Colors and Emotions in Video Games

Abstract

People experience emotions when playing videogames, and these emotions are a main reason for playing. In our research, we examine whether colors can be used in videogames to elicit specific emotions. In an experiment we used a videogame in which four different colors, associated with four specific emotions, were used in four different conditions. After each condition we measured the players' emotional responses by means of a Self-Assessment Manikin questionnaire. We found that the color red evoked a highly-aroused, negative emotional response, while the color yellow evoked a positive emotional response. These results were significantly different from the emotional responses measured for other colors. Furthermore, we found that inexperienced players showed much more explicit reactions to colors than experienced players. We conclude that the use of colors is a suitable method for game designers to elicit specific emotional responses from the players.

Full text

COLORS AND EMOTIONS IN VIDEOGAMES
Evi Joosten, Giel van Lankveld, and Pieter Spronck
Tilburg University / TiCC
P.O. Box 90153, NL-5000 LE Tilburg, The Netherlands
E-mail: evijoosten@gmail.com, g.lankveld@uvt.nl, p.spronck@uvt.nl
KEYWORDS
Emotions, videogames, Self-Assessment Manikin.
ABSTRACT
People experience emotions when playing videogames, and
these emotions are a main reason for playing. In our research,
we examine whether colors can be used in videogames to
elicit specific emotions. In an experiment we used a
videogame in which four different colors, associated with
four specific emotions, were used in four different
conditions. After each condition we measured the players’
emotional responses by means of a Self-Assessment Manikin
questionnaire. We found that the color red evoked a highly-
aroused, negative emotional response, while the color yellow
evoked a positive emotional response. These results were
significantly different from the emotional responses
measured for other colors. Furthermore, we found that
inexperienced players showed much more explicit reactions
to colors than experienced players. We conclude that the use
of colors is a suitable method for game designers to elicit
specific emotional responses from the players.
INTRODUCTION
“Emotions” are feelings, caused by persons or events, that
are experienced over a short period of time and that can
change rapidly (Robbins and Judge 2008). Experiencing
emotions tends to be the main reason for people to play
videogames (Ravaja et al 2004). Therefore, manipulating
emotions is a game designer’s prime concern, and playing a
game becomes an enjoyable experience if players experience
emotions that they find satisfying.
Psychological research has shown that music and color
affect people’s emotions. There is evidence that this holds in
particular for videogames (Livingstone and Brown 2005).
Stark et al (1982) found that colors influence emotions
experienced during play of a gambling game. Wolfson and
Case (2000) found a link between the colors red and blue,
and emotions experienced while playing a “breakout”-like
game. Previous research on the effect of colors on emotions
in videogames was, however, generally limited in three ways:
(1) only few colors were examined, (2) rather simple games
were used, and (3) the adaptation of colors in order to
influence emotions was considered to be outside the scope of
the research.
In the present research we examine to what extent colors
can influence a player’s emotions in a relatively complex
role-playing game. We do not measure emotions directly, but
instead focus on emotional responses, i.e., a player’s arousal
and enjoyment experienced during game playing.
We first provide background information on research into
emotions, emotional responses, emotions in video games, and
emotions and colors. We then outline our experimental setup
and describe and discuss results.
BACKGROUND
This section discusses how emotions can be measured, what
the most salient emotions in video games are, and to what
extent emotions can be evoked by the use of colors.
Measuring emotions
Theories of emotions state that humans are evolutionary
endowed with a limited set of basic emotions, namely anger,
fear, disgust, happiness, sadness, and surprise (Ekman 1993).
Each basic emotion is independent of the others in its
behavioral, psychological, and physiological manifestations,
and each arises from activation within unique neural
pathways of the central nervous system (Posner et al 2005).
One should therefore be able to measure basic emotions by
examining facial expressions and physiological responses. At
present, however, insufficient empirical foundation exists for
defining which emotions are basic and how they correlate to
what can be observed (Ortony and Turner 1990).
Emotions can be measured indirectly in terms of emotional
responses. Studies have repeatedly yielded two-dimensional
models of emotional responses (Larsen and Diener 1992). In
these models all affective states (emotions) are understood to
arise from common, overlapping, neuropsychological
systems (Posner et al 2005). One often-used model is the
‘circumplex model of affect’ (Russell 1980). It characterizes
emotions in terms of two dimensions of emotional responses,
namely (1) arousal and (2) valence. Arousal is the
physiological and psychological state of being proactive
(activation) or reactive (deactivation) to stimuli. Valence is
an intrinsic positive (pleasant) or negative (unpleasant)
feeling that is evoked by an event, object, or situation.
Russell posits that each affective state is the consequence of a
linear combination of these two independent dimensions
(Figure 1). The present research measures emotions based on
emotional responses (arousal and valence), derived from the
circumplex model of affect.
Figure 1: The circumplex model of affect (Russell 1980) with
valence on the horizontal and arousal on the vertical axis.

Salient emotions in videogames
Ravaja et al (2004) examined whether there are reliable
differences in the emotional response patterns elicited by
four videogames with diverse characteristics, namely: Tetris,
Super Monkey Ball 2, Monkey Bowling 2, and James Bond
007: NightFire. Emotions elicited by the games were defined
by arousal and valence in terms of six affective states,
namely: fear, anger, relaxation, pleasure, joy, and depression.
They showed that arousal and valence are two important
factors in enhancing emotions, especially in James Bond
007: NightFire. They demonstrated that arousal and valence
are suitable measurements to identify emotions in a
videogame.
Perron (2005) attempted to characterize some prototypical
videogame emotions. Based on film theory he identified the
following seven emotions: interest, enjoyment, worry, fear,
surprise, anger, and frustration. All these emotions score high
on the arousal level. Therefore, there is evidence that
videogames generate highly-aroused emotions.
Effect of color on emotions
Color is a high predictor for emotions (D'Andrade 1974). El-
Nasr et al (2006) measured the impact of changing color (in
terms of saturation, brightness, and warmth) and contrast on
tension felt in videogames. They reported a significant effect.
Their results suggest that the use of lighting patterns attaches
players emotionally to a game.
Wolfson and Case (2000) examined the effects of the
colors red and blue on players’ emotions. They used five
simple videogames where the color of the screen was
manipulated. The results suggested that participants in the
red group were more aroused than participants in the blue
group. Therefore they found effects on arousal by
manipulating color in videogames, though limited to the
colors red and blue. Furthermore, the simple games they used
can be considered less arousing compared to the more
complex videogames that are common today.
Plutchik (2001), Oberascher and Gallmetzer (2003), and
Valdez (1994) stated that every basic emotion can be linked
to a color. Plutchik (2001) linked colors to basic emotions as
shown in Table 1. Oberascher and Gallmetzer (2003)
confirmed Plutchik’s findings. Of the emotions used in
Plutchik’s research, the highly-aroused ones are surprise,
fear, joy, and anger. These should be prevalent in
videogames. Therefore our research focuses on their
associated colors: light blue, dark green, yellow, and red.
Emotion Color
Surprise Light blue
Fear Dark green
Acceptance Light green
Joy Yellow
Anticipation Orange
Anger Red
Disgust Purple
Sadness Dark blue
Table 1: Emotions and corresponding colors as identified by
Plutchik (2001). The colors and emotions in italics are those used
in the present research.
EXPERIMENTAL SETUP
For the purpose of this research, we designed and built a
videogame. In this game background colors were
manipulated. The emotional responses of players were
measured through a questionnaire, and analyzed. This section
describes the participants, the questionnaire, the game, and
the analysis method, respectively.
Participants
Students of the School of Humanities of Tilburg University
were approached to participate in our experiment. A total of
68 students signed up, 60 of which were used during the
analysis (the remaining 8 students either did not show up, or
failed to complete the game). 25 of the participants were
male, 34 female, and one did not specify gender. Their age
ranged from 18 to 31 years, with a mean age of 21.5 years.
Questionnaire
To measure players’ emotional responses a picture-oriented
instrument called Self-Assessment Manikin (SAM) was used
(Lang 1980). The SAM, displayed in Figure 2, has been used
to effectively measure emotional responses in a variety of
situations, including reactions to videos (which can be
compared to videogames). It is an easy, nonverbal (and
therefore language-independent) method for quickly
assessing people’s reports of emotional responses (Bradley
and Lang 1994). By asking participants to indicate which
image from a row of pictures best approaches their current
emotional state, the SAM directly assesses the valence,
arousal, and dominance associated with a response to an
object or event. The arousal and valence dimensions are as
discussed in the Background section. The dominance
dimension is not related to an emotional response, but is used
for indexing the relationship of control that exists between
the perceiver and the perceived situation. Scores ranged from
1 to 9 (participants could also indicate a value between two
neighboring pictures). For the first row, 9 indicated “most
aroused” (left-most picture). For the second row, 9 indicated
“most pleasant” (left-most picture). For the third row, 9
indicated “most in control” (right-most picture).
Arousal
Valence
Dominance
Figure 2: Self-Assessment Manikin.

Game
For the experiment we built a game as a Neverwinter Nights
module using BioWare’s Aurora toolset. The game starts
with a short introduction that introduces the participant to the
game controls. After that he encounters five conditions
(situations). In each condition the player enters a different
room, in which he has an interaction with a non-player
character (NPC), and must collect an item, which he needs to
open the next room. When all five items are collected, the
participant can open a treasure chest and “win” the game.
After leaving each of the rooms, the game asks him to fill out
the SAM questionnaire. After the game is won, he is asked to
fill out a short questionnaire, which includes a question on
his experience with videogames.
The visual characteristics of the five conditions do not
change in the game, apart from the ambient light color. The
first condition is the control condition in which no
manipulation takes place. In the other four conditions the
color is manipulated. Color is therefore our independent
variable. In order to prevent carry-over effects of the
response set, the order of the colors of the four manipulated
conditions are counterbalanced completely. As there are 24
different orderings for the four colors used in the experiment,
our 60 participants were sufficient to test each ordering two
or three times.
By manipulating the conditions’ colors our aim was to
stimulate feelings of surprise, fear, joy, and anger, which,
according to Plutchik (2001), are associated with the colors
light blue, dark green, yellow, and red (see Table 1). These
feelings all should score high on the arousal dimension.
However, whereas surprise and joy (light blue and yellow)
should score high on the valence dimension, fear and anger
(dark green and red) should score low on the valence
dimension.
Analysis
To analyse the results of the measured emotional responses, a
one-way ANOVA, an independent-samples T-test, and
descriptive statistics were performed using SPSS 16.0. The
scores on the SAM were the three dependent variables:
arousal, valence, and dominance. Descriptive statistics of
these variables were performed over the control condition,
the rooms, and the four colors.
Descriptive statistics were also used to measure the number
of participants, the mean score of the SAM dimensions, and
the standard deviation for the four rooms, the control
condition, and the four colors. A manipulation check was
performed to analyze the differences between the rooms
(conditions) by a one-way ANOVA. To test whether there is
a main effect for color (including the control condition) on
the arousal, valence, and dominance scores a one-way
ANOVA was performed. We also examined whether
experience or inexperience with videogames had an effect on
emotional responses using one-way ANOVA tests.
RESULTS
In this section we discuss the SAM scores for the conditions
(rooms), for the colors (and thus the emotions), and for the
differences between experienced and inexperienced
participants, respectively. Note that exact numerical details
of the results are given by Joosten (2010).
SAM-scores for the rooms
The mean scores on the SAM for the four manipulated rooms
are presented in Figure 3. From the results we conclude that a
significant effect for the rooms was found on the valence and
dominance scores. Scores for valence increased with the
room numbers, with significant differences between rooms 1
and 4. Scores for dominance increased with the room
numbers, with significant differences between rooms 1 and 3,
rooms 1 and 4, and rooms 2 and 4. This means that, on
average, a player’s experience of pleasure and feeling of
control increased while the game progressed.
Figure 3: Mean scores on arousal, valence, and dominance for the
four manipulated rooms.
Figure 4: Mean scores on arousal, valence, and dominance for the
control condition and the colors.
SAM-scores for the color conditions
The mean scores on the SAM for the four color conditions
are presented in Figure 4. From the results we conclude that,
on average, arousal scores for the color red were significantly
higher than for any of the other colors or the control
condition. There were no significant differences on arousal
for the other colors. On average, valence scores for the color
yellow were significantly higher than for the colors dark
green and red, and for the control condition. There was no
significant difference between the valence scores for yellow
and light blue. Neither was there a significant difference for
light blue and either dark green or red, or the control
condition. The valence score for red, however, was
significantly lower than those for yellow, light blue, or the

control condition. For dominance no significant effects were
found for any of the colors. This means that, on average,
more so than any other color used in the experiment, the
color red arouses a player, but is experienced as negative.
The color yellow, more than any other color used in the
experiment, is experienced as positive. These results are
statistically significant.
SAM-scores and experience
When examining our data, we hypothesized that there were
differences in scores between participants who considered
themselves experienced videogame players, and participants
who considered themselves inexperienced in this area. We
therefore calculated the mean scores on the SAM for the four
color conditions for these two groups separately. The results
are presented in Figures 5 and 6.
From the results we conclude the following: For the colors
dark green, red, and the control condition the 38
inexperienced participants scored significantly lower on
valence than the 21 experienced players (one participant did
not answer the question on experience), i.e., they derived less
pleasure from these conditions. For the colors yellow, red,
and the control condition the dominance scores of
inexperienced participants were also significantly lower than
for experienced participants, i.e., they felt less in control in
these conditions.
For experienced participants we find no significant
differences between the scores on any of the colors for any of
the dimensions. In contrast, for inexperienced participants we
find three significant effects. First, the arousal score for the
color red is significantly higher than the scores for any of the
other colors, or the control conditions. One might even call
the score ‘surprisingly high.’ Second, the valence score for
the color yellow is significantly higher than the scores for the
colors dark green, red, or the control condition. There was no
significant difference with the color light blue, although the
score for yellow is still quite a bit higher than the score for
light blue. Third, the valence score for red was significantly
lower than all the other colors or the control condition. For
dominance, no significant effects are found.
The somewhat surprising conclusion that we must draw is
that the effects we found were mainly the result of the
answers given by the inexperienced participants.
Figure 5: Mean scores on arousal, valence, and dominance of
experienced videogame players for the control condition and the
colors.
Figure 6: Mean scores on arousal, valence, and dominance of
inexperienced videogame players for the control condition and the
colors.
DISCUSSION
In this section we discuss our findings on the effects of the
rooms, the colors, and player experience on emotional
responses.
Effects of the rooms on emotional responses
Results on the differences between the rooms showed an
effect on the valence and dominance dimensions. Differences
on the valence scores between the rooms may be due to the
enjoyment participants felt on winning the game. Participants
felt more enjoyment in the last condition than in the first
condition. Holbrook et al (1984) state that performance is an
important factor for enjoyment in videogames.
The reason for the differences on the dominance scores
between the rooms may be due to the fact that the
participants got more experienced during the game, therefore
navigation became easier (Griffith et al 1983) and
participants felt more in control. This explanation is
supported by the fact that experienced participants showed
significantly higher dominance scores than inexperienced
participants.
Effects of the colors on emotional responses
While we found effects on emotional responses for the colors
yellow and red, we failed to find effects for the colors dark
green and light blue. We provide three possible reasons for
this lack of effect.
First, the lack of effect may be due to the environment used
to measure emotional responses. The natural environments
used by Plutchik (2001), who identified the relationship
between colors and emotional responses, differ from ours. It
is very well possible that color manipulation in a videogame
leads to different emotional response effects than in a natural
environment.
Second, the lack of effect may be due to the limited options
for colour selection in the Aurora toolset. The colors in the
light blue and dark green condition were less intense than the
colours in the red and yellow condition. Valdez and
Mehrabian (1994) stated that colors of high intensity elicit
stronger emotional effects in humans than those of low
intensity. Furthermore, the color of the light blue condition
was almost identical to the color of the control condition, and
the color dark green made the condition very hazy, which

might have a stronger effect on how a player experiences the
game than just the color effect.
Third, the lack of effect may be due to the fact that in this
experiment only the background light of a condition was
manipulated. The objects in the condition and the condition
itself were in standard colors.
Effects of experience on emotional responses
Emotional response effects for videogame experience were
found on the valence and the dominance scores. In general
experienced participants scored higher on these dimensions
than inexperienced players. The higher scores for the
experienced participants may be due to the perceived
complexity of the videogame. Players who perceive less
complexity during a videogame feel more enjoyment and
have higher feelings of dominance (Holbrook et al 1984).
Emotional response effects for color were only found for
the inexperienced participants. The reason for the lack of
emotional response effects of color for the experienced
participants may be due to the fact that they are familiar with
videogame situations. They know what to expect and how to
navigate through the game; their experience guides them
more than the colors they see. Jørgensen (2008) stated that
videogame players who are unfamiliar with a videogame
situation use visual aspects (including color) of the
videogame to make sense of the situation. Therefore, they
take more note of visual cues, even subconsciously.
A further explanation for the lack of emotional response
effects for experienced participants may be the time they
spent in the conditions. In general, they were very fast in
playing the game. Landers and Boutcher (1993) found that
the duration of a videogame is an important factor in eliciting
emotional responses.
CONCLUSION
We investigated to what extent the use of colors in a
relatively complex videogame may influence a player’s
emotional responses. We found significant effects on
emotional responses for the colors red and yellow. According
to Plutchik (2001), the color red is associated with a feeling
of anger, which evokes a highly-aroused, negative emotional
response, and the color yellow is associated with a feeling of
joy, which evokes a highly-aroused, positive emotional
response. Indeed, in our experiment we found that red
elicited a highly-aroused, negative emotional response, and
yellow elicited a positive emotional response. We found
these color effects prevalent mainly with inexperienced
videogame players. We conclude that game designers can
employ certain colors to manipulate player’s emotions,
specifically those of novice players.
REFERENCES
Bradley, M.M. and Lang, P.J. 1994. “Measuring emotion: The Self-
Assessment Manikin (SAM) and the semantic differential.”
Journal of Experimental Psychiatry & Behavior Therapy, 25,
49-59.
D'Andrade, R., and Egan, M. 1974. “The Color of Emotions.”
American Ethnologist, 1(1), 49-63.
Ekman, P. 1993. “Facial Expression and Emotion.” American
Psychologist, 48, 384-392.
El-Nasr, M., Niedenthal, S., Knez, I., Almeida, P., and Zupko, J.
2006. “Dynamic lighting for tension in games.” Game Studies,
7(1).
Griffith, J.L., Voloschin, P., Gibb, G.D., and Bailey, J.R. 1983.
“Differences in eye-hand motor coordination of video-game
users and non-users.” Perceptual and Motor Skills, 57(1), 155-
158.
Holbrook, M.B., Chestnut, R.W., Oliva, T.A., and Greenleaf, E.A.
1984. “Play as a Consumption Experience: The Roles of
Emotions, Performance, and Personality in the Enjoyment of
Games.” Journal of Consumer Research, 11(2), 728-739
Joosten, E. 2010. The Emotions of Colour in Videogames. MA
thesis, Tilburg University, Tilburg, The Netherlands.
Jørgensen, K. 2008. “Audio and gameplay: an analysis of PvP
battlegrounds in World of Warcraft.” Game Studies, 8(2).
Landers, D.M. and Boutcher, S.H. 1993. “Arousal-performance
relationships.” Applied sport psychology, Personal growth to
peak performance (J.M. Williams, ed.), 2
nd
edition, Mayfield
Publishing Company, Palo Alto, CA.
Lang, P.J. 1980. “Behavioral treatment and bio-behavioral
assessment: Computer applications.” Technology in mental
health care delivery systems, 119-167, Ablex, Norwood, NY.
Larsen, R.J. and Diener, E. 1992. “Promises and problems with the
circumplex model of emotion.” Review of personality and
social psychology, 13, 25-59.
Livingstone, S.R. and Brown, A.R. 2005. “Dynamic response: real-
time adaptation for music emotion.” Proceedings of the Second
Australasian Conference on Interactive Entertainment, 105-
111.
Oberascher L. and Gallmetzer M. 2003. “Colour and emotion.”
Proceedings of AIC 2003 Bangkok: Color Communication
Management, 370-374.
Ortony, A., and Turner, T.J. 1990. “What's basic about basic
emotions?” Psychology Review, 97, 315-331.
Perron, B. 2005. “A cognitive psychological approach to gameplay
emotions.” Conference Proceedings of DiGRA’05, Changing
Views: Worlds in Play, Digital Games Research Association.
Plutchik, R. (2001). The nature of emotions. American Scientist,
89, 344-350.
Posner, J., Russell, J.A., & Peterson, B.S. 2005. “The circumplex
model of affect: An integrative approach to affective
neuroscience, cognitive development, and psychopathology.”
Development and Psychopathology, 17, 715-734.
Ravaja, N., Salminen, M., Holopainen, J., Saari, T., and Laarni,
J.J.A. 2004. “Emotional response patterns and sense of
presence during videogames: potential criterion variables for
game design.” Proceedings of the Third Nordic Conference on
Human-Computer Interaction, 339-347.
Robbins, S.P. and Judge 2008. Organizational Behavior. Prentice
Hall International, Upper Saddle River, NJ.
Russell, J.A. 1980. “A circumplex model of affect.” Journal of
Personality and Social Psychology, 39, 1161-1178.
Stark, G.M., Saunders, D.M., and Wookey, P.E. 1982. “Differential
effects of red and blue coloured lighting on gambling
behavior.” Current Psychological Research, 2, 95-100.
Valdez, P. and Mehrabian, A. 1994. “Effects of colour on
emotions.” Journal of Experimental Psychology, 123, 394-409.
Wolfson, S. and Case, G. 2000. “The effects of sound and colour on
responses to a computer game.” Interacting with computers,
13(2), 183-192.