (KK) and for the three experiments in Israel.
Among determinants of inherent usability, only keypad type and the number of grouped elements were
correlated with apparent usability. Table 2 presents a contingency table in which 4 equal interval categories
are used to rank the 26 designs according to their apparent usability rating. Despite the similarity in
magnitude of the coefficient of contingency between Israeli and Japanese participants, a closer look at the
contingency table reveals that, while the Japanese associated better usability with the horizontal keypad
layout, Israelis related it to the telephone keypad. In addition, significant correlations were found in this
experiment between apparent usability and the number of grouping elements in the ATM design. Higher
apparent usability was associated with less groups, defying conventional advice in the usability literature,
which calls for the separation of functionally unrelated controls (e.g., [18]).
http://www.sigchi.org/chi97/proceedings/paper/nt.htm (6 sur 12)27/11/2007 17:28:40
CHI 97: Aesthetics and Apparent Usability: Empirically Assessing Cultural and Methodological Issues
Table 2. Ranking of Numeric Keyboard
Types in Japan vs. Experiment 1 in Israel
Experiment 2: Testing for potential response dependency
As noted above, in both KK's study and Experiment 1, 26 different designs of an ATM interface were
displayed using an overhead projector. Participants were asked to rate each design on two consecutive 1-10
scales before advancing to the next design. Given this procedure, it is possible that the strong correlation
between apparent usability and aesthetics is an artifact of a potential dependency between the responses to
these questions. In Experiment 2, the procedure was modified to overcome this potential problem. The 26
designs were projected in two separated rounds. The order of presentation of the designs was determined
randomly for each round. A different group of eighty-one first-year engineering students participated in this
study. The students were assigned randomly to one of two conditions. In one condition, participants evaluated
the design aesthetics for all 26 designs in the first round, and the apparent usability for the 26 designs in the
second round. In the other condition, the order of evaluation was reversed. Because only one evaluation per
overhead was needed, presentation time was cut to about 15 seconds per overhead.
Results
Correlations and contingency coefficients are presented in Table 1. The results resemble those of Experiment
1, weakening the alternative explanation that the relationships between aesthetics and apparent usability are
primarily the result of a response dependency bias caused by the method used in KK and in Experiment 1.
Experiment 1's results, regarding apparent usability relationships with keypad type and number of groupings
are also replicated here. Thus, experiment 2 lends further support to the overall strong correlation between
apparent usability and aesthetics, and to the differences between Israelis and Japanese concerning these
relationships.
Experiment 3: Testing for medium bias
The two previous experiments—like KK's experiment—used an overhead projector to display the designs on
a large screen. Participants saw the designs from different distances (which were generally large), and from
different angles. The third experiment was designed to test whether judgments would differ if the ATM
designs were presented on a terminal display rather than on a large public screen. Working with personal
computers had additional advantages from the experimenter's viewpoint. It increased the uniformity of the
viewing conditions, enabled participants to work in their own pace, and allowed for full randomization of the
stimuli presentation (i.e., both the order of the questions and the designs). Participants were 108 3rd year
engineering students who participated in the study in part fulfillment of their requirements for an HCI course.
Procedure
http://www.sigchi.org/chi97/proceedings/paper/nt.htm (7 sur 12)27/11/2007 17:28:40
CHI 97: Aesthetics and Apparent Usability: Empirically Assessing Cultural and Methodological Issues
Participants were seated in front of a personal computer. A computer program, written in Visual Basic was
used to present the stimuli material, accept user responses, and measure response times. The program started
with a short on-line introduction, after which participants were presented with the 26 ATM designs. The
designs were displayed in a random order. At the bottom of the screen, one of the two questions regarding
aesthetics and apparent usability was presented. The participants answered the question on the 1-10 scale by
selecting one of ten available response buttons. To proceed to the next design, they had to select a "Continue"
button. After responding to the first question for all 26 designs, the 26 designs were presented once again (in
a newly randomized order) and the participants answered the other question (apparent usability or aesthetics)
for each design.
Results
The results from Experiment 3 are presented in Table 1. By and large, the results are consistent with those
obtained in Experiments 1 and 2, reinforcing the patterns observed thus far. The use of computers in this
experiment allowed for the measurement of the time it took participants to evaluate the designs. On the
average, people took more time to evaluate usability than aesthetics (mean evaluation times = 8.68 sec. vs.
7.58 sec. respectively; t=2.49, df=107, p=.014). Though not very large, the difference in latencies supports
the intuitive expectation that evaluating apparent usability is more complex, and hence more time consuming,
than the evaluation of the interface aesthetics.
DISCUSSION
The results of this study replicate the basic findings of Kurosu and Kashimura [12]. However, because of the
unexpectedly high correlations between apparent usability and aesthetics found in Israel, it is still premature
to rule out the possibility of method variance as a major cause of the obtained results. When traits are
expected to correlate, it is preferable to add to the study traits that are expected to be independent of each
other, in order to rule out method effects. Unfortunately, this was not done in this case, because apparent
usability and aesthetics were not expected to correlate in Israel (recall the original proposition of this study).
Some support exists, though, for the dominance of true trait correlation between apparent usability and
aesthetics. Table 3 presents the correlations obtained in the 3 experiments between these variables in a
multitrait-multimethod matrix format [2]. The resulting matrix indicates that, at the very least, the results
cannot be attributed solely to method variance. Clearly, all validity diagonals (italicized numbers) are very
large, providing evidence of convergent validity (cf. [2]). Support for discriminant validity is more tentative.
Each validity value is higher than the values lying in its column and row within the heteromethod blocks
(enclosed by broken lines). On the other hand, to establish discriminant validity correlations of independent
measures of the same trait (e.g., A1 and A2, U2 and U3) need to exceed the correlations between different
traits measured by the same method (e.g., A1 and U1, A3 and U3). This happens in only 5 out of 12
comparisons. However, most violations of this requirement occur with comparisons that involve Experiment
1 (the experiment most likely to include method bias). Comparisons involving only Experiments 2 and 3 meet
this requirements 3 out of 4 times. Thus, it appears that some method bias indeed existed in Experiment 1 and
was removed in the subsequent experiments. Another argument for a true correlation of apparent usability
and aesthetics stems from the fact that very similar results were obtained under three considerably different
experimental contexts and procedures. Thus, it can be concluded with some confidence that despite the
potential method bias, people's perceptions of apparent usability and aesthetics are quite high in general.
Unfortunately, the experiments reported above are too exploratory in nature to explain the process by which
people associate usability and aesthetics. Nevertheless, the high correlations across cultures and experimental
conditions challenge our assumptions regarding the dimensions of system acceptability in general and the
http://www.sigchi.org/chi97/proceedings/paper/nt.htm (8 sur 12)27/11/2007 17:28:40
CHI 97: Aesthetics and Apparent Usability: Empirically Assessing Cultural and Methodological Issues
relationships between aesthetics and usability in particular. The various design disciplines have long been
occupied with the fragile equilibrium between form and function, aesthetics and usability (e.g. [23, 25]). The
field of HCI has taken an unequivocal stand on this matter, concentrating on usability. The results of this
study, however, suggest that to achieve the ultimate goal of an acceptable system, a more balanced approach
may be needed. The influence of attractive appearance on attitudes and behavior has been documented by
social psychologists (e.g., [3, 19, 29]), and has been used by advertisers and persuaders of all sorts. The
potential effect of aesthetic experience has not escaped software vendors as well, nor is it ignored by the trade
literature in its evaluation of computer products. In their attempts to shift the balance back towards a more
user-oriented — rather than customer-oriented — design, it seems that HCI researchers have thus far ignored
the possible interplay between aesthetics and usability. Clearly, future research is needed to discriminate
between different concepts of usability (for example, intended-, apparent-, and measured usability) and to
evaluate the effects of aesthetics on each and on the overall acceptability of the system. The results obtained
in this study, together with the potential effect of apparent usability on post- use satisfaction [10,13], strongly
suggest that we pay more attention to people's perceptions of the interface aesthetics then we have done thus
far. In a sense, this study provides empirical support to Laurel’s [15] call for asking “not what the users are
willing to endure, but what the ideal user experience might be, and what sort of interface might provide it” (p.
69).
Table 3. A multitrait-multimethod matrix presentation of apparent usability and
aesthetics as measured in three experiments. The table is arranged according to
Campbell and Fiske [2]. Validity diagonals are marked in bold italics; heteromethod
blocks enclosed within broken lines.
The other major finding of this study suggests that the degree to which aesthetics relate to apparent usability
is culturally dependent. We hypothesized that Japanese are more sensitive to aesthetics than Israelis, and
would therefore emphasize more the role of aesthetics in interface design. Apparently, this is not the case. In
Israel, correlations between apparent usability and aesthetics were considerably higher than in Japan.
Apparently, our knowledge of how culture-specific aesthetics affect HCI issues is still limited. Possibly, with
greater aesthetic sensitivity come also greater sophistication and critical skills which perhaps allow the
Japanese in KK's study to not completely associate aesthetics with apparent usability as did their Israeli
counterparts. An alternative explanation can be based on the larger variability in scores exhibited by Israeli
students. The tendency by Israelis to provide more extreme evaluations could be attributed to their tendency
http://www.sigchi.org/chi97/proceedings/paper/nt.htm (9 sur 12)27/11/2007 17:28:40
CHI 97: Aesthetics and Apparent Usability: Empirically Assessing Cultural and Methodological Issues
to freely express opinions and preferences [11]. Another possible explanation, suggested by an anonymous
reviewer, relates the larger variance among Israeli participants to their more diverse cultural background.
Thus, the relatively unrestricted range of ratings provided by the Israelis for both variables may have led to
higher correlations between aesthetics and apparent usability. Clearly, relating aesthetics to HCI in general,
and to cross-cultural issues in particular should benefit from a more disciplined approach than has been taken
thus far. The HCI literature on cross-cultural issues is scarce and limited [20, 26]. There is no theory of cross-
cultural HCI, and recommendations are not based on thorough empirical investigations. Consequently, its
usefulness is quite limited. For example, Graphical User Interfaces (GUIs), once considered a panacea for
interface globalization, now appear to be just as culturally specific and limited as older interfaces. And
attempts to apply rules of thumb for GUIs aimed at culturally diverse users have failed when tested
empirically, as demonstrated effectively by Teasley et al. [31]. Thus, while cultural aesthetics might serve as
a good starting point for a cross-cultural research agenda in HCI, developing a cross-cultural theory of HCI
requires us to apply more rigorous research methods and to empirically test our hunches and conventional
wisdom. Future research should also focus more on identifying and measuring relevant aesthetic components
(cf. [5]) that might help explain our experience with, and evaluation of, computer aesthetics.
CONCLUSION
This study was designed with the prospect of demonstrating that high correlations between aesthetics and
apparent usability are culture specific. It was expected that the correlations in Israel would be lower than
those obtained in Japan. Surprisingly, the results indicated the opposite. This leads to three major
conclusions: First, aesthetic perception and its relations to HCI relevant constructs are culturally dependent.
Second, our current knowledge limits our ability to accurately predict how culture influences HCI related
issues. Third, the results provide further support for the contention that perceptions of interface aesthetic are
closely related to apparent usability and thus increase the likelihood that aesthetics may considerably affect
system acceptability. The first two conclusions call for future efforts in defining the areas in which culture
interacts with the domain of HCI and in systematically studying these areas. The third conclusion postulates
that objective measures of system behavior and use may not suffice in predicting system acceptability.
Perhaps a more holistic approach towards understanding how people experience and judge information
systems is needed [30].
ACKNOWLEDGMENTS
I am indebted to Masaaki Kurosu for his help throughout this study. Thanks to Sheizaf Rafaeli for a good
advice and to Marina Azarova and Tania Bernshtein for their help in conducting the experiments and
analyzing the data.
REFERENCES
1. Baecker, R.M, Grudin, J., Buxton, W.A.S., and Greenberg, S. (eds.)Readings in Human-Computer
Interaction: Toward the Year 2000 (2nd ed.). Morgan Kaufman, 1995.
2. Campbell, D.T. and Fiske, D.W. Convergence and discriminant validation by the multitrait-
multimethod matrix. Psychological Bulletin, 56,2 (1959), 81-105.
3. Chaiken, S. Communicator physical attractiveness and persuasion. Journal of Personality and Social
Psychology, 37 (1979), 1387-1397.
4. Dix, A., Finlay, J., Abowd, G., and Beale, R. Human-Computer Interaction, Prentice Hall, 1993.
5. Eysenck, H.J. Visual aesthetic sensitivity and its measurement. In Ross, M. (ed.) The Arts: A Way of
http://www.sigchi.org/chi97/proceedings/paper/nt.htm (10 sur 12)27/11/2007 17:28:40
CHI 97: Aesthetics and Apparent Usability: Empirically Assessing Cultural and Methodological Issues
Knowing. Pergamon Press , 1983.
6. Fernandes, T. Global Interface Design. AP Professional, 1995.
7. Foley, J.D., van Dam, A., Feiner, S.K., and Hughes, J.F. Computer Graphics: Principles and Practice
(2nd ed.). Addison-Wesley, 1990.
8. Goodman, S. and Ariav, G. Israel: Of swords and software plowshares, Communications of the ACM,
37,6 (1994), 17-21.
9. Guilford, J.P. and Pruchter, B. Fundamental Statistics in Psychology and Education (6th ed.) McGraw-
Hill, 1978.
10. Hiltz, S.R. and Johnson, K. User satisfaction with computer mediated communication systems,
Management Science, 30,6 (1990), 739-764.
11. Katriel, T. Talking Straight: Dugri Speech in Israeli Sabra Culture. Cambridge University Press, 1986.
12. Kurosu, M. and Kashimura, K. Apparent usability vs. Inherent usability, CHI '95 Conference
Companion, (1995), 292-293.
13. Kurosu, M. and Kashimura, K. Determinants of the Apparent Usability, Proceedings of IEEE SMC,
(1995), 1509- 1513.
14. Landauer, T.K. The Trouble with Computers. MIT Press, 1994.
15. Laurel, B.K. Interface as Mimesis. In Norman, D.A. and Draper, S.W. (eds.) User Centered System
Design. Lawrence Erlbaum, 1986.
16. Laurel, B. Computers as Theatre. Addison-Wesley, 1991.
17. Maquet, J. The Aesthetic Experience. Yale University Press, 1986.
18. Marcus, A. Graphic Design for Electronic Documents And User Interfaces. ACM Press, 1992.
19. Mills , J. and Aronson, E. Opinion change as a function of communicator's attractiveness and desire to
influence. Journal of Personality and Social Psychology, 1 (1965), 173-177.
20. Nielsen, J. Designing user interfaces for international use. Elsevier, Amsterdam, 1990.
21. Nielsen, J. Usability Engineering. AP Professional, 1993.
22. Norman, D.A. The Psychology of Everyday Things. Basic Books, 1988.
23. Petroski, H. The Evolution of Useful Things. Knopf, New York, 1993.
24. Preece, J., Rogers, Y., Sharp, H., Benyon, D., Holland, S., and Carey, T. Human-Computer
Interaction. Addison- Wesley, 1994.
25. Read, H. Art and Industry (3rd ed.). Faber and Faber, London, 1953.
26. Russo, P., and Boor, S. How fluent is your interface? Designing for international users. INTERCHI
'93 Conference Proceedings, (1993), 342-347.
27. Shackel, B. Usability—Context, framework, definition, design and evaluation. In Shackel, B. and
Richardson, S. (eds.) Human Factors for Informatics Usability. Cambridge University Press, 1991.
28. Shneiderman, B. Designing the User Interface (2nd ed.). Addison Wesley, 1992.
29. Stewart, J.E. Defendant's attractiveness as a factor in the outcome of trials. Journal of Applied Social
Psychology, 10 (1980), 48-361.
30. Stolterman, E. The aesthetics of information systems. Unpublished manuscript, Institute of
Information Processing, University of Umea, Sweden ,1995.
31. Teasley, B. et al. Cultural diversity in user interface design: Are intuitions enough? SIGCHI Bulletin,
January 1994, 36-40.
Aesthetics :In the context of this study, we adopt a narrow definition of aesthetics in which aesthetics can be
seen as a synonym for visual beauty.
http://www.sigchi.org/chi97/proceedings/paper/nt.htm (11 sur 12)27/11/2007 17:28:40
CHI 97: Aesthetics and Apparent Usability: Empirically Assessing Cultural and Methodological Issues
CHI 97 Electronic Publications: Papers
http://www.sigchi.org/chi97/proceedings/paper/nt.htm (12 sur 12)27/11/2007 17:28:40
... Aesthetic experience involves one or multiple sensory modalities in order to delight during interaction. Aesthetic aspect of an experience includes not only visual of a product, but also other attributes such as sound and tactual [16]. This aesthetic approaches may transform a product into an object of desire by creating impression s in the first contact [17]. ...
... usability [16]. For example, feeling frustrated from using complex product or disappointed with a small battery capacity of a smartphone. ...
... A product may fail if it is easy to use and easy to understand but has deficient technology that has limited ability and obsolescent, or the product has tremendous technology but is too difficult to use [6]. Study conducted by Dianne Cyr et al. demonstrated that aesthetic elements of the product do in fact have a significant impact on perceived usefulness, ease of use and enjoyment [16,29]. ...
Article
Full-text available
User experience, known as the users' cognition responses to the look and feel of a product, has been evolved and recognized as one of the prominent aspects in product design and development. Theoretically, the studies of user experience have been expanding, adapting from psychological research to design research. user experience, a concept which places users as a central focus, has been used for creating products that are provide aesthetically pleasing and pleasurable to use as it focuses beyond usability and traditional design strategies. However, since the concept of user experience is still broad and arguably unclear in definition, there are a few numbers of practical researches of user experience that stress the benefit of this emphasizes in different directions, such as creation of product differentiation by user experience, brand building and longevity of a product. This study provides an understanding of user experience concept and its relationship with product design by product specific example and how roles of user experience are generated. In addition, framework and model to explain the stages of user experience in product interaction is proposed. This study identifies three main roles and contributions of user experience in product design: user experience as product differentiator, decoding of complex technology and product longevity. Finally, a framework to explain the mechanism of user experience stages during an interaction with a product is proposed. This research expects to elicit the design researchers to further explore and investigate this topic in order to compels and expand the benefit of user experience in another direction. Particular design research effort is needed for developing tools and methods applicable in design practice that would enable designers to create differentiations of their products based on user experience approach.
... First impression of a product is a key to create interaction with the product to be more engaging or immersive. Aesthetic impression that Norman defined as visceral, can lead into product engagement [13]. Interaction which is created from first impression could make user interested or engaged and lead to the continuous usage of the product [14]. ...
... This indicates that respondents tend to anticipate the car they want to buy by considering the aesthetic aspects of the car, and how the car may suit their style and personality. We can say that the more a person impressed with particular aesthetic attributes of car, their tendency to create further interaction with the car will be higher [13,14]. In addition, respondents build their anticipation not only through the car's aesthetic appeal its impression on durability since a car is considered as a product with extortionate price. ...
Article
Full-text available
In user-product interaction, there are different stages of user experiences which are constructed from various sources during interaction. These stages are occurred before, during, and after the usage of a product. This research aims to identify the dominant, instrumental, symbolic, and affective factors that influences in three different stage of experience in car usage and how the experience is shifting from before usage, during usage, and after usage. This paper reports the results of a questionnaire study aimed towards different kinds of user's experience responses during interaction with a car in three different conditions of interaction stages. We found that there are significant relationships between product attachment and the length of immersive interaction which mainly created by three of five experience attributes. This research is expected to help designers to identify which important factors for user experience and as well as to foster greater user experience within their product.
Article
Full-text available
Computation is in the phase of its jet revolution where most users want best service performance algorithm. A user paid for a service, it is important to him that the performance of the service paid for is as expected. In today's world, most users tend to employ a zero-training system design algorithm in executing their functions, but some of these zero-training systems may not be easy to come by due to wrong or poor user interface design algorithm (UIDA) employed by concerned programmer(s). This has to do with the view of the analyst, the designer and the programmer in relation to ergonomic rules as a body. In regards to this, every aspect of computation has to concern itself with ergonomic even in cloud computing environment as well. This paper "Zero-Training User Interface Design Algorithm using Ergonomics Rule in Cloud Computing Environment" thus, will bring to light the importance of total understanding of the system analyst; the designer and the programmer in bringing out the best user interface design as specified in a given user requirements using ergonomic rules, which can be used without stress at the end of system implementation by the concerned users in cloud computing environment, and any other application alike. Not just that, the service call and response time algorithm should also be considerably fast to avoid much wait time on the side of the end users.
Article
Full-text available
Computation is in the phase of its jet revolution where most users want best service performance algorithm. A user paid for a service, it is important to him that the performance of the service paid for is as expected. In today's world, most users tend to employ a zero-training system design algorithm in executing their functions, but some of these zero-training systems may not be easy to come by due to wrong or poor user interface design algorithm (UIDA) employed by concerned programmer(s). This has to do with the view of the analyst, the designer and the programmer in relation to ergonomic rules as a body. In regards to this, every aspect of computation has to concern itself with ergonomic even in cloud computing environment as well. This paper "Zero-Training User Interface Design Algorithm using Ergonomics Rule in Cloud Computing Environment" thus, will bring to light the importance of total understanding of the system analyst; the designer and the programmer in bringing out the best user interface design as specified in a given user requirements using ergonomic rules, which can be used without stress at the end of system implementation by the concerned users in cloud computing environment, and any other application alike. Not just that, the service call and response time algorithm should also be considerably fast to avoid much wait time on the side of the end users.
Article
Full-text available
Computation is in the phase of its jet revolution where most users want best service performance algorithm. A user paid for a service, it is important to him that the performance of the service paid for is as expected. In today's world, most users tend to employ a zero-training system design algorithm in executing their functions, but some of these zero-training systems may not be easy to come by due to wrong or poor user interface design algorithm (UIDA) employed by concerned programmer(s). This has to do with the view of the analyst, the designer and the programmer in relation to ergonomic rules as a body. In regards to this, every aspect of computation has to concern itself with ergonomic even in cloud computing environment as well. This paper "Zero-Training User Interface Design Algorithm using Ergonomics Rule in Cloud Computing Environment" thus, will bring to light the importance of total understanding of the system analyst; the designer and the programmer in bringing out the best user interface design as specified in a given user requirements using ergonomic rules, which can be used without stress at the end of system implementation by the concerned users in cloud computing environment, and any other application alike. Not just that, the service call and response time algorithm should also be considerably fast to avoid much wait time on the side of the end users.
Chapter
As increasingly complex in-vehicle infotainment systems (IVIS) are available in the automotive marketplace, there is a growing tendency to move from usability to user experience, the latter emphasizing more strongly the importance of users’ emotions during system design. Emotions have long been recognized as cognitive processes that affect decision making, perception, and attention. Recent literature reveals that design based on emotions can influence the overall user experience of interacting with IVIS. Therefore, this paper presents an exploratory co-design study aimed at identifying the emotional needs of car occupants (particularly for passengers inside the vehicle) and potential directions for the design of future passenger targeted IVIS. 16 participants were invited to participate in the co-design process, which mainly included the sessions of brainstorming and concept generation. The results showed that emotional design plays a vital role in user engagement with IVIS. In line with Norman’s emotional design model, a thematic network was constructed to illustrate how the visceral, behavioral, and reflective levels of mental processing can be supported and enhanced by the appropriate emotional design of passenger targeted IVIS.
Chapter
An end-use’s experience of any software is typically influenced by the interface presented by the application to the user. For Mixed Reality Environments such as Augmented and Virtual Reality, the user interface is highly visual, and a poor interface can significantly degrade the user experience. Adequate attention is required when designing or creating interfaces and user experience within Mixed Reality Environments as traditional interface design goals and specifications often need to be adjusted. Furthermore, for mixed reality environments on Mobile devices there are additional interface constraints and considerations that would considerably improve the user experience when properly addressed. This research paper discusses the evolution(s) of user interface(s) and user experience of Augmented and Virtual Reality applications on Mobile devices and contributes a framework for improving user interfaces and experience when using Mixed Reality Environments.
Article
Full-text available
For the purpose of comparing the concept of the apparent usability with the concept of the intended inherent usability and the experienced inherent usability, two psychological experiments were conducted to see how much the apparent usability is related to the inherent usability. In both experiments, two kinds of usability concepts showed low correlations with each other. The result indicated that the apparent usability is a different concept from the inherent usability and is rather related to the visual impression of the interface. This result also means that the inherent usability is difficult to be understood by just looking at the interface and suggests that the interface designers have to make efforts to make the interface 'look' usable as well as to make it 'actually' usable.
Book
The idea that one day robots may have emotions has captured the imagination of many and has been dramatized by robots and androids in many famous movies. This book tackles the issue of whether robots can have emotions from a purely scientific point of view. The study of the brain now usefully informs study of the social, communicative, adaptive, regulatory, and experiential aspects of emotion and offers support for the idea that we exploit our own psychological responses in order to feel others' emotions. The chapters in this book show the many ways in which the brain can be analyzed to shed light on emotions. Fear, reward, and punishment provide structuring concepts for a number of investigations. Neurochemistry reveals the ways in which different 'neuromodulators' such as serotonin, dopamine, and opioids can affect the emotional balance of the brain. And studies of different regions such as the amygdala and orbitofrontal cortex provide a view of the brain as a network of interacting subsystems. Related studies in artificial intelligence and robotics are discussed, and new multi-level architectures are proposed that make it possible for emotions to be implanted. It is now an accepted task in robotics to build robots that perceive human expressions of emotion and can 'express' simulated emotions to ease interactions with humans. Looking towards future innovations, some scientists posit roles for emotion as a powerful self-motivational tool as well as a way to work effectively in a group. But daunting questions remain as we ask what may be the nature of emotions in future generations of robots that share neither our biological heritage nor our need to share emotions with our fellow humans.