Interactive Art : Effects on User Identity and
User Satisfaction
Amy Gonzales
Abstract
Communication Dept.
Interactive art questions relationships between creator
and audience in a direct and provocative manner. Our
goal was to explore how interactive art either promotes
or constrains audiences’ sense of creative involvement,
and satisfaction. We provide a brief background on
interactive art before describing an interactive music
installation, Terpsichore, designed to test two
hypotheses in an experimental context: 1.) How does
interactive art impact user satisfaction, and 2.) How
does interactive art shape the self-concept of the user
as creative? Self-perception theory is used to predict
identity change. We examine future implications for
interactive artistic design, audience enjoyment, and
identity construction.
Cornell University
330 Kennedy Hall
Ithaca, NY 14853 USA
alg49@cornell.edu
Thomas Finley
Computer Science Dept.
Cornell University
5156 Upson Hall
Ithaca, NY 14853 USA
tomf@cs.cornell.edu
Stuart Duncan
Music Dept.
Cornell University
Keywords
Lincoln Hall
interactive art, creativity, identity, user satisfaction,
music installation
Ithaca, NY 14853 USA
spd34@cornell.edu
ACM Classification Keywords
H5.1 Multimedia Information Systems
Copyright is held by the author/owner(s).
CHI 2008, April 5 – April 10, 2008, Florence, Italy
ACM 1-xxxxxxxxxxxxxxxxxx.
H5.5 Sound and Music Computing
2
Introduction
A trend in new media research is to build interactive art
installations in which audiences have some control over
interplay with the system. Although interactivity is
defined as a neutral concept (Erickson, 2006), often it
is assumed that interaction will lead to more positive
experiences. Previous research on interactive systems
does suggest that the psychological outcomes
associated with digital interaction include enhanced
learning, entertainment, and persuasive effects (Bucy &
Tao, 2007; Richards, 2006; Vorderer, 2001). Our
interest, in particular, is in the positive impact of
interactivity on user experience and identity as a
creative person.
We use an artistic installation to examine how and why
interacting with art, versus being a passive observer,
may alter the experience of the audience. Other
interactive artistic installations have been developed for
the purposes of questioning specific cultural identities
(Gluck, 2005), or to shift the role of users from
‘audience’ to collaborator (Arrasvouri & Holm, 2007).
This is one of the fundamental purposes of interactive
art, to redefine the user as a creative source, and one
in which we hoped to further explore with the design of
our interactive sound installation, Terpsichore.
the observer. Dynamic-Passive works are those that
change in response to the physical environment, but
are not influenced by user inputs. Dynamic-Interactive
pieces generate outputs that correspond directly to
input from audiences.
Edmonds and colleagues have updated this taxonomy
(Edmonds, Turner & Candy, 2004) to include DynamicInteractive (Varying) pieces, which distinguish articles
that change over time, either through automated
learning or through updates from the artist. Such
pieces have been used to explore the potential of
technology to aid users’ reflections on identity and
intentionality (Hook, Sengers & Andersson, 2003). Our
installation, a musical system that reacts to the
physical motion of the user through spatial movement
and gesture, falls into this category of interactive art.
A Taxonomy of Interactive Art
Around the time that digitally influenced art first
became possible, Cornock and Edmonds developed a
taxonomy for categorizing art along a continuum of
interactivity (Cornock & Edmonds, 1973). They
classified interactive art as being static, dynamicpassive, or dynamic-interactive. Static artworks are
those that contain no opportunities for interaction from
Figure 1 – Participants use Wiimotes to generate sound
during interaction with the installation.
3
Implications of Interactive Art
In their original piece, Cornock and Edmonds note the
potential for interactive art to inspire creativity from
users. More recently others have argued this same
point, suggesting that interaction with a creative piece
may ultimately generate a sense of authorship on the
part of the audience (Willis, 2006). At the same time,
some artists have raised concerns about the false sense
of creative ownership that interactive art may give to
audiences (Campbell, 2000). Technical constraints on
the number of outcomes a user can create in an
artwork may actually have the opposite effect on
creativity. By necessarily trying to predict a user
behavior, the artist may unintentionally constrain a
user’s sense of autonomy, perhaps even to a larger
degree than would otherwise be experienced in a noninteractive installation (Campbell, 2000).
Psychological Research on Identity Construction
In addition to artists’ and designers’ ideas that
interactive art can generate perceptions of creativity,
there is a history of psychological research on identity
construction that supports this notion. Self-perception
theory is a theoretical paradigm based on the premise
that through self-observation people come to determine
their own identities, even when behaviors are
externally induced (Bem, 1972). So, for example, if I
eat toast every morning for breakfast because my
partner serves it to me, I decide that I am “a toast
liker.” Additional work suggests that this effect is
especially powerful when behaviors are enacted publicly
(Schlenker, Dlugolecki, Doherty, 1994). Our question is
will individuals that observe themselves generating
sound in a creative environment, come to think of
themselves as creative individuals? If this is true, it
could have implications for interaction effects, other
creative and non-creative environments, alike.
The mechanisms of both self-perception theory and
public commitment operate by activating an individual’s
sense of responsibility to social norms. Social norms
require individuals to maintain a certain level of
consistency between self perceived identity and social
behavior. By having participants experience Terpsichore
in pairs, we hope to evoke this identity construction
mechanism in the current experiment.
Design
Our study, aside from its artistic component, was
guided by two hypotheses. The first hypothesis was
that those that could interact with and change the
music would find the experience more enjoyable than
those who could not. The second was that users
affecting a piece of music would in turn gain a sense
that they themselves are intrinsically more creative and
would thus rate themselves as being more creative
than individuals that did not influence the music.
To test these two hypotheses, we designed an
experimental study in which participants would be
exposed to the installation, being either able to affect
the system’s sound output (Interactive Condition), or to
the music in the same space without user interaction
(Non-Interactive Condition).
Our design of this system consisted of three major
components: (1) the sensor component which detected
users’ efforts to affect the system, (2) the music
generation component which took the sensor input and
produced the resulting sounds, and (3) the
4
environment in which this experiment was performed.
The sensors used to detect user motion in the
installation interaction were Nintendo Wiimotes.
positions) were perturbed, and by striking in time with
the chimes the rhythmic pattern of the chimes would
change.
Wiimotes are force-sensing controllers originally
developed for the Nintendo Wii game system, but they
use the generic communication protocol Bluetooth, and
so may be adapted to entirely different contexts
beyond the game system. The Wiimotes transmitted
user acceleration data, allowing us to infer rhythmic
movements of Wiimote-holding users. Also, the system
reacted when users “struck” the Wiimote like a
drumstick by playing a bell sample.1
The experiment was only meant to last for two minutes
per participant group, and then the system would fade
out to silence. If people were able to figure out the
pattern of change that was evoked by striking the bells
in rhythmic unison, causing each participants individual
pitch to gravitate towards a central unison pitch, the
interactive conditions could potentially complete the
entire installation’s process in less than the allotted two
minutes, although this rarely occurred.
The music generation component was a Pure Data
program that read the sensor information and produced
an audio signal. Pure Data is a graphical programming
environment, often used in audio processing.2 The
particular program for this design produced sounds that
were not overtly musical, but rather were atypical
sounds that are often associated with experimental
composition. These sounds included bell-like sounds, a
low-level bell ringing overlaid by chimes, and chimes
that rang at a higher frequency corresponding to user
actions.
Each participant in the pair had their own distinct pitch
associated with his or her motions. Rhythmic patterns
became more complex as the users cycled through the
system gradually shifting their respective pitches
towards a central pitch. In other words, by performing
certain actions and striking the Wiimotes, the frequency
of two individual pitches (primarily set at high and low
1
http://sourceforge.net/projects/darwiin-remote/
2
http://puredata.info/
Figure 2 – Many interactive groups were highly active in their
use of the system.
5
Methods
We ran our experiment in random, alternating
interactive and non-interactive modes. 71 pairs in total
participated in this study. 34 pairs were exposed to the
interactive installation, and 37 pairs were exposed to
the non-interactive version of the installation.
In both conditions, users were brought into a space,
and told to “reflect on the space, the sounds, and each
other” during the process. As mentioned, in the
interactive conditions, users were given Wiimotes in
order to generate changes in the sound quality and
pitch. In the non-interactive mode, users were not
given the Wiimotes, and no sensor was turned on.
They heard the sounds that the system produced, but
this was a recording from a previous interactive session
that was played back. We did not furnish them with
Wiimotes, nor give any indication that the system did
or did not have an interactive component.
The environment itself was a music studio, roughly
thirty feet by fifteen feet. The use of the room was a
practical choice given the need to keep Wiimotes within
range of the Bluetooth sensors. The participants stood
in a square area within a quadraphonic speaker set up
at the corners. A mirror was placed within the room so
that the participants could see themselves, and a video
camera was set up on the side to monitor and record
participants’ reactions. The mirror and the video
camera were also intended to help heighten
participants’ sense of being public.
Figure 3 – This interactive group was conspicuous in that they
did not seem to bother using the system at all.
A pair of participants would enter the space, whereupon
we would introduce them to the system. We told the
participants that we were pilot testing a digital music
installation, and we were interested in their feedback.
The sound was explained as coming from experimental
bell-like sounds. We instructed them to stay within the
square bounded by the quadraphonic speakers for
optimum sound enjoyment. Those in the interactive
condition received the Wiimotes and were given a brief
primer as to their use. We deliberately did not explain
exactly how exactly they would affect the system, to
help avoid any perception of the system as being
somehow “goal driven.” Then we left them alone in the
room with each other, and came back in a few minutes
to collect them once the system completed its run.
After being exposed to the system, we gave
participants multiple questionnaires to fill-out in
response to the experience.
6
Measures
Check on Perceptions of Interactivity
To ensure that subjects perceive that the interactive
design was in fact more interactive than the noninteractive design two questions were posed to users:
(How much did you affect the sound? Was this
installation interactive?) Responses were summed for
each person to create a single Likert based interactivity
rating (range 2-14, high score=perceived interactive).
User Enjoyment
Figure 4 – The majority of non-interactive groups merely
“enjoyed” the experience simply by standing around, listening.
The quality of the experience was assessed using 3
items asking users about their enjoyment of the
installation, how much fun they had, and whether or
not they would recommend the installation to friends.
These items were summed to create a single Likert
based rating of the experience (range 3-21, high
score=positive experience).
User Identity
Figure 5 – This innovator non-interactive group decided they
would both lie down.
Two different scales were used to assess identity shift
as a function of interacting with the installation. The
first was a summary measure of 5 questions developed
by the experimenters intended to assess different
aspects of user’s self-perceptions as creative individuals
(e.g. How much do you enjoy participating in creative
activities?; range 5-35, high score=creative selfperception). The second measure was an adjective
checklist designed to assess creative personality
(Gough, 1979).
7
Results
Check on Perceptions of Interactivity
In order to test our hypotheses about the implications
of interactive art, we first conducted a t-test to ensure
that participants in the interactive conditions did
perceive the experience to be more interactive than
participants in the non-interactive conditions. As
expected, participants in the interactive condition did
find that the experience was more interactive (M=9.71,
SD=2.42) than those in the non-interactive condition
(M=5.27, SD=3.07), t(69)=6.79, p<.01.
User Enjoyment
We also used a t-test to evaluate whether participants
that interacted with the installation like it more than
those that did not interact. By summing 3 items, we
found that on average the interactive participants liked
their experience more (M=12.62, SD=3.40) than
participants that did not interact (M=10.78, SD=4.10)
with the piece, t(69)=2.67, p=.01. Further, there is a
significant correlation between participants’ perceptions
of the installation as interactive and their enjoyment of
the installation, (r=.53, p<.01). This correlation was
consistent across both interactive (r=.42, p=.01) and
non-interactive (r=.57, p<.01) conditions.
User Identity
The impact of interaction with a music installation on
participants’ perceptions of being creative was assessed
by comparing the mean rating of self-perceived
creativity for participants in both conditions. In contrast
to our hypothesis, participants exposed to the
interactive installation were not more likely to rate
themselves as being creative according to either our
scale (M=26.08, SD=4.58), or the Gough scale
(M=5.62, SD=3.10), relative to participants that did
not interact with the sound (our scale: M=26.26,
SD=5.21; Gough scale: M=5.41, SD=3.24).
Discussion
User Experience
Our results support our first hypotheses, in that users
do enjoy the experience more when they are able to
interact with the system. This result is demonstrated by
a significantly higher rating of user enjoyment by
subjects in the interactive conditions. Perhaps more
striking is the additional support for this hypothesis
provided by the positive correlation between participant
ratings of interactivity and user satisfaction. This
finding is consistent with previous studies that both
empirically demonstrate and anecdotally note that
people do appreciate interactivity. And although it is
unclear how well this finding generalizes to non-artistic
spaces, or even to other artistic spaces, the fact that
this correlation was established within each condition
separately suggests that these findings do have
implications for the interactive design.
First, the fact that participants in the non-interactive
conditions rated the installation along a range of 1 (Not
very interactive) to 7 (highly interactive) supports the
notion that, at some level, all art involves a degree of
interaction, including psychological interaction
(Campbell, 2000; Cornock & Edmonds, 1973). And
while this perspective arises in an artistic context it
may be a useful perspective to take when designing for
interactivity in non-artistic contexts as well. That is,
traditional interactive spaces, such as hypertext
8
interfaces, may not only be perceived as interactive
based on the number of opportunities for users to
physically engage the space. Our findings suggest that
interactivity may also be defined by psychological
engagement with a piece, which may not involve
physical change.
Second, because all people in the interactive condition
had the same amount of control, yet also displayed a
range of interactivity ratings further reinforces the idea
that perceptions of interactivity, rather than objectively
defined factors, are what shape a user’s experience.
Given that interactivity ratings were strongly correlated
with user enjoyment across all conditions, these
findings suggest that future research would benefit
from continued exploration of why users view a
particular piece as interactive in addition to why it may
be enjoyable.
Finally, we must make a note of caution that although
an experimental test of this system revealed that user
enjoyment increases as a function of interactivity, we
can not be sure that these findings would generalize to
other forms of interactive design, or even to additional
interactive art installations. Specifically, future tests of
this design should distinguish the effects of the
Wiimotes from actual interactivity, by allowing
participants to hold Wiimotes without impacting sound
output. In all, we do feel that this finding encourages
continued exploration into the relationship of
interactivity to user satisfaction.
User Identity
In contrast to our hypotheses, we did not find that
interaction with the art installation had any effect on
users’ self-perceptions of creativity. Although it is
difficult to make generalizations based on null results,
in brief, we note some possible explanations for the
lack of effects.
First, it is possible that our manipulation failed to
activate the participants’ creative self-concept. Previous
research on self-concept change has generally given
individuals explicit instructions about how to present
themselves. Individuals asked to present themselves in
a particular manner (e.g. extroverted) leads to
internalization of that behavior (Schlenker et al., 1994).
Perhaps if we had explicitly prompted users to explore
their “creative potential” while interacting with the
artwork they may have altered their perceptions of
themselves as creative.
On the other hand, consistent with what Campbell and
others have argued, it is also possible that interactive
art does not actually induce a sense of creativity as
otherwise presumed. Of course, neither does these
findings suggest that art constrains creativity, as there
was no significant difference in self-reports of creativity
between groups.
Finally, it may be that interactive art inspires a
temporary state of creativity without actually shifting
users’ concepts of themselves as creative. However,
this would be somewhat inconsistent with assumptions
by designers that user generated creativity would
increase their own sense of creative authorship
(Edmonds et al., 2004; Willis, 2006). Because this is an
9
important component of many interactive designs, we
feel that the question of whether or not interactive art
actually can influence identity (Edmonds et al., 2004;
Willis, 2006) deserves continued attention.
additional questions about the impact of interactivity on
user identity, and whether or not interacting with art
can change self-perceptions of one’s sense of creativity.
References
As mentioned, future tests using this system may
involve a different set of instructions to users. An
additional alternative is to alter the design so as to
increase users’ sense of control. Work by Erickson has
noted that interactive art must attain a fine balance
between work that is intuitive, yet challenging enough
to remain interesting (Erickson, 2006). It is possible
that if the system were more intuitive, user’s would
have a more immediate sense of themselves as
‘composers’ which may in turn be more likely to shift
self-perceptions of creativity, although we can not be
sure of this without additional tests. Finally,
adjustments to our measurements of what it means to
be creative may also be necessary.
Conclusion
Campbell writes that expecting different creative
ouputs from objects that are constructed out of discrete
inputs may be impractical (Campbell, 2000). As a
culture, we’ve begun to look to technology to express
ourselves artistically, yet technology is generally built
to limit ambiguity, and maximize efficiency. One might
argue that applying an experimental test of differences
in user experiences with interactive and non-interactive
art continues to inappropriately constrain what should
be an open, undefined process- the production and
experience of art. However, knowing that interactivity
does increase user enjoyment, at least in this case,
may serve to prompt continued research at more
nuanced levels, and in other realms. Also, this study
has given us a starting point from which to ask
[1] Arrasvuori, J., & Holm, J. (2007). Designing
interactive music mixing applications for mobile
devices. Proceedings of the 2nd international
conference on Digital interactive media in
entertainment and arts, ACM: NY
[2] Bem, D.J. (1972). Self-perception theory. In L.
Berkowitz (Ed.), Advances in Experimental Social
Psychology, 13, 420-432
[3] Bucy, E.P. & Tao, C. (2007). The mediated
moderation model of interactivity. Media Psychology, 9,
647-672
[4] Campbell, J. (2000). Delusions of dialogue: Control
and choice in interactive art. Leonardo, 33, 133-136
[5] Cornock, S. & Edmonds, E. (1973). The creative
process where the artist is amplified or superseded by
the computer. Leonardo, 1, 11-16
[6] Edmonds, E., Turner, G., & Candy, L. (2004).
Approaches to interactive art systems. Proceedings of
the 2nd international conference on Computer graphics
and interactive techniques in Australasia and South
East Asia, Singapore
[7] Erickson, T. Five Lenses: Towards a Toolkit for
Interaction Design. (2006), Theories and Practice in
Interaction Design (ed. S. Bagnara, G. CramptonSmith, and G. Salvendy.) Lawrence Erlbaum: p.301310.
[8] Gluck, R. J. (2005). Sounds of a community:
Cultural identity and interactive art. Leonardo Music
Journal, 15, 37-43
[9] Gough, H. G. (1979). A creative personality scale
for the Adjective Check List. Journal of Personality and
Social Psychology, 37, 1398-1405.
10
[10] Hook, K., Sengers, P., & Andersson, G. (2003).
Sense and sensibility: evaluations and interactive art.
Proceedings of the SIGCHI conference on Human
factors in computing systems, Ft.Lauderdale, Florida,
241-248
[11] Richards, D. (2006). Is interactivity actually
important?, Proc. of 3rd Australasian Conf. on
Interactive Entertainment (IE’2005), Murdoch
University, Perth
[12] Schlenker, B.R., Dlugolecki, D.W., & Doherty, K.
(1994). The impact of self-presentations on selfappraisals and behavior: The power of public
commitment. Personality and Social Psychology
Bulletin, 20, 20-33.
[13] Vorderer, P., Knobloch, S., & Schramm, H. (2001).
Does entertainment suffer from interactivity? The
impact of watching an interactive TV movie on viewers’
experience of entertainment. Media Psychology, 3, 343363.
[14] Willis, K.D.D. (2006). User authorship and
creativity within interactivity. Proceedings of the 14th
annual ACM international conference on Multimedia,
731-735