The effect of Colour on Reaction Time Performance
Conducted by Eva Loth, Angela Martin, Rachel Powell, Christine Ruschak and Rebecca Tipper
The effects of environmental cues on the processing of information and task performance have received
much attention. The cognitive tuning approach (see Soldat, Sinclair and Mark, 1997, for review)
suggests positive environmental affective cues signal a situation to be benign and hence heuristic or non-
systematic processing strategies to be adequate. In contrast, neutral or negative cues imply that more
systematic, detailed processing is necessary to allow an appropriate response to be made. Th is approach
is supported by Bless, Bohner, Schwarz and Strack (1990): happy subjects were found to be influenced
equally by weak or strong arguments whereas sad subjects were influenced only if an argument was
strong. It was suggested that subjects in a p ositive mood were less likely to elaborate the projected
message for themselves. Ottati, Terkildsen and Hubbard (1997) showed that, under low motivation
conditions, subjects exposed to happy facial displays showed systematic processing under all condition
s. The subjects’ moods were suggested to affect processing style, as proposed by the cognitive tuning
model. However, angry facial displays yielded a processing style between the two extremes. Angry faces
should provoke negative moods and signal a threate ning situation requiring systematic processing more
so than neutral faces. It is possible that the relative ambiguity of the neutral facial displays was more
important in triggering systematic processing than the mood provoked by the faces.
Soldat et al. (1997) examined the effect of different paper colours on the ability to solve complex
problems. It was expected that red would be seen to be a warm, happy colour whereas blue would be
more likely to elicit negative affect. Hence red would lead to heuristic processing and poorer talk
performance whereas blue would cause more appropriate systematic processing, according to the
Cognitive Tuning Model. Subjects with low motivation presented with red paper were found to give
fewer accurate res ponses than those receiving blue paper. However, although Soldat et al (1997)
measured subjects’ mood, no effect of colour on affect was found; the supposed mediating effect of
mood would appear not to be necessary.
In nature, the colour red appears as a danger signal, warning of an unpleasant taste or toxicity. Also, in
Western society, red means ‘stop’, ‘danger’ or ‘be alert’ rather than implying a situation to be benign or
happy. Studies finding red to be more arousing than other colours are reviewed by Kwallek, Lewis and
Robbins (1988). Perhaps those of Soldat et al.’s (1997) subjects given red paper were more stressed and
aroused and less able to concentrate on complex problems.
The colour red is also associated with racing and speed. Macrae, Bodenhausen, Milne, Castelli,
Schloerscheid and Greco (in press) found subjects took longer to read a word list entitled ‘The
Shimuhuru Word Reading List’ than ‘The Schumacher Word Readin g List’. It was suggested that the
exemplar ‘Michael Schumacher’ was activated, causing the accessing of information about Schumacher.
http://psy.st-andrews.ac.uk/resources/colour1.html (1 of 5)09/03/2006 09:54:02
Given red paper, Soldat et al.’s (1997) subjects could have felt under more pressure to perform
efficiently and quickly, although no effect on time to complete the problems was found.
The present experiment examines the effect of colour on a simple reaction time task. It is aimed to find
whether colour affects responses where complex processing is not required, to see if an effect exists on a
lower level. If the cognitive tuning mod el can be applied to this task, the colour red would signal a
happy mood and benign situation, lessening alertness and slowing responses. Nevertheless, following
other work (e.g. see Kwallek et al., 1988, Macrae et al., in press), it is expected that red will increase
alertness, arousal and speed, shortening reaction times. It would not seem to be necessary to involve
mood or affective cues in explaining the effects of colour on task performance.
The colour variable was manipulated by changing the background colour of the computer screen
displaying the test stimuli. It is likely that, as subjects are accustomed to working with different coloured
screens, this manipulation could be less noticeab le than if unusual paper colours were used. Following
Soldat et al. (1997), the main colour variables were red or blue backgrounds. The test stimuli were
presented in either white or black to determine whether any effects found were due to differences in
contrast between the target and the varying background colours used; the target could be more easily
discriminated against one of the backgrounds than the others. A control condition with the maximum
contrast of white stimuli on a black background was inc luded. If a contrast effect exists, white stimuli
would be expected to yield shorter reaction times than black stimuli for both red and blue backgrounds,
with the black background producing the fastest results.
However, it is hypothesised that subjects presented with a red screen will give the shortest reaction times
to the target stimuli.
Subjects and Design:
Participants were 50 undergraduate psychology students at the University of St Andrews. The
experiment used a 3x2 factorial between-subject design with the variables background colour (red,
black, blue) and stimulus colour (black, white). This led to a total of five conditions.
The stimuli were presented on a computer screen and consisted of each a cross and a circle in either
black or white colour, respectively. The background colour of the computer monitor was either blue, red
or black. The red and blue colours were roughly matched in terms of their intensity and brightness.
http://psy.st-andrews.ac.uk/resources/colour1.html (2 of 5)09/03/2006 09:54:02
Subjects were randomly assigned to one of the five conditions and seated individually in front of a PC
computer screen. The instruction was presented on the computer screen and was held constant across all
conditions. After having read the inst ruction, the subject determined the start of the experiment by
pressing a key. After 6 test-trials, 20 experimental trials followed. The task consisted of pressing a key
as soon as the subject perceived the change of the cross into a circle. Both circle a nd cross were located
on the center of the computer screen. The time interval between the appearance of the circle varied
between 4,5 or 6 seconds.
The reaction-times (seconds) were recorded by the computer and a mean-reaction time appeared on the
screen upon completion of all trials.
What do the results show?
There was no significant effect of either screen colour, therefore the results do not confirm our
hypothesis that a red screen would produce faster reaction times.
There was a non-significant trend which was in the expected direction to some extent. The red screen
with white cross was responded to faster than the blue screen with white cross, and the red screen with
black cross faster than the blue screen with bl ack cross. This would support our prediction that red
screens evoke faster response times than blue screens, perhaps by increasing level of arousal. This effect
of colour, however, seems to be only present within the stronger effect of contrast, i.e.. in high contrast
(white crosses) conditions, red produced faster responses than blue, and within the low contrast
condition (black crosses) red was again faster than blue.
http://psy.st-andrews.ac.uk/resources/colour1.html (3 of 5)09/03/2006 09:54:02
Overall the screens with white crosses (i.e. high contrast) were responded to faster to than screens with
black crosses (i.e. low contrast). Also, the black screen with white cross, (maximum contrast) was the
fastest of all the conditions. Again, this was a non-significant trend.
Although these results were not at significance level, they do raise the question that it may not be colour
alone causing the differences in processing or reaction times, but discriminability of the target. This may
be one reason why Soldat et al. foun d more systematic strategies on their blue papers. It may not have
been blue paper inducing negative affect but rather that a conscious, careful effort to discriminate the
text (as it could not be seen so clearly), that caused an increase in general conce ntration resulting in
more careful processing of the problems.
Is colour a factor in reaction time tasks?
Perhaps our study showed that colour is not a factor in this type of simple reflex task. As we did not find
any effect, it may then be possible that the influence of colour is only observable in more complex
cognitive processes, (as Soldat implied) and that in such a simple task as reacting to a stimulus, no
problem solving strategies need to be employed . However, this does not necessarily mean that colour
affects mood which in turn affects selection of processing strategy, as he claimed. Other mechan isms
may be at play, for instance colour may prime concepts which in turn affect strategies (Macrae et al.)
without activating mood. This sort of priming perhaps is also not activated on such a simple task which
involves no cognitive effort.
However, a study by Kwallek et al (1988) showed that red did increase arousal, and it is possible that in
our experiment too colour did affect arousal, but that we did not discern this because any effect was
canceled out by differences between subjects ’ resting levels of arousal. The Yerkes-Dodson law states
that there is an optimal level of arousal needed to perform well; lower or higher than this the
performance is impaired. Perhaps subjects with a high resting level of arousal were excessively arous ed
by red screens so that their performance was impaired, whereas with blue screens they were calmed and
performed well. However, those with a low resting level were conversely over-calmed by blue and
aroused to the optimal level by red. Assuming that the re was an equal number of subjects of each type,
then there may appear to be no difference between the colours, yet if resting levelshad been taken into
account colour may have shown an activating effect.
Another possibility is that the task was so simple that even if colour had increased arousal or induced a
"fast response" prime or a particular mood, the effect would have been minimal since the subjects were
able to perform at a very high level alread y. In other words; there may have been a ceiling effect and
any subtle differences caused by colour would thus not be discernible.
How might the study have been improved?
There was a high degree of variation between subjects, but this does not challenge the ceiling effect
hypothesis. Subjects may each have performed at their personal best possible levels. To ascertain this for
http://psy.st-andrews.ac.uk/resources/colour1.html (4 of 5)09/03/2006 09:54:02
certain, a within subjects design could have been used. This would have given the advantage of allowing
us to compare the effects of colours on individuals, allowing measures of initial resting levels of arousal
to be considered as mentioned earlier. The priming and mood hypotheses too allow f or the possibility
that moods or associated concepts induced or primed by colours may vary between individuals.
However, a within subjects design would have been impractical because it would then have become
obvious to the subjects that we were testing the effect of colour. There would then have been a risk of
the study being confounded by their simply respondin g as they thought was expected of them. We
would not then have been measuring the effect of coloured backgrounds alone.
Also, perhaps an effect of colour was not found because the subjects were not exposed to the
background screens for a long enough time for any arousal, priming or mood to be induced. On a similar
point the screen constituted only a small part of the su bjects’ environment. Perhaps perceptual
immersion, for instance through one’s surroundings, e.g.. walls, carpets, in a colour would produce a
more dramatic effect.
Future research should disentangle the issues of contrast and colour. If an accurate measure of contrast
between target (or text) and background was used and an effect of colour still found, then the possibility
that high contrast aiding processing cou ld be rejected.
The issue of how exactly colour may affect responses or problem solving should be addressed directly.
Is Soldat’s reference to a mediating effect of mood necessary?
The question of whether colour might affect individuals simply by arousing them (without any
mediating variable), could be clarified by measuring galvanic skin response during the task. This would
clarify also whether individuals did indeed react diffe rently to colours in terms of arousal.
In the case of priming, employing questionnaires asking subjects about the kind of concepts that come to
mind when looking at a certain colour could show whether relevant constructs were in fact primed.
Caution needs to be taken not to influence people in the way questions are phrased, thus ‘channeling’ the
answers. An alternative way of measuring mood, rather than using questionnaires, would be to take an
indirect measure, maybe by presenting an open ended short story on coloured paper and analyzing r
esponses to it. This too would avoid guiding responses by asking, e.g.. "how does this colour make you
feel, happy or sad?"
Furthermore, the effect of the proportion of the visual field that is exposed to colour could be
investigated. Numerous workplaces and institutions have adopted certain coloured decor with the aim of
improving harmony or efficiency. Systematically inve stigating the effect of the amount of colour one is
exposed to could show whether the intuitive expectations that colour should have an effect on our
behaviour are experimentally replicable.
http://psy.st-andrews.ac.uk/resources/colour1.html (5 of 5)09/03/2006 09:54:02