Engineering, 2013, 5, 189-194
http://dx.doi.org/10.4236/eng.2013.510B040 Published Online October 2013 (http://www.scirp.org/journal/eng)
Positive-Negative Emotional Categorization of Clothing
Color Based on Brightness
Xiao-Feng Jiang1,2, Xiao-Pei Bian1
1College of Textile and Clothing Engineering, Soochow University, Suzhou, China
2National Engineering Laboratory for Modern Silk, Soochow University, Suzhou, China
Received May 2013
ABSTRACT
In current study, behavioral measures were conducted to investigate clothing color. The purpose was to focus on the
rule that color brightness influenced positive-negative emotional categorization. Results showed that the effect of
brightness on clothing color emotion categorization was significant. With the increase of brightness, the variation curve
of positive emotion appears to be a “U-shaped”, whereas that of the negative emotion shows an upside down “U-
shaped”. Compared with the low brightness colors, the emotion reaction to the high brightness colors was more positive;
Most of the colors with different brightness scales were classified as positive emotions and the minors were classified as
negative emotions; the positive colors could be done much faster than the negative ones.
Keywords: Color Emotion; Response Time; Brightness; Categorization; Positive-Negative
1. Introduction
Color is ubiquitous in our perceptual experience of the
world when we encounter objects like clothing items. It
is widely recognized that color is an essential ingredient
in the perception of fashion, since it plays an imperative
role in wearer’s emotional expression and the image
building on various occasions. For example, for sports-
wear, vivid and medium brightness are more used in
representing active and vigorous activities. For under-
wear, colors of high brightness are usually chosen to
keep in harmony with the relaxed conditions and to give
relatively comfortable and simple impression. However,
for working uniform, the occupational characteristic is
mainly emphasized, and it is necessary to select the col-
ors of lower brightness for office clothing in order to
show solemnities [1-3].
The brightness can be changed by adding black or
white, which may evoke dissimilar human feelings [4].
Much research on brightness stimuli of color indicated
that individuals associate colors with emotions, for ex-
ample, a color of higher brightness will evoke light and
pure feelings; that of medium brightness may be easily
perceived as an implicit and featureless sense; that of
lower brightness represents mysterious and depressed
sensation, etc. [5-7]. These feelings, evoked by either
colors or color combinations, are called color emotions.
However, the relation between color and emotion is very
intricate, because it may be influenced by the factors
such as genders, cultural backgrounds and so on. Never-
theless, some researchers found that the influence of cul-
tural background was very limited, whereas brightness
and saturation were the most important factors influen-
cing color emotion. For example, J. H. Xin et al. investi-
gated cross-regional color emotions, and found that the
hue was much less predominant for color emotions than
brightness and saturation [8].
It is notable that color emotions can also be divided
into the positive and negative parts in line with the
arousal by different colors. For instance, Red has both
positive and negative impressions such as passionate,
active, and conversely raging and offensive. Blue also
has both positive and negative impressions such as faith-
ful and calm, but on the other hand impassive and de-
pressed. Shirley Willett’s color codification of emotions
intuitively suggested the positive-negative emotions of
colors. That is, the outside circle includes positive traits
and the inside circle indicates negative ones [9]. Some
researchers believed that positive-negative emotions of
color are typically expressed with semantic word pairs,
such as “vivid/sombre”, “gaudy/plain”, “striking-sub-
dued” etc [10]. Except for the hue of colors, some studies
aimed at the brightness, from which the effects of posi-
tive-negative emotions also have been found. Boyatzis
and Varghese reported that bright colors evoke mainly
positive emotional associations, while dark colors evoke
negative emotional associations [11]. Numerous re-
searches focus on the relation between emotion and the
color perceived by human beings. Unfortunately, most of

X.-F. JIANG, X.-P. BIAN
190
them have used only square color patches in the experi-
ments and do not consider any effect of shape [12,13],
while a few studies used contextualized colors such as
full scale simulations of clothing. In fact, these investiga-
tive results were unrealistic to apply to clothing [14].
Furthermore, only few researches have provided experi-
mental evidence on positive-negative emotional catego-
rization of clothing color based on human behavior. Thus,
in this study, clothing as stimuli were selected, and the
behavioral cognition of colors based on brightness was
evaluated using psychology software tools. This study
has three main objectives as follows:
1) To investigate the influence of different brightness
on the positive-negative emotional categorization;
2) To explore the rule of positive-negative emotional
categorization;
3) To evaluate the time course during the categoriza-
tion.
2. Experiment 1
2.1. Participants
50 undergraduate students from Soochow University (18
males, 32 females, aged in 20 - 22) participated in the
experiment. All the subjects reported that they are right-
handled and they have normal corrected vision. The par-
ticipants were tested individu ally.
2.2. Stimuli
60 words were picked out from a large number of litera-
tures to describe colors as stimu li and were formed into
30 word pairs preliminarily.
2.3. Procedure
The experiment consisted of two phases. In one phase the
suitability of words were evaluated and in another phases
the positivity and negativity of words were evaluated. In
Phase 1, each trial had only one word, and it began with
the presentation of a fixed cross in the center of the
screen for 100 ms. After 400 ms, a word was displayed
for 300 ms on a white background randomly. After that,
the participants must evaluate the suitability or positivi-
ty-negativity of the word (show n in Figure 1). Responses
were collected via keyboard. Before the experiment, 10
trial words were used to acquaint with the manipulative
method by the participants. Then, an instruction page
informed the participants that they had done with the test
and that the phase 1 was about to begin. The stimuli and
the procedure in Phase 2 were the same as those in Ph ase
1.
2.4. Results and Discussion
The results are shown in Figure 1. When the acceptance
Figure 1. Procedure of presenting stimuli.
rate is above 50%, it means that more than half of the
subjects thought that the evaluated word is suitable or
positive. If it is below 50%, it means that half of them
thought that the evaluated word is unsuitable or negative.
The picked words show in Figure 2. It is obvious that the
acceptance rate of each admitted positive word was much
larger than 50%, while the admitted negative word was
smaller than 50%. The seven word pairs chosen from the
measurement were “light/dark”, “warm/cool”, “vivid/
sombre”, “graceful/vulgar”, “vigorous/weary”, “high-
class/low-class” and “dynamic/passive”.
The behavioral test in this experiment answered prop-
erly what we concerned. That is, which words are more
suitable to be used to describe clothing colors and which
words are positive and negative? Therefore, we are firm-
ly convinced that these word pairs assessed from the ex-
periments are effective to be used as stimuli in Experi-
ment 2. Behavioral test has some advantages. On the one
hand, it makes subjects decide the best words, which
reflected the subjects’ objective cognitive level of cloth-
ing colors and avoided our subjective intention to pick
out color words. On the other hand, each word had an
independent test, which was not subject to interference
from the other words.
3. Experiment 2
3.1. Participants
38 undergraduate students who have already finished
experiment 1 (14 males, 24 females, aged in 20 - 22)
participated in th is experiment, whose majors have noth-
ing to do with clothing. All subjects had normal or cor-
rected-to-normal vision and were unaware of the purpose
of the experiment.
3.2. Equipment and Stimuli
The stimuli were presented on a 17-inch monitor con-
nected to a 2 GHz Pentium computer in a dimmed room,
the distance between participants and monitor was about
60 cm, and the visual angle was 12.3˚ × 4.9˚, using
E-Prime 2.0 (Psychology Software Tools, Inc.). In this
experiment, the color images were used as stimuli. Pic-

X.-F. JIANG, X.-P. BIAN
191
Figure 2. Acceptance rates of the seven word pairs.
tures were produced with the Photoshop 7.0 procedure.
Five basic colors were selected evenly from the 360˚
color wheel at the interval of 72 ˚. Then, according to the
hue/saturation pattern, every basic color was divided into
9 sub-colors with different brightness (change interval:
18), in which the brightness at level 1 was lowest while it
was highest at level 9. Finally, all these colors were pre-
sented in clothing (shown in Figure 3). All the colors are
based on the RGB color mode.
3.3. Procedure
There were 45 pictures used as stimuli. The experiment
included 630 trials and was divided into three phases for
alleviating the subjects’ fatigue. Each trial had only one
word and one color picture, and it began with the pres-
entation of a fixed cross in the center of the screen for
100 ms. After 400 ms, a word was displayed for 300 ms
on a uniform white background, and then followed by a
color image which appeared randomly. After that, the
participants must evaluate the emotion of the color of the
image (shown in Figure 4). Responses were collected via
keyboard. The subjects pressed “A” in keyboard if he
thought the color emotion was in accordance with the
word. Otherwise, he should press “L” in keyboard. Be-
fore the experiment, 10 trial images were also provided
to acquaint with the manipulative method. Then, an in-
struction page informed the participants that they had
done with the test and that the phase 1 was about to begin.
The stimuli and the procedure in Phase 2 and Phase 3
were the same with those in Phase 1.
3.4. Results and Discussion
3.4.1. Differences of Positive-Negative Emotional
Categorization
Figure 5 presents the mean categorization accuracy of
positive-negative emotions for seven word pairs. In gen-
eral, In 630 colors, about two-thirds of them (M = 0.59,
MSE = 0.02) were classified as the traits of positive
emotion. On the contrary, only around one-thirds of them
(M = 0.31, MSE = 0.02) were classified as the traits of
Figure 3. The sample of brightness changes for clothing.
Figure 4. Procedure of presenting stimuli.
Figure 5. Differences of positive-negative emotional catego-
rization among the word pairs.
negative emotion. The data were fed into multiple-factor
repeated-measures ANOVA, with Type (positive, nega-
tive) × Brightness (1-9). There was a significant main
effect of type [F (1, 37) = 113.37, p < 0.00], implying
that participants deemed that most of the colors are posi-
tive. In addition, there is also a significant interaction
0%
20%
40%
60%
80%
100%
120%
light
dark
warm
c ool
vivid
sombre
grace ful
vulgar
vigorous
weary
high-class
low-class
dynamic
passive
suitable positive-negative

X.-F. JIANG, X.-P. BIAN
192
effect between type and word pair. However, there was
no significant main effect of word pair [F (6, 222) = 1.85,
p = 0.11], suggesting that the differences of emotional
categor i z a tion are not re markabl e among the w or d pa i rs.
3.4.2. Rules of Positive-Negative Emotional
Categorization Based on Brightness
Mean categorization ratings for colors are plotted in
Figure 6. There was a significant main effect of bright-
ness [F (2, 296) = 23.05, p < 0.00], implying that there
were prominent differences of emotional categorization
among various clothing brightness.
Apparently, with the increase of brightness, the varia-
tion curve of negative emotion shows a “U-shaped”,
whereas that of the positive emotion shows an upside
down “U-shaped”. That is, they presented a pattern that
the positive emotions gradually weakened after enhanc-
ing according to brightness scales from 1 to 9, the lower
brightness colors had weaker positive emotions, but the
medium brightness colors had stronger positive emotion.
But for negative emotion, it shows the inverse trend. In
fact, the “U-shaped” is not asymmetrical. Compared with
the low brightness colors, the high brightness colors were
easier to be categorized as the positive emotions. In addi-
tion, there is also a significant interaction effect among
the type, word pair and brightness, [F (48, 1776) = 9.04,
p < 0.00]. In order to investigate the emotional categori-
zation for every word pair, we conducted an ANOVA of
Type (positive, negative) × Brightness (1 - 9) with re-
peated measurement, respectively. The results show that
the main effects of brightness are very significant for
seven word pairs, as shown in Table 1.
3.4.3. RTs of Positive-Negative Emot ional
Categorization
The response times (RTs) for positive and negative emo-
tions are shown in Figure 5. There was a significant
main effect of type [F (1, 37) = 61 .01, p < 0. 00], indicat-
ing that significant differences of RTs existed between
the two emotional types during the Participants catego-
rizing them. Compared with categorizing the negative
emotions (M = 736.7 ms, MSE = 23.1 ms), categorizing
the positive emotions (M = 677.8 ms, MSE = 19.4 ms)
were faster. This evidence revealed a fact that it is diffi-
cult to classify negative emotions. For the word pairs,
there was also a significant main effect in RTs [F (6, 222)
= 7.10, p < 0.00]. It suggests that the differences of RTs
were distinct among the word pairs. For example, the
speed of emotional categorization after the word pair
“light/dark” Primi n g (M = 690.7 ms, MSE = 20.4 ms)
was faster than that of emotional categorization after the
word pair “dynamic/passive” priming (M = 728.1 ms,
MSE = 24.3 ms).
We also used a two-way ANOVA of Type (positive,
Figure 6. Tendency of positive-negative emotional categori-
zation based on brightness.
negative) × Brightness (1 - 9) with repeated measurement
respectively so as to verify this phenomenon. The results
show that the main effects are very significant for all the
seven word pairs. As shown in Table 2. Further results
confirm the fact that categorizing the positive emotions is
easier than categorizing the negative emotions.
The classification results of seven words consistently
reflected a rule. That is, no matter what the brightness is,
the subjects responded more quickly in classifying the
colors as positive emotion than as negative emotion.
There are many explanations about this phenomenon.
The reasonable ones are as follows: Firstly, based on the
theory of H. Bless [15], the participants were more likely
to adopt the heuristic strategy in a positive mood, which
is a top-down way of processing information, and more
depends on the color schema in the brain and ignores the
details. On the contrary, the participants were more likely
to adopt the strategy in a negative mood. That is a bot-
tom-up way of processing , which less depends on the
color schema in the brain and gives more attention to the
details. Therefore, the time is less in matching the color
schema than in processing the details. Accordingly, it
prolonged the response time.
Secondly, there are two methods for handling envi-
ronment or stimuli for human. They are behavioral ap-
proach and behavioral avoidance. When facing with a
positive environment or stimulus, he can respond with
approach, while when facing with a negative environ-
ment or stimulus, he may makes a response with avoid-
ance, which must have triggered a delayed-response [16].
According to this theory, the participants spontaneously
adopt the way of approach to complete the classification
task after the onset of positive emotion, while the avoid-
ance is used after the onset of negative emotion, which
would increase the response time. Based on affective-
matching theory, if the affective word is in line with the
evaluation of target stimuli, rationality which promotes
positive response will be added. Otherwise, it will create
a sense of unreasonable, repressing the response [17].

X.-F. JIANG, X.-P. BIAN
193
Table 1. Main effect of emotional categorization for the seven word pairs.
Source Sum of squares Mean square F Sig. Partial eta squared
Light/dark 18.03 18.03 141.61 0.000 0.79
Warm/cool 3.27 3.27 20.44 0.000 0.36
Vivid/sombre 14.94 14.94 115.63 0.000 0.76
Graceful/vulgar 21.80 21.80 88.36 0.000 0.70
Vigorous/weary 9.81 9.81 44.03 0.000 0.54
High-class/low-class 20.95 20.95 114.25 0.000 0.76
Dynamic/passive 17.37 17.37 90.36 0.000 0.71
Table 2. Main effect for RTs of emotional categorization for the seven word pairs.
Source Sum of squares Mean square F Sig. Partial eta squared
Light/dark 527913.44 527913.44 23.83 0.000 0.39
Warm/cool 706203.19 706203.19 18.89 0.000 0.34
Vivid/sombre 281005.11 281005.11 17.31 0.000 0.32
Graceful/vulgar 464731.30 464731.30 13.34 0.001 0.26
Vigorous/weary 566694.10 566694.10 15.90 0.000 0.30
High-class/low-class 686642.96 686642.96 28.25 0.000 0.43
Dynamic/passive 1052850.04 1052850.04 35.32 0.000 0.49
Figure 7. RTs of positive-negative emotional categorization
for the seven word pairs.
Figure 8. RTs of positive-negative emotional categorization
based on bright ness.
Finally, in this experiment, the probability of positive
and negative words was the same, but the subjects pre-
ferred to classify most of color as positive, which has
been mentioned in the previous discussion. So the con-
flicts were less between positive words and clothing col-
or than those between negative words and clothing color,
resulting in longer RTs. This is in line with the Stroop
effect and we believe this explanation is most persuasive
[18].
4. Conclusion
Several significant findings have been revealed from the
behavioral. Firstly, the effect of brightness on clothing
color emotion categorization was incredibly remarkable
and stable. With the increase of brightness, the variation
curve of positive emotion presented a “U-shaped”, whe-
reas that of the negative emotion showed an upside down
“U-shaped”. And the emotion reaction to the high bright-
ness colors was more positive, compared to the low
brightness colors. Secondly, most of the colors with dif-
ferent brightness levels were classified into the group of
positive colors and the minors were classified into the
group of negative colors. Finally, categorizing the posi-
tive colors was much faster than categorizing the nega-
tive ones.
5. Acknowledgements
This work was funded by A Project Funded by the Prior-
ity Academic Program Development of Jiangsu Higher

X.-F. JIANG, X.-P. BIAN
194
Education Institutions (PAPD), and was supported by
JSNSF Grant (No. BK2012196).
REFERENCES
[1] A. Naitou and S. Kobayashi, “A Dressing Image and a
Color Effect of Women’s Suits,” Journal of the Japan
Research Association for Textile End-Use, Vol. 43, No.
10, 2002, pp. 52-62.
[2] C. Mizutani, S. Otsuki, C. Ued and S. Park, “A Compari-
son of the Relationship between Color of Wearing Clothes
and Fashion Consciousness in Japan and Korea,” Journal
of Textile Engineering, Vol. 54, No. 1, 2008, pp. 9-14.
http://dx.doi.org/10.4188/jte.54.9
[3] K. Ito, “Color Affection Produced by Two-Color Combi-
nations Appearing in Fashion Magazines,” Journal of the
Japan Research Association for Textile End-Use, Vol. 48,
No. 11, 2007, pp. 32-41.
[4] T. Nakamura, T. Sato and K. Teraji, “Arrangement of Co-
lour Image Words into the Non-Luminous Object Colour
Space,” Journal of Color Science Association of Japan,
Vol. 18, 1994, pp. 10-18.
[5] X. P. Gao, et al., “Analy sis of Cross-Cultural Color Emo-
tion,” Color Research and Application, Vol. 32, No. 3,
2007, pp. 223-229.
http://dx.doi.org/10.1002/col.20321
[6] L. C. Ou and M. R. Luo, “A St udy of Color Emotion and
Color Preference. Part I: Color Emotions for Single Col-
ors,” Color Research and Application, Vol. 29, No. 3,
2004, pp. 232-240. http://dx.doi.org/10.1002/col.20010
[7] X. P. Gao and J. H. Xin, “Investigation of Human’s Emo-
tional Responses on Colors,” Color Research and Appli-
cation, Vol. 31, No. 5, 2006, pp. 411-417.
http://dx.doi.org/10.1002/col.20246
[8] J. H. Xin, K. M. Cheng, G. Taylor, T. Sato and A. Han-
suebsai, “Cross-Regional Comparison of Color Emotions.
II. Quantitative Analysis,” Color Research and Applica-
tion, Vol. 29, No. 6, 2004, pp. 458-466.
http://dx.doi.org/10.1002/col.20063
[9] Shirley Willet Color Codification.
http://home.earthlink.net/~smwillett/color of emo tions.html
[10] A. Naitou and S. Kobayashi, “A Dressing Image and a
Color Effect of Women’s Suits,” Journal of the Japan
Research Association for Textile End-Use, Vol. 43, No.
10, 2002, pp. 52-62.
[11] C. F. Boyatzis and R. Varghese, “Children’s Emotional
Associations with Colors,” Journal of Genetic Psycholo-
gy, Vol. 155, 1994, pp. 77-85.
http://dx.doi.org/10.1002/col.20063
[12] L.C. Ou, M. R. Luo, A. Woodcock and W. Angela, “A
Study of Colour Emotion and Colour Preference. Part II:
Colour Emotions for Two-Colour Combinations,” Colour
Research and Application, Vol. 29, 2004, pp. 292-298.
[13] A. C. Moore, A. K. Romney and T. L. Hsia, “Cultural,
Gender, and Individual Differences in Perceptual and
Semantic Structures of Basic Colors in Chinese and Eng-
lish,” Journal of Cognition and Culture, Koninklijke Brill
NV, Leiden, 2004, pp. 1-28.
[14] T. W. Whitfield and T. J. Wiltshire, “Color Psychology:
A Critical Review,” Genetic, Social, and General Psy-
chology Monographs, Vol. 116, No. 4, 1990, pp. 385-
411.
[15] H. Bless, G. L. Clore, N. Schwarz, V. Golisano, C. Rabe
and M. Wölk, “Mood and the Use of Scripts: Does a
Happy Mood Really Lead to Mindlessness?” Journal of
Personality and Social Psychology, Vol. 71, No. 4, 1996,
pp. 665-679.
http://dx.doi.org/10.1002/col.20063
[16] R. J. Davidson, D. C. Jackson and N. H. Kalin, “Emotion
Plasticity, Context and Regulation: Perspectives from Af-
fective Neuroseience,” Psychological Bulletin, Vol. 126,
No. 4, 2000, pp. 890-909.
http://dx.doi.org/10.1037/0033-2909.126.6.890
[17] K. C. Klauer and J. Musc h, “Affective Priming: Findings
and Theories,” In: J. Musch and K. C. Klauer, Eds., The
Psychology of Evaluation: Affective Processes in Cogni-
tion and Emotion, Lawrence Erlbaum, Mahwah, NJ, 2003,
pp. 7-49.
[18] J. R. Stroop, “Studies of Interference in Serial Verbal Re-
actions,” Journal of Experimental Psychology, Vol. 18,
1935, pp. 643-662. http://dx.doi.org/10.1037/h0054651