I Smell Creativity: Exploring the Effects of Olfactory and Auditory Cues to Support Creative Writing Tasks

Abstract

Humans perceive different objects, scenes or places using all their senses. Our sensory richness also plays an important role for creative activities. Humans also recall those sensory experiences in order to spark creativity, e.g. while writing a text. This paper presents a study with 100 students, divided in groups, that explores the effect of auditory and olfactory cues and their combination during a creative writing exercise. Our results provide useful insights suggesting that olfactory cues have an important role in the creative process of users and even when this type of cues are combined with auditory cues. We believe, that this type of modalities should gain more relevance on the development of creativity support tools and environments for supporting the creative writing process .

Full text

I Smell Creativity: Exploring the Effects of Olfactory and
Auditory Cues to Support Creative Writing Tasks
Frederica Gonçalves1, Diogo Cabral1, Pedro Campos1, Johannes Schöning1,2
1Madeira-ITI, University of Madeira, Funchal, Portugal
frederica.goncalves@m-iti.org; diogo.cabral@staff.uma.pt;
pcampos@uma.pt
2University of Bremen, Human Computer Interaction, Bremen, Germany
schoening@uni-bremen.de
Abstract. Humans perceive different objects, scenes or places using all their
senses. Our sensory richness also plays an important role for creative activities.
Humans also recall those sensory experiences in order to spark creativity, e.g.
while writing a text. This paper presents a study with 100 students, divided in
groups, that explores the effect of auditory and olfactory cues and their combina-
tion during a creative writing exercise. Our results provide useful insights sug-
gesting that olfactory cues have an important role in the creative process of users
and even when this type of cues are combined with auditory cues. We believe,
that this type of modalities should gain more relevance on the development of
creativity support tools and environments for supporting the creative writing pro-
cess.
Keywords. Creativity Support Tools; Creativity; Olfaction; Odor; Sound; User
studies; Creative Writing;
1 INTRODUCTION
Writing is among one of the top forms of human artistic expression [1]. Creative writing
is also a central component to a range of different industries and disciplines (e.g. jour-
nalism, science fiction, advertisement, etc.) [2]. Even so the process of writing a text is
often much unstructured, one can observe the following steps in a creation of a text: a
writer starts with a prewriting, finishes the first draft, revises the draft, and then edits
and finishes a final version of the work [1]. During this process, writers try to obtain
their creativity from a very wide range of sources ranging from memories of dreams or
television news reports, in a way in which they can imagine new characters or situations
that can be included in their writings [3].
In our digital world 80% of our texts are created with the help of computer systems
[4]. Therefore, different creativity support tools help people engage creatively with the
world [5], and some researchers claim [6] that it is a challenge for human-computer
interaction researchers and user interface designers to construct information technolo-
gies that support creativity. There are currently some possible solutions in creative writ-
ing tools to foster the creativity of writers. One example, is writing prompts [7,8] that

are simple phrases meant to help writers trigger their creativity and start writing fluently
without losing their time with a "writer's block", not being able to write a single line of
text. Shneiderman [9] argues that there’s an effort for developing creativity support
tools, which enable us to explore, discover, imagine, innovate, compose and collabo-
rate. Even with decades of creativity research, there is no single, agreed upon method-
ology for evaluating how well a creativity support tools to aid the creativity of its users
[10]. Creative workers use artificial stimuli as inspirational guides, e.g., listening to
music while writing or looking at images while drawing.
Certainly, well-designed creative writing tools can help users in generating multiple
levels of creativity during the process of writing, particularly tools that can also gener-
ate different stimuli [11].
Olfactory cues are well known to have a strong emotional effect on arousal level and
task on arousal level and task performance have also been suspected but not well ex-
plored in the literature [11]. Among other researchers, Seo [12] states that even though
we often perceive odours in the presence of various background sounds, there is limited
knowledge about the effects of background sound on odour perception. Spence [13],
has studied the effect of background noise on the sensory-discriminative aspects of
taste/flavour perception and on people’s hedonic responses to food and beverage items.
He highlights the impact of background music and/or soundscapes on food and bever-
age perception/consumption on people’s sensory-discriminative and hedonic re-
sponses.
Therefore, this paper presents the first study on the effects of different modalities
(sound and smell) on the creative writing process. The core contribution of this research
is a novel between-subjects study to discover patterns in the influence of smell and
sound on the participants' creativity while using a word processor in different environ-
ments. The research described below makes two supporting contributions. First, we
triangulate qualitative and quantitative data from different sources to assess creativity
of users in different writing conditions, giving us useful insights to develop a novel
prototype for supporting the creative writing process. Second, although research on
creativity has thus largely occurred within areas such as psychology and neuroscience
and HCI, we believe other areas of computer science can also contribute to this domain.
2 RELATED WORK
Researchers have investigated how to improve the creative writing process [14], e.g.
by giving students the opportunity to interact with real readers of their work, showing
that this could lead to an increased motivation to write. Advances in creative writing
tools have been mostly made in very specific areas. Yannopoulos [15] proposed a sym-
bolic language intended to express the content of films (motion pictures) much as notes
provide a language for the writing of music, therefore bringing a new approach to the
creative process of filmmaking. Goulet [16] focused on computer writing tools used
during the production of documents in a professional setting. They report on a focus
group conducted with professional writers, in which writers narrated their experiences
using computer tools to write documents, describing their practices, pointing out the

most important problems they encountered, and analyzing their own needs. Based on
this work, they describe LinguisTech, a reference website for helping language profes-
sionals. Keeping in mind that one goal of digital tools for creative writing is to help
users produce greater quantities of writing, Coughlan and Johnson [17] present three
perspectives on creative interaction that have emerged from four years of empirical and
design research. They argue that creative interaction can be usefully viewed in terms of
Productive Interaction – focused engagement on the development of a creative out-
come; Structural Interaction – the development of the structures in which production
occurs; and Longitudinal Interaction – the long-term development of resources and re-
lationships that increase creative potential.
Perception of olfactory and auditory cues
Ho and Spence [18] investigate the differential effects of olfactory stimulation on
dual-task performance under conditions of varying task difficulty. Their results provide
the first empirical demonstration that olfactory stimulation can facilitate tactile perfor-
mance, and highlight the potential modulatory role of task-difficulty in odour-induced
task performance facilitation.
Some researchers, such as Xiang et al. [19] presented in their study a prototype sys-
tem – Olfaction - that emits odour emoticons and it was applied in two contexts: online
text chatting and voicemail receiving. Their results suggested that odour emoticons in-
duced more chatting, and were easy to use, and helped participants to better perceive
and convey emotions. Studies have been conducted with auditory cues in different areas
such as consumer behaviours [20] and consumers' perception of food texture and qual-
ity [21].
Other studies were conducted to investigate the effect of visual cues on olfactory
perception in humans [22]. Guest et al. [23], investigated whether similar auditory ma-
nipulations change people’s perception of the roughness of abrasive surfaces and rep-
licated the rubbing-hands manipulation of previous experimenters while participants
rated either the perceived roughness or wetness of their hands. In these experiments, it
was possible to demonstrate that auditory frequency manipulations can have an influ-
ence on the perceived tactile roughness and moistness of surfaces.
Creativity support tools
To properly investigate creativity, it is appropriate to adopt a variety of methods and
perspectives to make it plausible and understanding. Creativity is a complex and mul-
tifaceted phenomenon [24] and includes discovery or invention of a significant idea,
pattern, method, or device that gains recognition from accepted leaders in a field [25].
Over the last decades of creativity research there is no consensus on how to evaluate
how well a Creativity Support Tool (CST) supports the creativity of its users [10]. As
Joyce states [26] emerging computer-based tools can develop better and more creative
solutions to the problems that we face in our days. Other researchers [27] considers that
success during software development, despite of being a conceptually complex,
knowledge intensive and cognitive activity, depends on the creativity of software engi-
neers. Shneiderman [25] argues that it is a challenge to construct information technol-
ogies that support creativity and the goal of developing new CST can be obtained by
building upon an adequate understanding of creative process. Also, the goal of CST is

to develop improved software and user interfaces that make users become more pro-
ductive, and more innovative [6]. Creativity and motivation enhancement can easily be
aligned with the design of high-quality human-computer interaction and creativity can
be viewed as any process which results in a novel and useful product, as stated by [28].
Researchers have also targeted other stimuli to support creativity, such as the visual
stimuli, images and text, increase both originality and diversity of ideas during brain-
storming [29,30,31]. Other such as Gonçalves et al. [32], studied UI Zen-based themes,
composed of sound and images, foster inspiration, focus and immersion on creative
writing tasks.
In the next two sections, we will describe the evaluations of the different environ-
ments, including the methods, participants, procedures and results of each. All data
taken from the experiment was made completely anonymous.
3 USER STUDY
We conducted a between-subjects experiment to investigate if olfactory or auditory
cues affect peoples' creativity during a creative writing exercise. Since writing is part
of everyday routine of our sample, the activity in this experiment involved writing a
short story using Microsoft Word (MS Word) as the writing environment. We chose
MS Word because the participants (high school students) were all familiar with.
3.1 Conditions
The writing activity was conducted under four different conditions: Neutral Environ-
ment (no cues); Smell Environment (cues alert/relax), Sound Environment (cues
alert/relax) and Smell+ Sound (both cues, sound/smell combined).
Neutral Environment. Our baseline (session 1, N=14) without any cues.
Sound Environment. In this condition, we used two auditory cues: an alerting
sound (session 2, N= 17) through which users might feel that they would be sitting in
a cafe with the constant bustle of movement of people, machines and dishwashers, and
a relaxing sound (session 3, N=14), achieved using a natural soundtrack featuring wa-
ter, birds and foliage.
Smell Environment. We used two types of olfactory cues: an alerting smell (session
6, N= 12) which was achieved by the actual fragrance of hot coffee that was spread
around the room, and a relaxing smell (session 4, N=10) for which we used a laurel
fragrance.
Smell + Sound Environment. We used the combination of sound and smell in each
condition – alert and relaxed. One group of students have smelled the real coffee and
listen the sound of a café (session 7, N=21), and the other group of students smelled
the relaxing fragrance and listen the sound of nature (session 5, N=12).
Smell Spread: To make sure that all students could smell the sense that was in the
classroom, ten minutes before the session six and session seven started, we made coffee
in the school's kitchen, and placed the coffee pot in the classroom, with doors and win-
dows closed. To spread it we used a fan and small cups of coffee were placed next to

each computer in case participants wanted to smell it more during the writing. In session
four and session five we spread olfactory laurel fragrances using air fresheners' room
spray by Air Wick
1
, that remains for up to one hour and is propelled by 100% filtered
air. Participants were exposed to smell during the session. Sessions of the different ol-
factory cues were done on different days.
Audio: To play back the audio we used two sound columns of 2x10 Watt output power.
Since this was a study conducted in a school environment, including students that
were minors of age, a protocol was setup between parents, school and researchers re-
garding all the data gathered, even if anonymous, in order not to raise any privacy or
security issues.
3.2 Task
Students had to write a short story about an “imaginary path on an island” [7] [33]
Therefore, map
2
(Figure 1) was handed out to all students before the activity. They were
instructed to choose two points from the map, and through their own imagination and
creativity they had to write a story about their path from a point A to point B. They had
to initiate their writing from their own ideas. Participants were given 15 minutes as a
time limit to complete the writing task in all conditions.
Fig. 1. Map used to trigger the writing task.
1
http://www.airwick.us/products/room-sprays/
2
All rights reserved to Sofia Vasconcelos http://frommadeiratomars.com/pt/

3.3 Participants
The study involved 100 students (45 female, 55 male), from the institution’s population,
aged between 15 and 19 years old (M=15.9; SD=0.94). They were recruited through
the school's mailing list. All subjects were naïve to the experimental conditions. Eve-
ryone reported having a normal olfactory and auditory acuity.
3.4 Measures
We used the following measures for our experiment: Creative Behavior Inventory [34];
Flow Theory dimensions and Self-assessments; Creativity Support Index and Post-ex-
periment interviews.
Creative Behaviour Inventory (CBI) is a psychometric tool to investigate ones past
creative behaviour and activities [34]. The CBI was used to access the different creative
levels of the participants. We used a subset of thirty items from the original CBI [34]
[35], particularly the specific examples of creative activities such as: literature, miscel-
laneous, performing arts, science and music [36]. As Hocevar states [34], an item score
was based on response category: zero points for never, one point for once, two points
for twice, three points for 3-4 times and four points for more than five times. We clas-
sified subjects into two classes: subjects that had less than 30 points on the CBI as being
less creative, and participants that had more than 30 points on the CBI as highly creative
[10].
The Creativity Support Index (CSI) is a survey to assess a tool’s creativity support
which users provide ratings for six dimensions of creativity support: Enjoyment, Ex-
ploration, Expressiveness, Immersion, Results Worth Effort, and Collaboration [5].
We asked participants to self-rate their creativity [37], answering the question “I
consider myself a creative person” ranking a seven-point Likert with the evidence scale
for 1 (totally disagree) and 7 for (totally agree).
Another different approach to measuring creativity is the one of the Flow theory
[38]. Csiksentmihalyi defines flow as “a state in which people are so involved in an
activity that nothing else seems to matter; the experience is so enjoyable that people
will continue to do it even at great cost, for the sheer sake of doing it” [38]. As the
author argues, we all experience flow from time to time and we recognize its charac-
teristics. When we are “in the flow” we typically feel strong, alert, in effortless control,
unselfconscious, and at the peak of our abilities. Therefore, we focused on the following
Flow Theory dimensions: concentration, sense of control, losing track of time, and loss
of self-consciousness. These dimensions were evaluated by the participants while rank-
ing the following 7-point Likert scale questions: "I felt very concentrated during this
task”; “I was able to solve this task without any problem”; “I lost track of time during
this task”; and “I lost my attention during this task”.
To access participants' mental well-being, they were asked to choose up to three
adjectives from the following list: Surprised, Delighted, Laid back, Depressed, Pacific,
Happy, Tired, Bored, Sad, Satisfied, Frustrated, Angry, Serious, Animated, Distressed,
Creative, and Frightened.

Finally, we collected qualitative data from all users with a set of questions such as:
“Did you enjoyed to write in this condition and why?”; “This condition gave you more
immersive and emotionally engaging experience and why?”; “This condition felt more
natural to you to write and why?”; “In this condition you felt more creative and why?”;
“Did you fell more relaxed and why?”; “Did you fell more stressed and why?”; “Is
there any comment that you would like to add?”, in order to know the participant's
opinion about the whole experience.
3.5 Procedure
The experiment was conducted in a classroom in the secondary high school (Figure 2),
during two weeks.
Fig. 2. Students from session 7, performing the study in the moment of the experience.
A preliminary evaluation was conducted with six participants to examine the feasi-
bility and accuracy of the smell spread in the classroom. Three participants were in one
classroom with the smell of coffee and the other three were in a classroom with the
smell of laurel. For this purpose, we used a simple creative writing technique [36] that
participants were presented with an image. They were asked to write a simple story
during fifteen minutes. Following, participants were asked whether they could smell
the odour during the writing task. The smell was already spread into the classroom as
previously mentioned (see 3.1 Conditions, Smell Spread). Since participants reported
that they smelled the odour, it was considered that the study may be conducted accord-
ing the procedure.
In the main experiment, participants were brought to the classroom (Figure 2), pre-
viously prepared for it. The experiment was conducted in a classroom equipped with
computers on desks, and the time requirement for each session, including pre-question-
naire, instructions, experiment and debriefing took over an hour.

When participants entered the classroom, they delivered the document with the ex-
periment protocol and the authorization to participate, which was previous delivered to
them. Before starting each session, it was asked if the smell bothered any of the partic-
ipants. First, participants were requested to fill the Creative Behaviour Inventory [31],
and to self-rate their creativity [33]. After filling the inventory, the writing task was
explained. Then, when they finished the writing task, they had to fill the Creativity
Support Index [3], together with the very short survey based on Flow Theory [35] and
to select up to three adjectives from the list that was mentioned before. Lastly, we in-
terviewed them.
4 RESULTS
To inquire the impact of olfactory and auditory cues on user's apparent and experienced
creativity, we triangulated different data sources, such as behavioural data, users' verbal
accounts during task execution, self-reports using psychometric scales of creativity and
data from our exit interviews. We will refer to the creative writing conditions previously
presented and shown in Table 1.
Table 1. Description of the creative writing conditions used.
Creative Writing
Condition
Session
N
Baseline
Base
14
Sound Alert
SA
14
Sound Relax
SR
17
Smell Relax
SmR
10
Smell + Sound Relax
SmSR
12
Smell Alert
SmA
12
Smell + Sound Alert
SmSA
21
Are our samples equally creative?
Participants self-rated their creativity (M=4.85; SD= 1.17) in a seven-point Likert
Scale before starting the experience. Results are shown in Tables 2 and 3.
Table 2. Frequency and percentages "I consider myself a creative person".
Seven-point Likert scale
Frequency
Percent
1
0
0%
2
0
0%
3
11
11%
4
31
31%
5
32
32%
6
14
14%
7
12
12%
Total
100
100%

Table 3. Descriptive of self-assessment about creativity in each creative writing conditions.
Sessions
Base
SA
SR
SmR
Sm
SR
Sm
A
SmS
A
Mean
4.36
4.71
5.06
5.00
5.17
4.67
4.95
St. Devia-
tion
.84
1.38
1.19
1.49
1.03
.88
1.24
Std. Error
.225
.370
.290
.471
.297
.256
.271
Minimum
3
3
3
3
3
3
3
Maximum
6
7
7
7
7
6
7
A chi-square test of independence was performed to examine the relation between
students in each condition and their self-assessment of creativity. The relation between
these variables was not significant, X (24) =17.992, p=.803. Therefore, there is no sta-
tistically significant association between participants and their self-assessment of crea-
tivity, i.e., students consider themselves creative persons.
Through Cronbach’s alpha, the CBI inventory was found to be highly reliable (30
items; α= .84). Regardless of the self-report scale in the 30-item CBI that could capture
a creative accomplishments and activities in past behaviours, our results suggest in
terms of past creative actions that 42 students had more than 30 points on the CBI
("highly creative") and 58 students had less than 30 points on the CBI ("less creative").
We compared the gender of the subjects with their creative level of the self-report scale
in CBI to see if there were any difference between genders. On average, female partic-
ipants reported greater creativity in past activities (M=30.78; SD=16.40) than male par-
ticipants (M=25.86; SD=14.99). This difference was not statistically significant (t (98)
= 1.57, p>.05). However, it did represent a small-sized effect r=.13.
We also compared the subjects in each condition (session 1 to session 7) and their
creative level of the self-report scale in CBI. Some descriptive statistics are shown in
Table 4. Table 4. Average (SD) from CBI in each creative writing condition.
Sessions
Base
SA
SR
SmR
SmSR
SmA
SmSA
M (SD)
20.64
(12.73)
20.01
(14.93)
23.71
(9.50)
38.20
(15.55)
35.58
(14.40)
33.83
(23.26)
29.48
(13.38)
A chi-square test of independence was performed to examine the relation between
the participants in each condition and their results in CBI (creativity in past activities).
The relation between these variables was not significant, X (288) =296.04, p=.360.
Therefore, there is no statistically significant association between participants and their
creativity and all groups were equally creative in past activities.
Did olfactory or auditory cues lead to increased flow?
To access the participants’ mental well-being, we asked them to select up to three
adjectives from the following list: Surprised, Delighted, Laid back, Depressed, Pacific,
Happy, Tired, Bored, Sad, Satisfied, Frustrated, Angry, Serious, Animated, Distressed,
Creative and Frightened. Table 5 displays the percentages for each adjective, as selected
by the participants on each creative writing environment.

We can see in Table 5 that Animated, Relaxed, Satisfied, Creative and Pacific were
the most chosen adjectives.
A chi-square test of independence was performed to examine the relation between
students and the adjectives selected in each creative writing condition. The relation be-
tween these variables was only significant for the adjective "Tired", X (6) =20.490,
p=.002. By looking at the data represented in Table 5, we can conclude that there is a
statistically significant association between participants and their state of mental well-
being in session one (baseline), session six (smell alert) and session seven (smell +
sound relax);, suggesting that students felt somehow tired in these sessions.
Table 5. Results in percentages for the adjectives chosen by participants in each creative writ-
ing condition.
Adjec-
tives/Sessions
Base
SA
SR
SmR
SmSR
SmA
SmSA
Distressed
-----
7.1
-------
-------
-----
-----
9.5
Animated
35.7
28.6
52.9
40.0
33.3
50.0
23.8
Satisfied
57.1
35.7
17.6
30.0
58.3
33.3
23.8
Bored
7.1
7.1
5.9
-------
-----
-----
4.8
Pacific
21.4
21.4
52.9
20.0
50.0
33.3
14.3
Relaxed
78.6
64.3
52.9
60.0
75.0
66.7
52.4
Creative
57.1
57.1
70.6
40.0
75.0
83.3
57.1
Astonished
21.4
7.1
17.6
-------
8.3
8.3
23.8
Serious
14.3
14.3
23.5
-------
-----
-----
14.3
Fear
-----
7.1
-------
-------
-----
-----
-----
Frustrated
-----
14.3
-------
10.0
-----
8.3
19
Happy
28.6
-------
29.4
20.0
16.7
8.3
28.6
Delighted
14.3
-------
-------
-------
-----
8.3
4.8
Tired
7.1
-------
-------
-------
-----
8.3
33.3
Angry
-----
-------
10.0
-------
-----
-----
-----
Sad
-----
-------
10.0
-------
-----
-----
-----
Depressed
-----
-------
10.0
-------
-----
-----
-----
TOTAL
N=14
N=14
N=17
N=10
N=12
N=12
N=21
Thus, when we asked about how they felt in the writing task, participants reported
similar thoughts to the creative writing conditions in this case, e.g., "I did not have much
creativity, so I felt tired of writing." [baseline, P12]; "I got tired of writing and imagin-
ing the end of the story I was writing (...)" [smell+sound relax, P10]; "I did not feel like
writing and I had no ideas to write." [smell alert, P2].
Some statistical results are shown in Table 6, to test differences between each crea-
tive writing environment from the answers in the survey based on Flow Theory [35]
dimensions and ranked by participants in a seven-point Likert scale.

Table 6. Average (SD) for Flow Dimensions in each creative writing condition.
Sessions
Base
SA
SR
SmR
SmSR
SmA
SmSA
(i) Concentra-
tion
4.64
(1.82)
5.50
(1.74)
4.47
(1.80)
4.00
(1.25)
5.25
(1.29)
5.25
(1.95)
5.00
(1.81)
(ii) Sense of
Control
5.00
(1.75)
5.42
(1.60)
5.11
(1.69)
4.90
(1.85)
5.91
(.99)
5.17
(1.67)
4.90
(1.73)
(iii) Loss of
self-conscious-
ness
5.79
(1.37)
5.78
(1.93)
4.65
(2.19)
4.60
(1.89)
5.00
(1.70)
4.50
(2.28)
5.38
(1.69)
(iv) Lost
Track of Time
3.79
(2.08)
2.71
(2.27)
3.35
(2.06)
3.50
(1.84)
2.41
(1.68)
2.67
(2.23)
3.47
(2.16)
A Skewness and Kurtosis and Kolmogorov-Smirnov test did not show a normal
distribution of Flow Dimensions scores in each creative writing condition and, there-
fore, non-parametric tests were conducted. A Kruskal-Wallis test, showed that there
was not a statistically significant difference in Concentration levels of participants' be-
tween the different creative writing conditions, H (6) = 8.36, p=.213 with a mean rank
Concentration score of 46.21 for session 1 (baseline), 61.71 for session 2 (sound alert),
43.74 for session 3 (sound relax), 33.20 for session 4 (smell relax), 54.96 for session 5
(smell+sound relax), 57.88 for session 6 (smell alert) and 52.83 for session 7
(smell+sound alert). The Sense of Control level of participants was not significantly
affected by the different sessions, H (6) = 3.192, p=.784 as well as the levels of Loss of
Self-Consciousness, H (6) =7.724, p=.259 and the levels of Lost of Track Time, H (6)
= 6.391, p=.381.
Yet, by looking at the Table 6, one can see that the Lost Track of Time dimension
was not a significant issue for any of the environments we evaluated. The Sense of
Control, Loss of self-consciousness, and Concentration dimensions, results were most
consistent in all environments.
Qualitative data revealed that students in the Smell + Sound Environment (33.3%)
felt more stressed and others claim that the noise interfered with their concentration.
7.1% of the users in the Neutral Environment felt stressed because they knew they were
contributing to a study. In the Sound Environment 12.9% of the users felt more stressed
as 18.8% in the Smell Environment. Participants in these environments felt pressure to
write under a limited time, but both the smell of the coffee and the laurel relaxed them,
and the same for the sound cues. For instance, e.g., "I felt stressed about knowing that
I am contributing to an investigation." [baseline, P10]; "Yes, I felt stressed, because I
had a time limit to finish." [sound relax, P13]; "Yes, the fact of having a time limit left
me stressed, but it did not stop me from expanding my imagination." [sound alert, P14];
"Yes a lot of pressure to write, however the smell of coffee allowed me to abstract a
little." [smell alert, P2].
Did olfactory or auditory cues lead to increased output?
Regarding the stories written and the data dispersion, results showed (Table 7) that
the participants in the session 4 (smell relax) wrote an average of 412.6 words
(SD=97.3), which contrasts with other creative writing conditions.

Table 7. Average (SD) for words written in each creative writing condition.
Sessions
Base
SA
SR
SmR
SmSR
SmA
SmSA
Words
Written
267.0
(132.1)
246.0
(90.23)
269.5
(105.7)
412.6
(97.3)
270.1
(90.0)
213.8
(79.7)
300.7
(79.6)
A Skewness and Kurtosis and Kolmogorov-Smirnov test did not show a normal
distribution of words written in each creative writing condition and, therefore, non-
parametric tests were conducted. A Kruskal-Wallis test, showed that there was a statis-
tically significant difference between the words written by the participants' and the dif-
ferent creative writing conditions, H (6) = 20.449, p=.002. Mann-Whitney tests were
used to follow up this finding with a correction at a .0083 level of significance, but we
did not find statistically differences between sessions.
Did olfactory or auditory cues lead to increased CSI?
We used the CSI [3] as a way of evaluating how well each environment (neutral,
smell, sound or smell + sound) supported the creativity of the participants. 100 students
generated an average overall CSI score as shown in Table 8 for creative writing task in
each environment. Tables 9, 10 and 11 show the average factor counts, average factor
score, and average weighted factor score for each of the six factors on the CSI in each
environment session.
Table 8. Overall CSI Score in each condition: Average (SD).
Sessions
Base
SA
SR
SmR
SmSR
SmA
SmSA
Overall
CSI
Score
54.43
(15.77)
71.03
(21.09)
58.49
(18.88)
51.37
(17.30)
77.60
(15.97)
74.03
(14.92)
51.73
(21.89)
Table 9. CSI Avg. Factor Counts (SD): Sum of Results of Two Questions by Factor.
Sessions
Base
SA
SR
SmR
SmSR
SmA
SmSA
Results
Worth Effort
11.29
(2.76)
14.50
(5.27)
11.88
(4.40)
9.70
(4.92)
17.00
(2.31)
15.50
(3.24)
12.14
(5.00)
Explora-
tion
10.71
(2.43)
13.40
(5.27)
11.47
(3.79)
10.50
(3.60)
15.10
(3.78)
13.60
(3.41)
10.86
(4.29)
Enjoyment
10.36
(4.57)
14.70
(4.45)
12.06
(4.58)
9.80
(4.78)
15.30
(3.97)
15.30
(4.16)
6.38
(5.69)
Expres-
siveness
9.64
(4.67)
12.90
(5.32)
11.35
(4.34)
10.90
(4.25)
15.30
(3.13)
13.70
(3.23)
11.43
(4.61)
Immersion
11.86
(5.67)
17.30
(3.06)
11.35
(3.89)
8.80
(3.68)
15.40
(4.50)
16.70
(4.16)
11.19
(6.03)
Collabora-
tion
11.57
(3.34)
14.50
(5.42)
12.18
(4.23)
10.50
(3.60)
13.90
(5.02)
14.20
(3.33)
7.43
(4.23)

Table 10. Avg. Factor Score (SD).
Sessions
Base
SA
SR
SmR
SmSR
SmA
SmSA
Results
Worth Effort
2.14
(1.10)
2.40
(1.43)
1.94
(1.20)
2.70
(1.34)
2.40
(1.26)
2.60
(1.35)
2.24
(1.30)
Explora-
tion
3.36
(1.15)
3.50
(1.08)
3.47
(1.46)
2.80
(0.92)
3.50
(1.08)
3.30
(1.16)
3.29
(1.15)
Enjoyment
2.00
(1.47)
1.10
(0.99)
2.18
(1.33)
2.30
(1.42)
2.20
(1.23)
1.10
(1.45)
1.71
(1.71)
Expres-
siveness
3.57
(1.60)
4.50
(0.53)
3.65
(1.22)
3.80
(1.14)
2.90
(1.66)
3.70
(1.25)
3.43
(1.57)
Immersion
1.71
(1.49)
2.40
(1.17)
2.18
(1.01)
2.00
(1.56)
2.20
(1.48)
2.40
(1.43)
2.90
(1.22)
Collabora-
tion
2.21
(1.89)
1.10
(1.66)
1.59
(1.62)
1.40
(1.35)
1.80
(1.23)
1.90
(1.79)
1.43
(1.63)
Table 11. Avg. Weighted (SD).
Sessions
Base
SA
SR
SmR
SmSR
SmA
SmSA
Results
Worth Effort
24.71
(14.96)
31.00
(15.61)
25.00
(21.95)
28.80
(22.58)
42.00
(23.72)
40.10
(25.31)
28.81
(19.18)
Explora-
tion
35.71
(14.35)
45.50
(22.42)
40.59
(20.77)
28.00
(10.91)
52.60
(22.10)
45.90
(21.88)
36.71
(19.23)
Enjoy-
ment
21.21
(20.36)
16.50
(18.45)
25.94
(15.95)
25.10
(17.43)
36.20
(27.60)
18.60
(26.50)
10.33
(15.63)
Expres-
siveness
35.29
(25.16)
58.90
(26.27)
43.65
(27.73)
40.20
(20.33)
43.90
(25.99)
49.50
(18.85)
40.10
(24.17)
Immer-
sion
19.43
(17.91)
43.20
(25.02)
25.24
(15.76)
18.60
(21.13)
33.00
(25.96)
39.70
(27.23)
32.29
(22.27)
Collabo-
ration
26.93
(23.07)
18.00
(31.45)
15.06
(13.81)
13.40
(14.52)
25.10
(19.89)
28.30
(26.60)
8.95
(11.42)
The Shapiro-Wilk test did not show a normal distribution of CSI scores and, there-
fore, non-parametric tests were conducted. The Kruskal-Wallis test showed that there
was a statistically significant difference in CSI scores between the different creative
writing conditions, H (6) = 24.046, p=.01 with a mean rank CSI score 39.21 for Base
(baseline), 60.71 for SA (sound alert), 45.29 for SR (sound relax), 37.50 for SmR (smell
relax), 73.42 for SmSR (smell+sound relax), 70.46 for SmA (smell alert) and 37.12 for
SmSA (smell+sound alert). Mann-Whitney tests were used to follow up this finding. A
correction was applied and so all effects are reported at a .0024 level of significance.
CSI scores in SA (Mdn=50.17) differ significantly from CSI scores in SmSR
(Mdn=81.33), U=25.5, p=.002, r=-.59.

There was a statistically significant difference in CSI scores from students in SmR
(Mdn=54.33) and SmSR (Mdn=81.33) U=12.0, p=.001, r=-.67. Also, CSI scores from
SmR differ significantly from SmA (Mdn=77.17), U=12.0, p=.001, r=-.67.
CSI scores in SmSR (Mdn=81.33) were statistically different from CSI scores in
SmSA (Mdn=50.33), U=41.5, p=.001, r=-.55.
Finally, CSI scores in SmA (Mdn=77.17), when compared with CSI scores in SmSA
(Mdn=50.33), U=42.5, p=.001, r=-.55.
From the results, we can conclude that the combination of olfactory cues with audi-
tory cues can significantly affect the level of support to the creative process of a digital
creativity support tool, when compared to a condition that does not use any cue (olfac-
tory or auditory). This holds true whether the cues are alerting or relaxing. Also, this is
verified when there are only smell cues (alert or relax) in the creative writing task.
However, when using only auditory cues, the results do not show any evidence of in-
creased creative process support.
Triangulating the results with semi-structured interviews allowed us to support some
of these results and observations. Students were curious to know what the writing task
was, since none of the users had ever taken an experiment such as this one. In general,
we observed that the participants were concentrated on the writing moment. It was clear
that all participants felt somewhat creative during the experiment.
In the semi-structured interview, participants showed mixed feelings about how they
felt creative in the creative writing conditions. 50% of students consider that they did
feel more creative in the Neutral environment, 54.8% in the Sound Environment and
63.6% in the Smell + Sound Environment. 72.7% of the students in the Smell Environ-
ment considered themselves more creative in this creative writing condition.
90.9% of the participants enjoyed to write in the Smell Environment, expressing
reasons such as: “It made me relax, without any pressure.”[smell_relax_P2]; “(...) be-
cause we were very excited and happy” [smell_relax_P5]; “(…) because I think it helps
to have more ideas” [smell_relax_P5]; “It was interesting - at first I was anguished
and then more relaxed.”[smell_relax_P10]; “it made me want to work.”[smell_re-
lax_P7]; “it allowed me to get immersed in the activity and be very creative. Although
I did not have enough time to finish my story, I managed to create a creative and funny
story, while not respecting some rules of grammar.”[smell_alert_P9]; “as the smell of
coffee is comforting and makes it easier to express my ideas.”[smell_alert_P7]; “(…)
because it allows a better flow of writing.”[smell_alert_P3]; “I think that it stimulated
my creativity, imposing it on the work” [smell_alert_P8].
In the Neutral Environment 57.1% of participants enjoyed writing in this environ-
ment as much as the 57.6% users in the Smell + Sound Environment. 77.4% of students
enjoyed to write in the Sound Environment, expressing “(…) because I felt completely
oblivious of what was happening around me and I was just concentrated on what I had
to do.”[sound_relax_P11]; “I was so focused writing that did not even notice the sounds
around me.”[sound_relax_P12]; “music gives a good environment for writing and is
relaxing.”[sound_alert_P9]; "(...) music has stimulated my creativity in writing"
[sound_alert_P3];
During the experience, we noted that participants were anxious and curious to know
what they were supposed to do. In a qualitative way, we also observed them focused

and in absolute concentration during the writing task in all conditions. According to the
interviews, some participants were deeply involved in the creative writing task.
5 DISCUSSION
This study aimed to explore which modalities, olfactory or auditory, were stronger trig-
gers for creativity.
We identify that most of the students considered themselves creative persons, even
though more than half of them were not usually engaged with creative activities, as
reported by the CBI. These results indicate that the study was not biased by a highly
critical self-assessment.
As stated by Csikszentmihalyi [35], every flow activity, whether they involved com-
petition or any dimension of experience, had in common the sense of discovery, the
creative feeling of transporting the person into a new reality. By giving a “writing chal-
lenge” we observed that students were immersed in the writing moment, and after they
finished the activity, they felt creative, animated, relaxed and peaceful. This was espe-
cially evident during the smell and sound cues and during the combination of both (the
smell + sound environment). From observation, in the Neutral Environment students
were feeling more apathetic and they made more pauses in writing during the timeout.
Although, through statistically analysis, we did not find significant results from the
Flow Theory dimensions, from our qualitative data we noticed that sound cues could
lead to lower levels of concentration. In addition, we noticed that they felt a little pres-
sure towards writing with a time limitation, and this was probably the reason for the
inexistence of substantial results in the Flow dimensions.
Triangulating the qualitative data from semi-structured interviews and results, indi-
cates that 72.7% of students said they felt more creative in the Smell Environment, in
contrast to other conditions that had higher values (50% in Neutral, 54.8% in Sound
and 63.6% in Smell + Sound Environment) as well as that 90.9 % of students enjoyed
writing in the Smell environment and 77.4% in the Sound environment.
CSI did show significant difference through statistical analysis, confirming the ben-
efit of olfactory and sound stimuli during a writing task. It is also important to observe
a decay on the CSI value for the combination of Alert smell+sound when compared to
the single conditions (only smell or sound), suggesting that such combination over-
whelmed the participants. Such disturbance is confirmed by the qualitative data results.
The different results from our study show a strong influence of smell and sound cues
during a creative writing task and with a highlight on olfactory cues. Therefore, an-
swering our main research question “Which modality, olfactory or auditory, sparks
stronger triggers for creativity?”. Such results become more relevant if we consider
that previous research did not show any significant effect of odours on performance, as
stated by Ho and Spence [17].
Our study suggests that olfactory cues should gain more relevance on the develop-
ment of creativity support tools and environments. Novel smell dispensers [39] and

their integration on current portable media devices as well as the olfactory augmenta-
tion of multisensory work environments could improve and foster creative tasks but
their combination with other stimuli must be designed carefully to avoid disturbance.
6 Conclusion
Creative writing is a constant activity in many sectors and professions in the modern
world. Because of today’s diversity of possible technological ways to write, designing
a creative writing user interface is hard work. Sometimes it is difficult to find a tool that
keeps users focused whilst eliminating some of the hard work. In this paper, we pre-
sented a study that addresses the use of different modalities such as smell and sound on
the creative writing process of users to enhance creativity.
In the between-subject study, we tested the influence of smell and sound on the par-
ticipant’s creativity in a writing task measured by the Creativity Support Index. We
have created two types of auditory and olfactory cues each: an alerting smell was cre-
ated using the actual fragrance of hot coffee that was spread around the room in small
cups. Regarding the alerting sound, users might feel that they would be sitting in a cafe
with the constant bustle of movement of people, machines and dishwashers. As for a
relaxing scent, we used the fragrance of laurel, and the sound mode used was a natural
soundtrack with water, birds and foliage.
We investigated which cue could promote higher levels of creativity and mental
well-being – as measured by a survey based on the dimensions of Flow Theory.
We compared four different environments (neutral, smell, sound and smell + sound)
using MS Word as a word processor. From a creative perspective, and triangulating
qualitative and the statistical results, it is suggested that users considered to feel more
expressive and more creative during the writing task, especially in the Smell Environ-
ment.
We found out that participants considered that the smell allowed them more creative
times, thus being one of the decisive factors to abstract and to become concentrated.
Our results provide interesting information regarding the smell and sound cues modal-
ities. At the same time, for the combination of smell and sound in each category, in
semi-structured interviews participants gave emphasis to these conditions in the crea-
tive moment of the writing task.
A significant problem faced by interaction designers that are involved in multisen-
sory interaction is the timing of each multisensory interaction (when to apply the mul-
tisensory stimulus). Writing prompts – among other techniques such as using images
like a map – are sometimes used to kick start the creative process, when writing. In this
research, we highlight the value of smell and sound cues as an alternative, more pow-
erful means to kick-start that same process more effectively.
There are many aspects of this study that remain open for future investigation such
as our results about creativity combined with different modalities. Measuring creativity
is an important approach that will lead us to different impacts of specific creative writ-
ing tools. These tools can have features to increase the creativity of writers and contrib-
ute to unblock writer’s block. In future work, we think it is important to intensify the

research on designing user interfaces that support creative writing and build novel tools
that can be used by several people.
ACKNOWLEDGMENTS
We would like to thank the students who participated in this research, as well as the
teachers and the institutional support from Escola da APEL. This work was partially
funded by ARDITI - Regional Agency for the Development of Research Technology
and Innovation through the M14-20 Project - 09-5369-FSE-000001- PhD Scholarship,
by FCT/MCTES LARSyS (UID/EEA/50009/2013 (2015-2017)), by IDE (ProCiência),
SENSE-SEAT, through the M1420-01-0247-FEDER-000001 and by ERA Chair: Grant
Agreement 621413.
References
1.
J. Kroll and G. Harper, Research Methods in Creative Writing, Palgrave
Macmillan, 2013.
2.
D. Morley, The Cambridge Introduction to Creative Writing, Cambridge:
Cambridge University Press, 2007.
3.
D. G. Myers, The Elephants Teach: Creative Writing since 1880., Chicago:
University of Chicago Press, 2006.
4.
D. Wang, “How People Write Together Now: Exploring and Supporting Today's
Computer-Supported Collaborative Writing,” in CSCW '16, Proceedings of the
19th ACM Conference on Computer Supported Cooperative Work and Social
Computing Companion, San Francisco, California, USA, 2016.
5.
E. Cherry and C. Latulipe, “Quantifying the Creative Support of Digital Tools
through the Creativity Support Index,” ACM Transactions on Computer-Human
Interaction, Vol. 21, N.º 4, Article 21, 2014.
6.
B. Shneiderman, “Creating Creativity: User Interfaces for Supporting Innovation,”
in ACM Transactions on Computer-Human Interaction, Vol. 7, N.º 1, 2000, pp.
114-138.
7.
W. Digest, The Writer's Guide to Creativity, F+W Media Inc., 2011.
8.
J. H. G. Kroll, Research Methods in Creative Writing, Palgrave Macmillan, 2013.
9.
B. Shneiderman, “Leonardo's laptop: Human Needs and the New Computing
Technologies,” MIT Press, Cambridge, 2002.
10.
E. Carrol and C. Latulipe, “Triangulating the personal creative experience: self-
report, external judments and physiology.,” In Proceedings the Graphics Interface

2012 (GI'12). Canadian Information Processing Society, Toronto, Ont. Canada.,
pp. 53-60, Canada, 2012.
11.
J. Millot, G. Brand and N. Morand, “Effects of ambient odors on reaction time in
humans.,” Neurosci, vol. 322, no. 2, pp. 79-82, 2002.
12.
H. Seo, V. Gudziol, A. Hähner and H. T., “Background sound modulates the
performance of odor discrimination task.,” Exp Brain Res. , vol. 212, no. 2, 2011.
13.
C. Spence, “Auditory contributions to flavour perception and feeding behaviour,”
Elsevier, Physiology & Behavior, vol. 107, no. 4, pp. 505-515, 2012.
14.
A. M. Magnifico, “"Getting other's perspectives": a case study of creative writing
environments and mentorship.,” in In Proceedings of the 9th Internacional
Conference of the Learning Sciences - Volume 1 (ICLS'10), 2010, pp. 1151-1157.
15.
A. Yannopoulos, “DirectorNotation: Artistic and technological system for
professional film directing,” J. Comput. Cult. Herit., 6, 1, , p. Article 2, 2013.
16.
M. D. A. Goulet, “Focus group on computer tools used for professional writing
and preliminary evaluation of LinguisTech.,” in In Proceedings of the Second
Workshop on Computational Linguistics and Writing (CLW 2012): Linguistic and
Cognitive Aspects of Document Creation and Document Engineering (EACL
2012), Stroudsburg, PA, USA, Linguistics, Association for Computational, 2012,
pp. 39-47.
17.
T. Coughlan and P. Johnson, “Understanding productive, structural and
longitudinal interactions in the design of tools for creative activities.,” In
Proceedings of the seventh ACM conference on Creativity and cognition
(C&C'09), pp. 155-164, ACM, New York, NY, USA., 2009.
18.
C. Ho and C. Spence, “Olfactory facilitation of dual-task performance,”
Neuroscience Letters, Elsevier , 2005.
19.
W. Xiang, S. Chen, L. Sun, S. Cheng and V. Bove, “Odor emoticon,” International
Journal of Human-Computer Studies, pp. 52-61 , 2016.
20.
M. Zampini and C. Spence, “The Role Of Auditory Cues In Modulating The
Perceived Crispness And Staleness Of Potato Chips,” Journal of Sensory Studies,
pp. 347-363, 2005.
21.
H. Seo and T. Hummel, “Influence of Auditory Cues on Chemosensory
Perception,” in The Chemical Sensory Informatics of Food: Measurement,
Analysis, Integration, American Chemical Society, 2015, p. 41–56.
22.
L. Demattè, D. Sanabria and C. Spence, “Olfactory Discrimination: When Vision
Matters?,” Chem. Senses 34, pp. 103-109, 2009.

23.
S. Guest, C. Catmur, D. Lloyd and S. C., “Audiotactile interactions in roughness
perception,” Experimental Brain Research, pp. 161-171, September 2002 .
24.
T. Ward, “Creative cognition as a window on creativity,” In Elsevier, Science
Direct, Methods 42 (2007), pp. 28-37, 2006.
25.
B. Shneiderman, “Accelerating Discovery and Innovation,” Communications of
the ACM, 50, 12, 2007.
26.
J. J. Elam and M. Mead, “Can Software Influence Creativity?,” Information
Systems Research, pp. 1-22, 1990.
27.
R. Hegde and G. Walia, “How to Enhance the Creativity of Software,” in
Proceedings of the 26th IEEE International Conference on Software Engineering
and Knowledge Engineering, Vancouver, Canada, 2014.
28.
T. Selker, “Fostering motivation and creativity for computer users,” International
Journal of Human-Computer Studies, vol. 63, no. 4-5, pp. 410-421, 2005.
29.
H. Wang, S. Fussell and D. Cosley, “From diversity to creativity: Stimulating
group brainstorming with cultural differences and conversationally-retrieved
pictures.,” in InProc. CSCW. ACM, 2011.
30.
S. Andolina, K. Klouche, D. Cabral, T. Ruotsalo and G. Jacucci, “InspirationWall:
Supporting Idea Generation Through Automatic Information Exploration.,” in In
Proceedings of the 2015 ACM SIGCHI Conference on Creativity and Cognition
(C&C´15) , New York, USA, 2015.
31.
L. Guzman, N. Kannan, K. Mendonza, S. Buenafe, N. Nunes, V. Nisi, P. Campos,
F. Gonçalves, M. Campos and P. Freitas, “Yarn: a Product for Unraveling
Stories.,” In NordiCHI'14 Industrial Program, Helsinki, 2014.
32.
F. Gonçalves, P. Campos, J. Hanna and S. Ashby, “You're the Voice: Evaluating
User Interfaces for Encouraging Underserved Youths to express themselves
through Creative Writing,” in C&C'15 Proceedings of the 2015 ACM SIGCHI
Conference on Creativity and Cognition, New York, USA, 2015.
33.
“Writers Digest,” [Online]. Available: http://www.writersdigest.com/qp7-
migration-books/writing-better-lyrics-excerpt.
34.
D. Hocevar, “The Development of the Creative Behavior Inventory,” in Annual
Meeting of the Rocky Mountain Psychological Association, 1979.
35.
D. Hocevar, “Measurement of creativity: Review and critique,” Journal of
Personality Assessment, vol. 45, no. 5, pp. 450-463, 1981.
36.
P. Silvia, B. Wigert, R. Reiter-Palmon and J. Kaufman, “Assessing Creativity
With Self-Report Scales : A Review and Empirical Evaluation,” Psychology
Faculty Publications, 2012.

37.
R. Wiseman, C. Watt and K. Gilhooly, “Creativity and ease of ambiguous figural
reversal,” British Journal of Psychology, vol. 102, pp. 615-622, 2011.
38.
M. Csikszentmihalyi, Flow: The Psychologic of Optimal Experience, New York:
Harper Perennial, 1991.
39.
J. Amores and P. Maes, “Essence: Olfactory Interfaces for Unconscious Influence
of Mood and Cognitive Performance,” in Proceedings of the 2017 CHI Conference
on Human Factors in Computing Systems, Denver, CO, USA, 2017.