Colour: Design & Creativity (2007) 1 (1): 9, 1–15 http://www.colour-journal.

org/2007/1/9/

Sign ifi c a n ce o f Co lo u r o n Ro o m Ch a rac te r: Stu d y o n

D o m in a n tly Re dd is h a n d Gre e n is h Co lo u rs in No r th - a n d
So u th -Facin g Ro o m s

Maud Hårlem an, Inga-Britt Werner and Monica Billger*

School of Architect ure and Built Env ironm ent, Roy al In st it ute of Technology, 10 0 44
Stockholm , Sw eden

Colour: Design & Creativity

*School of Architect ure, Chalm ers Universit y of Technology, Sw eden
Em ail: m [email protected]

Room character attributable to colour appearance and spatial evaluation in different compass orientations
has been studied. Rooms of the same colour, but observed in light from different compass directions
will appear differently, in particular their identity colours will differ. A study was carried out in Sweden
in two identical rooms, north- and south-facing. Seventy-two subjects evaluated and compared the
experimental rooms. The inherent colours were six pinkish and five greenish colours, all except one in
two nuances, plus one yellowish and one bluish colour, in total 13 colours. Description of room character
was aided by semantic scales. Data were statistically processed using the Statistical Package for the
Social Sciences (SPSS) program to analyse connections between spatial evaluation, inherent colour
and compass orientation. The NCS colour vocabulary was also used. Finally, verbal descriptions
using the subjects own words were used as a supplementary method to unfold nuances in response.
The study showed that differences in hue and nuance affected evaluation of room character. Subjects
reacted differently in pinkish and greenish rooms, describing distinct colour connotations. Differences
in direction of illumination caused weakening or strengthening of associations linked to the colours by
colour connotations.

1 Introduction

Identically coloured room s illum inated by light from different compass direction s will not
appear the same in the Nordic countr ies. In clear weather a south-facing room is dom inated
by sunlight while a nor th-facing room is illum inated by skylight and reflected light. As sunlight
and skylight differ in lum inous inten sit y, light direction and spectral character istics result in
a m ajor difference in light qualit y. Con sequently colour appearance differs between room s
even when they have been painted in the same colour, with the same paint. Traditionally these
different light qualities are refer red to as ‘war m light’ and ‘cold light’. How does this affect
evaluation of room and colour? Investigation s into the effect of daylight on colour and spatial
evaluation have been lacking. In teaching colour design, it is impor tant to con sider emotional
respon se to these situation s and to m ap out impression s of colour appearance and colour
design in different daylight qualities.
This study relates to research on environ mental colour design, concer n ing the relation ship
between m an, environment and colour. Our aim has been to understand connections between
visual appearance and the sign ificance of light and colour on room character. The m ain focus
of the work has exam ined room characters as a function of direction of illum ination. What
happen s when room s of the same in herent colour gain different identit y colours due to light
from different compass or ientation s? Chiefl y reddish and green ish colours have been used,

© 2007 Authors. Journal compilation © 2007 Society of Dyers and Colourists 1

Colour: Design & Creativity (2007) 1 (1): 9, 1–15 Hårleman et al.

together with one yellowish and one bluish reference. We have evaluated space and room
character. The question s to be an swered are:
1. If room colour ing affects evaluation of room character, and in what way?
2. If and in what way do different daylight qualities affect room character?

The m ain aim h as been to invest igate t he con nect ion bet ween verbal expression and
emotional impression of room colours due to different light qualities. A secondar y aim has
been to supply and test t wo new mean ing var iables: ‘embracing’ and ‘elevating’. The study is
named the ‘red-green study’, yet the reddish colours are referred to in term s of ‘pin kish’, since
reddish colours in the nuances used are generally called pin k.

2 Background

2 .1 D if fe re n t Appro ach e s to Co lo u r

Osgood et al. introduced a psychometr ic tech n ique with sem antic differentials used with
factor analysis [1]. In this techn ique, subjects m ark a representative value for evaluation with
the help of var iables or ‘item s’ on a graded scale. The problem was to fi nd appropr iate item s
for the study in hand. They found that item s could be grouped into ‘factors’ and showed that
three mutually independent factors represented the basis of most descriptive words; these were
value, activit y and potency. Research concer n ing a subject’s emotional reaction s on colour,
colour emotion, has a long histor y, especially among psychologists. Most studies on colour
emotion have been m ade for single colours or t wo-colour combination s [2– 6].
Küller developed a method for measur ing and descr ibing evaluation of built environ ment,
the SMB method [7]. He found eight factors: affection, complexit y, enclosedness, potency,
pleasantness, u n it y, social stat us and or igin alit y. Kü ller h im self par t icu larly pointed out
affection , complexit y and enclosed ness as impor tant contr ibutor s to research on spat ial
evaluation. Hogg et al. worked with simulated inter iors and found fi ve factors for colour in
spatial models: dynam ism, spatial quality, emotional tone, complexity and evaluation [8]. Over
the years, m any researchers have looked at factors in fluencing colour emotion [9– 17]. It was
concluded that item s can be classified under different factors in different areas of research,
for example isolated colours, colour combination s, colour in spatial models and colour on
exter iors. However, concer n ing inter ior spaces, the ‘value factor’ and a factor for activit y and
potency were still con sistently featured in most collation s of factors encountered.

2 .2 Co n n e c tio n be t w e e n Ite m s a n d Fac to rs

Many studies dealt with colour and temperature. Using colour samples it was found that a
subject’s reaction on colour is rooted in an experience of warm and cold colours [3,5]. Warmest
are red, (red, orange and yellow) and coldest are blue and green, (blue and blue-green). In
full-scale exper iments it was found that colours do not cause an exper ience of a physical
temperature, but can in stead cause a cogn itive exper ience with association s of war mth and
coldness [8,11– 13].
Kü ller searched for but fou nd no simple con nect ion bet ween t he ‘pleasantness’ factor
concern ing hue and nuance [11]. Greater differences were found between nuances within each
hue than bet ween different hues. Thus there was not a straightfor ward con nection bet ween

2 © 2007 Authors. Journal compilation © 2007 Society of Dyers and Colourists

Hårleman et al. Colour: Design & Creativity (2007) 1 (1): 9, 1–15

nu an ce an d pleasant n ess, an d in d ividu a ls eva lu ated t he pleasant n ess factor wit h gr eat
differences. It was found that light and whitish spaces contr ibute to a sen se of open ness, and
influence the size and form of the object. They also found that red colours had a more activating
effect on subjects.

2 .3 Re a l Ro o m s w ith Co m ple x In te rac tio n s

Billger has studied colour in real room s in ter m s of the complex interaction s bet ween for m ,
light and colour [18,19]. Her interest has been in colour appearance, colour change and colour
descr iption using complete room s and ar tificial light. Billger emphasised the direct attention
and con scious reflection of subjects when in room s, fi nding it valuable to use subject’s verbal
descr iption s on colour appearance.
Hårlem an accomplished two full-scale studies on colour appearance in natural daylight.
Both studies concerned nor th- and south-facing room s, and involved yellowish/ bluish [20 – 22]
and reddish/ green ish [23] colours in two nuances. Stahre et al. [24] compared t wo previous
studies on associations and emotional response on colours: Hårlem an’s 20 0 4 room study [23]
and a study using colour patches [14]. It showed that the colour patches had to be much more
colour ful to give comparable association s.

2 . 4 Te r m in o lo g y

We used the NCS system and adopted its ter m inology [25]. Colour is thus defi ned as: colour
percept, colour obser vation, colour perception; that which people see as colour and m ake it
possible to m ake out objects and fields on the basis of colour differences. A hue in the NCS
system is defi ned according to the relation to the four elementar y colours: red, blue, green,
yellow. Hue relation is shown through position in the colour circle. The colour tr iangle shows
colour nuance, descr ibed as par ts of blackness, whiteness and chrom aticness. Chrom aticness
is the sum of a colour’s chrom atic qualities. ‘Elementar y attr ibutes’ are ter med whiteness,
blackness, yellowness, red ness, blueness and green ness. Orange has t hus t wo ch rom at ic
elementar y attr ibutes: yellow and red.
The ter m ‘in her en t colou r ’ is u sed a s a ba se p oin t for m ea su r in g sh ift s in hue a n d
nuance. In herent colour refers to the colour a coloured object wou ld have if obser ved in
standardised obser vation condition s as denoted for NCS colour samples in agreement with
their specification s, i.e. with their colour code. This entails colour samples being placed at a
45° angle, obser ved in a light cabinet with simulated daylight consisting of six Lum a Colorette
fluorescent lights at 540 0 K [25,26]. Inherent colour is a constant characteristic independent of
exter nal circum stances, assum ing requirements are adhered to. In herent colour is compared
to the identit y colour that ar ises. The ter m ‘identit y colour’, as developed by Billger, is a ter m
which tallies with a holistic attitude and should thus be inter preted as the m ain impression of
what is apprehended as a single-coloured sur face in a room . The ter m ‘exper ience’ is used as
an ever yday term for mental and emotional excitation. Such exper ience can consist of feelings
or emotional states, which ar ise in the exper imental room . These exper iences or emotional
states m ay con sist of memor ies, association s and metaphors. Room character here function s
as an over r iding ter m for evaluation of a room with light, colour, space, sur faces and other
arch itecton ic elem ents. Fin ally, colour con notation, as Sivik and Taft defi ned it, is words
having mean ing not pr im ar ily related to colour [6]. We use colour con notations for the idea or
mean ing suggested by or associated with a colour, a room colour or a colour word.

© 2007 Authors. Journal compilation © 2007 Society of Dyers and Colourists 3

Colour: Design & Creativity (2007) 1 (1): 9, 1–15 Hårleman et al.

3 Study Design

3 .1 D e s ig n Appro ach

The exper imental per iod was J une to September in Stockholm , where the sun r ises bet ween
0 334 and 0 634 hours and sets between 220 5 and 18 45. The intention with the study design
was to create an environ mental colour design in order to catch reaction s to room and colour
as subtle descr iption s. The m ain focus dur ing the session s was to supply the subjects with a
standard situation, in this case a calm and ‘nor m al’ (in contrast to exper imental) situation,
a pause for refl ection per m it ting them to focus on room and colour. In total, 72 subjects
m ade 118 obser vat ion s. Sky cond it ion s for each session wer e stated in a quest ion n air e.
The methodology was designed for separate sequences, and to put subjects at ease in the
exper imental situation.

3 .2 Su bje c ts a n d Ex p e r im e n ta l Ro o m s

Subjects were architects and inter ior designers, plus students reading these subjects. This
choice of professional categor y, with people interested in both colour and spaces, was m ade so
as to obtain, as far as possible, in for med and detailed descr iption s 1. Each obser ver gave t wo
complete descr iption s of colour appearance, one in each room . These together took over one
hour and dur ing that time four fur ther par t-studies were m ade.
Two sim ilar full-scale room s were set up in a con str uction cabin positioned facing nor th–
south. The cabin was placed in a slope, with vegetation in front of the room in sun light and
other houses with yellow plaster outside the room in skylight. Room measurements were 4.2 ×
2.9 m . The in ner sur face of the walls con sisted of plywood roller-painted with a new in herent
colour for each test sequence. Floors were covered with beige-speckled lino, and ceilings were
white-painted. Both room s had sim ilar shor t-end windows with white-painted frames and
in ner reveals. The n atural daylight was tran sm itted from clear, double-glazed panes. The
colour temperature in the room s on a fairly cloudy day was approx 80 0 0 K in the nor th-facing
and approx 70 0 0 K in the south-facing room (Figure 1).

N S

Figure 1 Experimental rooms (two similar full-scale rooms were set up in a cabin positioned facing north–south)

1 The subjects were a rather homogeneous group and it was not intended to use observer’s age and gender
statistically. We started with a greater proportion of females and ended with a high proportion of students; otherwise
no analysis according to age, gender or occupation was attempted.

4 © 2007 Authors. Journal compilation © 2007 Society of Dyers and Colourists

Hårleman et al. Colour: Design & Creativity (2007) 1 (1): 9, 1–15

3 .3 Co lo u r S e le c tio n

The colours chosen were three pin kish and three green ish hues in two nuances com mon ly
used in architectural and inter ior colour design s. They were chosen to have perceptual equal
distance 2 . The nuances were whitish 1010 and more chrom atic 1030. The hues were yellowish-
pink (Y80R), pink (R), bluish-pink (R20B), bluish-green (B70 G), green (G) and yellowish-green
(G20Y). Due to un suitable weather condition s for a lengthy per iod, the pale yellowish-green
colour (1010 -G20Y) was om itted. One yellowish and one bluish in herent colour that were
used in a previous study were included [20 – 22]. These colours, reddish-yellow (10 30 -Y20R)
and reddish-blue (10 30 -R80B), showed interesting patter n s of hue shift, and could ser ve as
reference points bet ween the studies (Figure 2). The colours used are shown in Figure 3.
CIELAB coordinates for the colours used are given in Table 1.

W
Y
10

20
30
40
G R 50 C
60 100
90
80
70 70
80 60
50
90 40
30
100 20
B 10
S

Figure 2 Selected colours, in hue and nuance: three hues in two nuances, in total six pinkish and five greenish
colours

1010-Y80R 1030-Y80R 1030- G20Y 1030-Y20R

1010- R 1030- R 1010- G 1030- G

1010- R20B 1030- R20B 1010- B70Y 1030- B70G 1030- R80B

Figure 3 Colour display

2 In the NCS colour system colours are arranged according to similarity with the elementary colours: white, black,
yellow, red, blue and green. As a consequence, the NCS system does not define equal distance between elementary
colours, and the colour circle has in fact less perceptual difference between colours in the blue-green quadrant than
in the others. Consequently, we have chosen to adjust that situation by selecting one green inherent colour at a larger
distance from the elementary green, 30 steps instead of 20.

© 2007 Authors. Journal compilation © 2007 Society of Dyers and Colourists 5

Colour: Design & Creativity (2007) 1 (1): 9, 1–15 Hårleman et al.

Table 1 CIELAB coordinates for the used NCS colours*

NCS L a b

S 1010-Y80R 84.67 6.72 7.44

S 1010-R 83.39 5.72 3.63

S 1010-R20B 84.33 6.24 0.04

S 1010-B70G 85.94 –6.66 0.28

S 1010-G 85.43 –7.37 3.73

S 1030-B70G 81.17 –20.85 –0.04

S 1030-G 79.94 –20.33 9.89

S 1030-G20Y 80.03 –17.57 17.65

S 1030-Y20R 82.50 9.29 34.42

S 1030-Y80R 74.83 21.74 16.21

S 1030-R 75.47 21.31 7.12

S 1030-R20B 75.80 18.92 –0.05

S 1030-R80B 77.25 –3.57 –17.91

* Illuminant D65, 10° observer

3 . 4 Em pir ic a l Me th o d s

Room and colour were obser ved and evaluated with qualitative methods, such as sem antic
scales and verbal descr iption. We have used four out of fi ve factors developed by Hogg et al.:
temperature, emotional tone, spatial qualit y, and dynam ic factor [8] 3 . Established item s for
spatial exper ience were selected and grouped hypothetically under these factors. Two new
var iables for exper ience were added to the already established, em bracing and elevating,
associated with the emotional factor. The sem antic scales were graded from zero to six, with
six being the maximum, covering a value from ‘not at all’ to a value ‘to a high degree’, i.e. a type
called Liker t scales and not sem antic differentials in the sen se that they covered a scale from
one position to its opposite. The use of Liker t scales allows a better control and thus higher
data qualit y; Table 2 show item s in the or iginal grouping.

Table 2 Selected items grouped in original factors

Temperature Emotional Spatial Dynamic

Warm Merry Distinct Elevating

Cold Sombre Open Vivid

Sunny Hard Embracing Tranquil

Pleasant Small

Formal

Dry

3 The fifth factor, complexity, was not selected as it did not seem adequate in the unfurnished building shed.

6 © 2007 Authors. Journal compilation © 2007 Society of Dyers and Colourists

Hårleman et al. Colour: Design & Creativity (2007) 1 (1): 9, 1–15

A parallel method for verbal descr iption was developed to catch the in itial exper ience of
room and colour. Subjects were encouraged to note down own thoughts and descriptive words.
The given tasks were two-fold:
1. Descr ibe your impression of the wall colour
2. Descr ibe your impression of the room character.

The room descr iption was the prefer red option and the wall colour descr iption was used
as a complement. The subjects were in str ucted verbally to take their time for the process of
exper ience and association. It was impor tant that fi rst impression s should not be by-passed
and that evaluation s and exper ience of the room were allowed to develop. J ust as eyes must
adapt to a light situation, so must exper ience as a whole have time to adapt in a cer tain spatial
situation.

3 .5 P ro c e d u re

The walls in two full-scale room s were painted in the same in herent colour. Most session s
were separate but occasionally two subjects were present. After completion, the subject moved
to the second room and the process was repeated. The subjects stayed in each room without
any inter r uption, in order to adapt to prevailing light and exper iment situation. Obser vation
was thus con fi ned to one room at a time. The adaptation to room light condition s as well as
the session was m ade seated on a bench in a defi n ite position with a good over view of the
room . The exper iment super visor guided the subjects through the process and subsequent
question naire, and in str ucted the subject to reflect on the impression that room had m ade.
In t he in it ial phase verbal descr ipt ion was per for m ed, following by a per iod when m ore
concentrated attention was required. Subjects noted their exper ience with a value for each
item and 16 different item s were used.

3 .6 Me th o d s o f An a lys is

Data wer e accu m u lated fr om st ud ies in r oom s illu m in ated fr om t wo d iffer ent com pass
direction s, and with six colour hues in t wo nuances. Data were processed statistically using
the SPSS statistical program .
Factor analysis was applied to the whole data set, using data from all in herent colours. This
method generated new variables for fur ther analyses. Differences in evaluation between room s
in different colours and or ientation s were analysed by t-tests. In order to test the effect on
evaluation by colour and or ientation respectively an analysis of var iance was car r ied out. To
accentuate differences in factor loadings a Var im ax Rotation was applied.
Verbal descriptions from the studies were made use of. The most frequently used expressions
(at least five appearances) where put into three groups, one for each of the compass direction s
nor th and south respectively, and one group for expression s in com mon to both.

4 Study Results

4 .1 Ne w Va r iable s fo r Ro o m Ch a rac te r

The factor analysis was car r ied out for all obser vation s, through all hues and nuances, using
the pr incipal components extraction method. Using Var im ax Rotation, five new components
with eigenvalues greater than 1 where con str ucted. The fi rst three explained in all 54.6% of

© 2007 Authors. Journal compilation © 2007 Society of Dyers and Colourists 7

Colour: Design & Creativity (2007) 1 (1): 9, 1–15 Hårleman et al.

the total var iance, together with the rem ain ing two they explained 67.7% of the total var iance.
The percentage under the name of the component shows the amount of var iance explained by
the component respectively. The grouping of new var iables is presented in Table 3.
Following the factor analysis, four new variables were constructed to reduce data. The values
of these four new var iables were calculated as the mean of the included item s [27]. Item s with
negative factor loadings were not included in the calculation: working as m ir ror im ages of
their opposites they would have con founded the values of factor var iables by levelling out the
calculated mean values. Var iables included have factor loadings over 0 .4. New var iables and
character isation based on included item s are presented in Table 4.
The fi rst t wo var iables (elevating and harsh) in Table 4 descr ibe t wo different attitudes
towards the room , a positive and a negative. Temperature goes together with other emotional

Table 3 Factor matrix of room character (rotated component matrix)*

Items Components

1 (31.3%) 2 (13.6%) 3 (9.7%) 4 (6.7%) 5 (6.4%)

Elevating 0.805 –0.274 –0.086 0.134 0.191

Merry 0.796 –0.302 0.113 –0.054 0.098

Vivid 0.728 –0.176 0.041 –0.342 0.367

Sunny 0.687 –0.083 0.043 0.163 –0.167

Pleasant 0.600 –0.333 –0.231 0.500 0.074

Warm 0.456 (–0.623) 0.174 0.128 –0.148

Sombre (–0.693) 0.164 0.146 0.212 –0.093

Hard –0.170 0.782 –0.052 –0.251 0.041

Cold –0.286 0.741 –0.114 0.041 0.090

Formal –0.196 0.607 –0.189 0.143 –0.201

Dry –0.005 0.569 0.422 0.368 –0.118

Embracing –0.095 –0.294 0.703 0.077 0.321

Small 0.215 –0.022 0.597 –0.230 –0.153

Open 0.389 0.115 (–0.664) 0.316 0.033

Tranquil –0.055 –0.016 –0.198 0.821 –0.006

Distinct 0.165 0.045 0.004 0.007 0.881

* Extraction method: Principal component analysis

Rotation method: Varimax with Kaiser normalisation

Table 4 New variables for room character

Elevating Harsh Embracing Peaceful

Elevating Hard Embracing Tranquil

Merry Cold Small Pleasant

Vivid Formal Dry

Sunny Dry

Pleasant

Warm

8 © 2007 Authors. Journal compilation © 2007 Society of Dyers and Colourists

Hårleman et al. Colour: Design & Creativity (2007) 1 (1): 9, 1–15

expression s: war m , sun ny, mer r y, elevating compared with cold, harsh, for m al and dr y. The
third var iable descr ibes an embracing spatial exper ience, m aybe even tight or nar row. The
four th var iable br ings about a positive sen se of peace, a mood, where the ter m pleasant is
included as in the fi rst var iable.

4 .2 Im p ac t Cau s e d by Co lo u r o r Or ie n tatio n

The impact of hue and or ientation on spatial exper ience was assessed by t-tests. Table 5
shows compar ison s between categor ies according to or ientation and colour. The com ments
on ‘more’ or ‘less’ show the direction s of the differences. Test A concer n s or ientation on ly: all
cases were compared, divided into nor th- and south-facing room categor ies. Test B concer n s
colour and or ientation: all cases divided into pin k and green categor ies are compared with the
or ientation categor ies. Test C concer n s colour and or ientation: all cases divided into sm aller
colour categor ies are compared with the or ientation categor ies. Test D concer n s on ly pin kish
colours and or ientation: a yellowish-pin k and a bluish-pin k categor y are compared with the
or ientation categor ies.
There were some sign ificant differences in the perception of the room s, but mostly the
difference between different in herent colours over powered the difference between the identity
colours from the same in herent colour (tests B1 and C1). Using the whole data set there was an
indication, but not statistically significant (p < 0.10), that the or ientation facing nor th was less
related to an elevating character. When the data set was grouped into sub-samples of colour

Table 5 Impact of colour or orientation

Categories
compared Details Elevating Harsh Embracing Peaceful Distinct

A Orientation All cases divided p < 0.10 – – – –

into 2 nuances: (north less)
104 north and
104 south studies

B1 Hues: 100 cases divided – p < 0.01 p < 0.01 – p < 0.05
pink and green into 2 nuances: (pink less) (pink more) (pink more)
54 pink and 46
green studies

B2 Orientation in 100 cases divided – – – – –

pink and green into 50 north and
50 south studies

C1 Hues: 60 cases divided – – p < 0.05 – –

yellowish-pink into 2 nuances: (yellowish-
Y80R and bluish- 30 Y80R and 30 pink more)
green B70G B70G studies

C2 Orientation 60 cases divided p < 0.01 – – – –

in Y80R into 30 north and (north less)
and B70G 30 south studies

D1 Hues: 68 cases divided – – – – –

yellowish-pink into 2 nuances:
Y80R and bluish- 30 Y80R and 38
pink R20B R20B studies

D2 Orientation 68 cases divided – – – –

in Y80R into 34 north and
and R20B 34 south studies

© 2007 Authors. Journal compilation © 2007 Society of Dyers and Colourists 9

Colour: Design & Creativity (2007) 1 (1): 9, 1–15 Hårleman et al.

in two nuances, this difference was sign ificant for the group consisting of yellowish-pin ks and
bluish-green s 4 .
Pink room s were evaluated as less ‘harsh’ and more ‘embracing’ than were green room s. The
‘embracing’ character includes an allusion to a cramped condition, and this was more explicit
in both pin k and yellowish-pin k room s than in green and bluish-green room s. Also the pin k
room s were perceived as more distinct than the green ones. Compar ison between the sub-
group yellowish-pin k and bluish-pin k did not show any sign ificant difference.

4 .3 Ef fe c ts D u e to H u e a n d Or ie n tatio n

A two-way analysis of var iance was used to explain the effects of colour in relation to the
effects of or ientation, and their mutual interaction. The var iables studied were the same as in
the t-tests: elevating, harsh, embracing, peaceful and distinct. In Table 6 numbers in bold type
sign ify sign ificant con nections; numbers in italic are par tially sign ificant. Table 7 presents the
propor tion s of var iance explained.
The analysis showed an explicit effect by differences in colour. Regarding pin k and green
hues, the m ain par t of the var iance explained stem med from different hues. It mean s that
differences between room evaluations are greater between red and green colours than between
d ifferent or ientat ion s. Gener ally, t he nuance 10 30 d id affect bot h posit ive and negat ive
emotional item s more than the lighter 1010 nuances.

4 . 4 Co lo u r Co n n o tatio n s in Ve rba l D e s c r ip tio n

The subjects used a variety of expressions to describe room character and differences between
them. Most obvious was that greenish and pinkish room s evoked different colour connotations.
Pin kish room s caused associations of hum an skin, facial colour, strong emotional expressions
such as affection and defi ance and other mental character istics, while green ish room s caused
association s of nature that could be either pure nature or ar tificial. Within these fields of
association s subjects had all sor ts of different perspectives.
Some subjects descr ibed the pin kish room s as war m , gentle and stimulating but others
exper ienced t hose r oom s as pu shy, dem and in g and glar in g. Som e exper ienced t hem as
childish, young, fresh and funny while yet others described them as stale, tasteless, vulgar and
sloven ly. Pin kish room s produced var ious im ages of skin (dressed or naked, of sm all children
or adults), and views var ied on whether they were sweet or tr ying, in nocent or sin ful.
Green ish colour con notation s were of a completely different kind. Green ish room s evoked
con notat ion s of n at u re, and rela xat ion . These room s projected a sh adow, an im age of a
landscape, or an un restr icted space. This in tur n caused association s of calm , a retreat or
shelter; elements like water and wind and objects and room s like a pool or a bathroom were
mentioned. Greenish room s also caused associations related to health. Many subjects described
them as clean and pure, and opin ion s differed as to whether the room could be descr ibed as
living or plastic5 . Adjectives frequently applied to in green ish room s are calm , peaceful, light-
hear ted, con fident, soothing and tranquil.

4 Using the whole dataset the inherent colours were contra productive and the smaller groups tended to be too small
for statistics.
5 In Swedish: levande och plastig

10 © 2007 Authors. Journal compilation © 2007 Society of Dyers and Colourists

© 2007 Authors. Journal compilation © 2007 Society of Dyers and Colourists

Hårleman et al.
Table 6 Explanation of variance in effect

Elevating Harsh Embracing Peaceful Distinct

Categories
compared H O I H O I H O I H O I H O I
1010 red (R) and green (G) 0.05 1.25 0.75 0.71 0.61 0.11 4.11 0.10 2.68 0.57 0.24 0.76 † † †
1030* red (R) and green(G) 2.96 0.06 1.79 8.05 2.84 3.48 3.35 1.87 0.07 0.01 0.25 0.15 5.64 0.27 2.41
1010, 1030 all R and G 0.72 0.42 0.07 5.61 0.00 0.18 7.39 1.13 2.24 0.39 0.48 0.63 3.21 0.43 1.72
1010, 1030* all Y80R and B79G 0.04 6.48 0.04 0.90 0.19 2.00 4.47 0.01 0.01 1.29 0.03 1.13 1.66 3.08 1.11
1010, 1030* all Y80R and R20B 0.34 2.18 3.73 0.16 0.00 0.75 4.07 0.56 1.95 0.15 1.32 0.02 4.46 0.84 0.84

* Studies in cloudy weather included

† Studies excluded due to overcast and dark weather conditions
H hue
O orientation
I interactions

Table 7 Explanation of proportions of variance (%)

Elevating Harsh Embracing Peaceful Distinct

Categories
compared H O I H O I H O I H O I H O I
1010 red (R) and green (G) 0.2 2.4 1.5 1.3 2.1 0.4 8.4 0.3 7.1 2.0 0.4 1.4 † † †
1030* red (R) and green (G) 9.8 0.2 4.9 25.0 4.4 5.4 10.0 5.4 0.2 0.0 1.3 0.8 18.0 0.6 5.2
1010, 1030 all R and G 1.2 0.5 0.1 6.7 0.0 0.3 8.9 1.5 3.0 0.5 0.7 0.9 5.9 0.4 1.8
1010, 1030* all Y80R and B79G 0.1 20.0 0.1 3.8 0.2 2.1 12.0 0.0 0.0 5.0 0.0 1.5 6.5 3.3 1.2

Colour: Design & Creativity (2007) 1 (1): 9, 1–15

1010, 1030* all Y80R and R20B 1.6 3.8 6.5 0.9 0.0 0.9 16.0 0.9 3.1 0.6 4.0 0.1 16.0 1.8 1.8

* Studies in cloudy weather included

† Studies excluded due to overcast and dark weather conditions
H hue
O orientation
I interactions
11
Colour: Design & Creativity (2007) 1 (1): 9, 1–15 Hårleman et al.

4 .5 In fl u e n c e o f Or ie n tatio n o n Ve rba l D e s c r ip tio n s

As subjects reacted to situation s and different identit y colours according to the direction of
illum ination, this showed as var iation s in character. Different physical qualities emphasised
characters included in the colour con notation. In the strongly illum inated south-facing room ,
in herent colours with a yellowish content (G20Y, Y20R and Y80R) appeared with sur pr ising
chromatic intensity. Thus connotations linked to these colours were experienced and expressed
in ter m s of stronger verbal nuances. In the pin kish room s differences in visual impact m ade
subjects exper ience var iation s in character as: sweetness/ folly, pure/ sophisticated, etc. Y20R
and Y80R were con sidered as soft and fluffy in the south-facing room , while in the nor th-
facing room , having less (or null) yellowish attr ibute, the room was con sidered as more bare
and empt y, more hard and cold in compar ison. Elementar y red (R) in the south-facing room
appeared as yellowish-pin k and was por trayed as childish, pure and in nocent. In the nor th-
facing room it was bluish-pin k and represented impur it y and sophistication.
As t he green ish room s evoked ot her associat ion s, ch ar acter s also var ied. Nat u re and
qu iescence, slowness and t r anqu illit y were com m on descr ipt ion s for room s facin g bot h
compass direction s, also clean and fresh. Elementar y green (G) in the nor th-or iented room
(BG identit y colour) m ade subjects descr ibe the room exper ience as uneasy to grasp, alive or
plastic, disgusting or restful, pushy or relaxing. The south-or iented room of the same colour
(G identity colour) provoked descriptions such as warm, calm, soothing and gloomy. In Figures
4, 5, and 6 perceived colours and verbal descr iption of the room s in both compass direction s
are ar ranged together to illustrate their mutual relation ships.6

S 1030-Y80R S 1030- R S 1030- R20

I nt ense Cold Dignified

Darker Alive Nort h
Rest rained
Colder Adult

Snug Repet it ive

Secure Soft Nost algic descript ion for
Em bracing Pushy St rong room s in bot h
Pushy I m port unat e I ndependent com pass
Dom inant Trying Un- fresh direct ions

Soft Brave
Fluffy Warm Cheerful
Young Sout h
Energet ic Funny
Alive Relaxing Rom ant ic

Figure 4 Frequent verbal descriptions of the pinkish rooms, showing different experiences of hue as influenced by
the direction of illumination (descriptions in the middle section are in common to both)

6Here, mean values are not used but observers’ minimum and maximum notations have been chosen to provide a
wider spectrum of perceived colour in north- and south-facing rooms. Figures showing mean values of perceived
colour are presented separately [29].

12 © 2007 Authors. Journal compilation © 2007 Society of Dyers and Colourists

Hårleman et al. Colour: Design & Creativity (2007) 1 (1): 9, 1–15

S 1030- B70G S 1030- G S 1030- G20Y

Figure 5 Frequent verbal descriptions

of the greenish rooms, showing different
Alive versus experiences of hue as influenced by the
Nort h Dignified plast ic Dissociat ion
Rest rained Art ificial direction of illumination (descriptions
Disgust ing versus
rest ful in the middle section are in common
to both)

Repet it ive
Calm cool Fresh Calm
descript ion for
Fresh Clean Unsophist icat ed
room s in bot h
Pure Tranquil Clum sy
com pass
I nt ense Slow Sad
direct ions

Warm
Warm Calm
Sout h Warm
Em pt y Em bracing
Soot hing

S 1030-Y20R S 1030- R80B

Soft Dynam ic
Nort h Light er Excit ing
Act ivat ing Tranquil

Repet it ive
descript ion for Colourful Em bracing
room s in bot h Fun Enclosing
com pass Quiet som e Dem anding
direct ions

Warm Pale blue

Sout h Welcom ing Cool
Peaceful Soot hing

Figure 6 Frequent verbal descriptions of the yellowish

and bluish rooms

5 Discussion

5 .1 Co m pi le d An a lys is

Besides t he over r id in g effect cau sed by d ifferences in colou r on room evalu at ion , som e
d ifferences in or ientation were statistically sign ificant. It seem s that colour con notation s
pr ovide an in stant star t in g point fr om wher e associat ion s or igin ated. For exam ple, t he
elementar y qualities of pin k aroused cer tain expectation s of evaluation content. Perhaps the
colour did not suit the exper imental room; it was descr ibed fem in ine, pr ivate and childish.
Thereafter, the actual identit y colour led to a second impression, for example the descr iption
of nor th-facing room s as vivid was due to the intensely chrom atic bluish-pin k identity colours.
In t he act of r ecogn isin g, accept in g and exper iencin g r oom colou r, colou r con notat ion s
function as a basic idea again st which specific colour impression s are compared. The general
colour con notation seem s to be calibrated in ter m s of association s with identit y colour, as a

© 2007 Authors. Journal compilation © 2007 Society of Dyers and Colourists 13

Colour: Design & Creativity (2007) 1 (1): 9, 1–15 Hårleman et al.

con fi r m ation or a den ial of the basic character. In this way an overall evaluation can be m ade
through attaching descriptive words to a set context. It reflects cultural, traditional, sociological
or subjective mean ings, a colour con notation expressing (indirectly) a cer tain condition.

5 .2 Ne w Va r iable s a n d Ite m s

The SPSS methods provided clear result. However, without the supplementar y in for m ation
provided by the verbal descriptions, a full understanding of the complex relationships between
or ientation, hue, nuance and evaluation would not have been possible.
The original factors represented mutually exclusive categories: temperature, emotional tone,
spatial qualit y and dynam ism . None of these or iginal factors rem ain and most of the item s
have changed places. The new var iables represent four emotional var iables with overlapping
characters, and one separate var iable concer n ing spatial qualit y. They are not comparable
with the or iginal factors as they emerge from different ways of analysing data. The ‘elevating’
var iable was essential in descr ibing room character; war m colours and the more chrom atic
nuances increase happy and pleasing emotional exper iences. The ‘distinct’ var iable correlated
well with the obser vation of war m colours as proceed ing wh ile the cold ones seem to be
receding. As the cold bluish and green ish walls seemed to withdraw they con sequently m ade
an indistinct impression.

5 .3 Co m p a r is o n s be t w e e n Stu d ie s

Pin k aroused strongest emotional reaction s, regarding both positive and negative character.
The conclusion of Hogg et al. that the war mth factor more often expresses an emotional tone
than strict temperature is confi rmed [8]. Sivik has also pointed out that pink, and in par ticular
bluish-pin k, is often descr ibed in negative ter m s such as ugly, disagreeable and unappetising
[5]. We can see a clear relation sh ip bet ween these negative evaluation s from pin k colour
samples and room character.
The new var iable ‘embracing’, previously belonging to the emotional factor, was grouped
together with the ter m s sm all and closed (not open). This tallies well with Küller’s conclusion
concer n ing the word-scales closed/ air y and sm all/ large, as being not fully cor related [11].
Room s assessed as large could just as well be perceived as enclosed. Evidently ‘embracing’ is
not on ly a spatial scale for emotional tone but also for a cramped sen se.
However, tr ying to fi nd con nection s bet ween item s and factors can be hazardous. Since
m any item s often compose a factor, it is hard to fi nd a per fect name that entirely encapsulates
the tr ue content. There is no standard procedure to help with the nam ing process and this
m akes it even more difficult to compare con nection s bet ween colour emotion factors and
colour attr ibutes. A clearly positive evaluation in preference for war m colours did cr ystallise.
This could be expected with subjects as a cross-section of the population but is an interesting
result for arch itects and student arch itects. Th is unexpected tendency is wor th noting as
several researchers have presented evidence that the colour preferences of architects often
differ from those of non-architects, the for mer often prefer r ing str icter and cooler objects
[8,28]. This m ight indicate a wider application of the results presented.

6 Conclusion

The methods used in this study have helped to explore people’s experience of room and colour,
resulting in th ree new var iables that can be applied to the exper ience of room character:
elevating, harsh and embracing.

14 © 2007 Authors. Journal compilation © 2007 Society of Dyers and Colourists

Hårleman et al. Colour: Design & Creativity (2007) 1 (1): 9, 1–15

The study present offers a new perspective for colour exper ience in room s. We were able
to under take a great amount of system atic evaluation that could be ar ranged into a bigger
pattern, thus m aking it easier to grasp connections between visual appearance and evaluation
of colour.
Hue and nuance clearly rendered a com mon exper ience in colour con notation s. Spatial
evaluation was fi rm ly bound to the identity colour. Pinkish and greenish colours caused almost
opposing room characters. The difference bet ween war m and cold colours was clearest.
Light from different compass direction s resulted in var iances in room character. As nor th-
and south-facing room s of the same in herent colour gain different identity colours, such room s
m ake different impacts and subjects react differently. The direction of illum ination affected
spatial character, by mean s of a strengthen ing or weaken ing of colour at tr ibute. To some
degree, the nor th-facing and south-facing aspects imposed their own spatial characters.

7 References
1. C E Osgood, P H Tannenbaum and G J Suci, The measurement of meaning (Urbana: University of Illinois
Press, 1957).
2. B Wright and L Rainwater, J. General Psychol., 67 (1962) 89–99.
3. J A Hogg, J. General Psychol., 80 (1969) 129–140.
4. S Kobayashi, Col. Res. Appl., 6, (2) (1981) 93–107.
5. L Sivik, Proc. AIC Color 69, Stockholm, 1/2 (1969) 1064–1072.
6. L Sivik and C Taft, Colour combinations and associated meanings – semantic dimensions and colour
chords, Psychological Reports, Volume 1 (Gothenburg: University of Gothenburg, 1992).
7. R Küller, Doctoral dissertation, National Swedish Institute for Building Research, Stockholm (1972).
8. J Hogg, S Goodman, T Porter, B Mikkelides and D E Preddy, Br. J. Psychol., 70 (1979) 231–242.
9. M Kunishima and T Yanase, Ergonomics, 28 (6) (1985) 869–882.
10. R Küller in Architecture for people, Ed. B Mikellides (New York: Holt, Rinehart and Winston, 1980). ISBN
0-03-057491-9
11. R Küller, J. Light Vis. Environment, 10 (2) (1986) 33–37.
12. R Küller in Environment, cognition and action, Ed. T Gärling and G Evans (New York: Oxford University
Press, 1991). ISBN 0-19-506220-5
13. B Mikellides, Doctoral dissertation, Oxford Polytechnic (1989).
14. T Sato, K Kajiwara, H Hoshino and T Nakamura, Adv. Col. Sci. Technol., 3 (2000) 53–59.
15. L Oberasher and M Gallmetzer, Proc. AIC 2003, Bangkok (2003).
16. L C Ou, M R Luo, A Woodcock, and A Wright, Col. Res. Appl., 29 (3) (2004) 232–240.
17. L C Ou, M R Luo, A Woodcock, and A Wright, Col. Res. Appl., 29 (4) (2004) 292–298.
18. M Billger, Doctoral dissertation, Chalmers University of Technology, Gothenburg (1999). ISBN 91-7197-820
8
19. M Billger, Proc. AIC 2004, Porto Alegre, Brazil (2004).
20. M Hårleman, Färg i norr- och södervettande rum: kulörtonförskjutning genom varierande ljusförhållanden
(Stockholm: Royal institute of Technology (KTH), School of Architecture, 2000); ISBN 91-7170-590-2. In
Swedish.
21. M Hårleman, Nordic J. Architectural Res., 2 (2001) 41–48.
22. M Hårleman, Proc. AIC Color 2001, Rochester, USA, Vol. 4421 (2001) 69-71. ISBN 0-8194-4128-7
23. M Hårleman, Proc. AIC 2004, Porto Alegre, Brazil (2004).
24. B Stahre, M Hårleman and M Billger, Proc. AIC 2004, Porto Alegre, Brazil (2004).
25. A Hård and L Sivik, Col. Res. Appl., 6 (3) (1981) 129–138.
26. A Hård, L Sivik and G Tonnquist, Col. Res. Appl., 21 (3) (1996) 180–205, 206–220.
27. P Kline, The handbook of psychological testing (London: Routledge, 1995).
28. J Janssens, Looking at buildings – individual variations in the perception of building exteriors, (Lund,
Sweden: Lund Institute of Technology, 1984).
29. M Hårleman, Col. Des. Creativity, 1 (1) (2007) 8/1–15.

© 2007 Authors. Journal compilation © 2007 Society of Dyers and Colourists 15