Manipulation of body odour alters men’s self-confidence and judgements of their visual attractiveness by women

Abstract

Human body odour is important in modulating self-perception and interactions between individuals. Artificial fragrances have been used for thousands of years to manipulate personal odour, but the nature and extent of influences on person perception are relatively unexplored. Here we test the effects of a double-blind manipulation of personal odour on self-confidence and behaviour. We gave to male participants either an aerosol spray containing a formulation of fragrance and antimicrobial agents or an otherwise identical spray that lacked these active ingredients. Over several days, we found effects between treatment groups on psychometric self-confidence and self-perceived attractiveness. Furthermore, although there was no difference between groups in mean attractiveness ratings of men's photographs by a female panel, the same women judged men using the active spray as more attractive in video-clips, suggesting a behavioural difference between the groups. Attractiveness of an individual male's non-verbal behaviour, independent of structural facial features, was predicted by the men's self-reported proclivity towards the provided deodorant. Our results demonstrate the pervasive influence of personal odour on self-perception, and how this can extend to impressions on others even when these impressions are formed in the absence of odour cues.

Full text

Manipulation of body odour alters men’s
self-confidence and judgements of their visual
attractiveness by women
S. Craig Roberts*, A. C. Little, A. Lyndonà, J. Roberts*, J. Havlicek§ and R. L. Wrightà
*School of Biological Sciences, University of Liverpool, Liverpool L69 7BX, UK, School of Psychology, University of
Stirling, Stirling FK9 4LA, UK, àUnilever Research and Development, Port Sunlight Laboratory, Quarry Road East,
Bebington, Wirral, Merseyside, CH63 3JW, UK and §Department of Anthropology, Charles University, Husnikova 2075,
Prague, Czech Republic
Received 4 June 2008, Accepted 15 August 2008
Keywords: deodorant, evolutionary psychology, olfaction, perception, smell
Synopsis
Human body odour is important in modulating
self-perception and interactions between individu-
als. Artificial fragrances have been used for thou-
sands of years to manipulate personal odour, but
the nature and extent of influences on person per-
ception are relatively unexplored. Here we test the
effects of a double-blind manipulation of personal
odour on self-confidence and behaviour. We gave
to male participants either an aerosol spray con-
taining a formulation of fragrance and antimicro-
bial agents or an otherwise identical spray that
lacked these active ingredients. Over several days,
we found effects between treatment groups on psy-
chometric self-confidence and self-perceived attrac-
tiveness. Furthermore, although there was no
difference between groups in mean attractiveness
ratings of men’s photographs by a female panel,
the same women judged men using the active
spray as more attractive in video-clips, suggesting
a behavioural difference between the groups.
Attractiveness of an individual male’s non-verbal
behaviour, independent of structural facial fea-
tures, was predicted by the men’s self-reported pro-
clivity towards the provided deodorant. Our results
demonstrate the pervasive influence of personal
odour on self-perception, and how this can extend
to impressions on others even when these impres-
sions are formed in the absence of odour cues.
Re
´sume
´
L’odeur corporelle humaine a un ro
ˆle important
dans la modulation de la perception de soi et des
interactions entre individus. Les fragrances artifici-
elles sont utilise
´es pour manipuler l’odeur personn-
elle depuis des mille
´naires, mais la nature et
l’e
´tendue de leur influence sur la perception des
personnes restent relativement peu e
´tudie
´es. Nous
testons ici les effets d’une manipulation en double
aveugle de l’odeur personnelle sur la confiance en
soi et le comportement. Nous avons donne
´a
`des
participants masculins soit un spray ae
´rosol con-
tenant une formulation de parfum et des agents
antimicrobiens, soit un spray identique mais
exempt de ces ingre
´dients actifs. Sur plusieurs
jours, nous avons releve
´des diffe
´rences entre les
deux groupes expe
´rimentaux sur le plan des me-
sures psychome
´triques de confiance en soi et de
caracte
`re se
´duisant auto-e
´value
´. En outre, bien
que le caracte
`re se
´duisant des hommes e
´value
´
d’apre
`s leur photo par un panel fe
´minin ne re
´ve
`le
aucune diffe
´rence entre les deux groupes, ces
me
ˆme femmes ont juge
´les hommes utilisant le
Correspondence: S. Craig Roberts, School of Biological
Sciences, University of Liverpool, Liverpool L69 7BX, UK.
Tel.: +44 151 795 4514; fax: +44 151 795 4408;
e-mail: craig.roberts@liverpool.ac.uk
International Journal of Cosmetic Science, 2009, 31,47–54
ª2009 The Authors. Journal compilation
ª2009 Society of Cosmetic Scientists and the Socie
´te
´Franc¸aise de Cosme
´tologie 47

spray actif comme e
´tant plus se
´duisants dans un
clip vide
´o, ce qui sugge
`re une diffe
´rence de com-
portement entre les deux groupes. Le caracte
`re
se
´duisant lie
´au comportement, inde
´pendant de
traits faciaux structuraux, est pre
´dit par la propen-
sion des hommes pour le de
´sodorisant fournie. Nos
re
´sultats de
´montrent l’influence majeure de l’odeur
personnelle sur la perception de soi, et comment
elle peut s’e
´tendre a
`l’impression faite aux autres
me
ˆme lorsque ces impressions se forment en
l’absence d’indices odorants.
Introduction
The importance of the sense of smell in social
interactions and behaviour of animals is well
known [e.g. 1, 2]. In contrast, olfaction has often
been thought to be of minor relevance to primates
in general and humans in particular. Herrick [3]
classified both as microsmatic, on the basis that
olfaction played a minor role in their behaviour.
This perception has been dramatically changing
over the past two decades. Schaal and Porter [4]
directly challenged this classification, showing that
odour coordinates behaviour as wide-ranging as
maternal recognition by infants, peer-group rela-
tionships in pre-pubertal children and reproductive
behaviour (see also [5]). Human body odour also
provides cues of ovulatory status [6], psychometric
dominance [7] and genetic relatedness [8–10].
Despite the potential communicatory signifi-
cance of human body odour, many cultures have
a remarkably uneasy relationship with it. As Stod-
dart [11] puts it: ‘human beings behave as if they
are afraid of smelling like human beings, for
human beings smell bad’. Negative impressions of
body odour are widespread [12] and commonly
associated with outgroup prejudices such as racist
sentiment [13] or attribution of lower social class
[14]. Eli et al. [15] reported that oral malodour
(halitosis) can lead to low self-image and behavio-
ural changes.
At least since the ancient Greek and Egyptian
civilizations [11], individuals have employed exog-
enous fragrances to manipulate their body odour.
Although effects of odours on mood and behaviour
are well-known [16], relatively little research has
examined psychological effects of wearing fra-
grances specifically applied to the body to enhance
or mask personal odour. However, one recent
study familiarized women to a pleasant fragrance
within their skin care product and showed that
this odour later induced psychological and physio-
logical changes associated with a state of increased
relaxation [17]. Beyond effects on the wearer,
wearing fragrances can modulate personality attri-
bution and affect towards the wearer in other peo-
ple, particularly in terms of romantic relationships
and sexual attraction [18, 19] but also in other
contexts such as job interviews [20, 21].
Here we describe an experiment in which we set
out to test the effects of self-perceived personal
odour quality on the self-confidence of young men
and on the attributions made by others. We used
a repeated-measures, within-subject experimental
design in which, after having collected baseline
information, we asked participants to apply an
underarm spray each day. Measures of self-confi-
dence were recorded before application, 15 min
after application and 48 h later. This design
allowed for body odour development and associ-
ated behavioural changes to become manifest. We
divided men into two treatment groups: both
groups of men received visually-identical deodor-
ants specifically prepared for this experiment, but
one group’s spray contained a full commercial
deodorant formulation whereas the other lacked
vital ingredients responsible for fragrance and bac-
tericidal action. We predicted that use of the two
formulations would lead to differences between
groups in self-confidence and self-rated attractive-
ness. We also aimed to test whether any induced
behavioural effects might alter attributions made
by others. To do this, we filmed men while they
recorded an introductory video and showed these
videos to a female panel in the absence of any
odour cues.
Methods
Participants
Participants were 35 heterosexual and non-smok-
ing male students or staff at Liverpool University.
Advertisements were distributed throughout the
campus and on the University’s intranet. An
incentive of £15 payment was offered for partici-
pation. Participants were aged 19–35 (mean ± SD:
23.49 ± 3.43) years. Permission for the study was
granted by the University’s Biological Sciences
Human Ethics Review Board, and all participants
gave informed consent. Participants were informed
that the research aimed to examine the effects of
deodorant use on behaviour, but were provided
ª2009 The Authors. Journal compilation
ª2009 Society of Cosmetic Scientists and the Socie
´te
´Franc¸aise de Cosme
´tologie
International Journal of Cosmetic Science,31,47–5448
Manipulation of body odour alters men’s self-confidence S. Craig Roberts et al.

with no further details and were unaware of the
experimental hypotheses or design; as the same
information was given to all participants, this
could not have accounted for the observed differ-
ence between D+ and D)groups.
Experimental design
A mixed longitudinal design was used. In session
1, participants completed the questionnaires
(described below) to record baseline measure-
ments. At the beginning of session 2, approxi-
mately 24 h later, participants were allocated test
deodorant sprays (see Deodorant formulations) and
asked to apply them immediately. Participants
completed the questionnaires again 15 min after
deodorant application (we chose 15 min because it
balanced the two requirements of ensuring suffi-
cient exposure had occurred for any potential
immediate fragrance-induced effect to occur (cog-
nitive responses to androgen steroids are known to
occur within 6 min, for example, [22, 23]) and
minimizing inconvenience to participants). They
were then instructed to substitute the test deodor-
ant for their usual deodorant for the next 48 h,
allowing sufficient time for body odour to develop
even after thorough cleansing [24], but not too
long so as to inconvenience participants.
Two days later (session 3), they completed the
questionnaires for the final time, and were video-
recorded. Video clips were collected only once, after
2 days of deodorant use, to avoid practice effects.
Deodorants were collected from participants during
session 3. At the end of the study, we asked men to
rate the pleasantness of the deodorant spray that
had been allocated to them, using a Likert scale
(1 = unpleasant, 7 = pleasant). All men confirmed
they had used the allocated sprays each day.
The independent variable was the type of
deodorant given, with assignment alternating in
order of recruitment. Eighteen participants were
provided with a full deodorant (see below) to use
(the D+ group) and 17 were allocated the placebo
(D)group). The experimenter was unaware of the
group to which participants had been allocated.
In an initial background questionnaire, 32 of
the men indicated they used a deodorant spray
either every day or on most days; of the three who
rarely or never normally used deodorants, one
was allocated to the D+ group and two to the D)
group. None of the men used the formulation used
in this experiment.
Deodorant formulations
Formulations were either available as a commer-
cially available product (the D+ formulation) or
form part of one (the D)formulation). Both were
an ethanol solution pressurized with a butane/pro-
pane gas mix. D+ further contained a proprietary
fragrance oil and an antimicrobial ingredient
aimed at reducing malodour. Both deodorants
were prepared in white 150-mL spray cans
marked only with instructions for use and two
contact numbers of experimenters for use in the
event that any adverse reaction occurred; how-
ever, no such reactions were reported.
Questionnaires
In each session, participants completed a question-
naire measuring several variables related to self-
confidence and self-perceived attractiveness. This
included the physical attractiveness, dominance,
assertiveness, self-efficacy, competence and extra-
version scales [25] and Rosenberg’s Self-Esteem
Scale [26]. All scales were taken from the Interna-
tional Personality Items Pool: http://ipip.ori.org/
ipip. We used Spearman rank correlations (as raw
data are measured on an ordinal scale, despite all
variables being normally distributed Kolmogorov–
Smirnov tests, all P> 0.2) to describe correlations
between scores (all from Session 1) on these seven
primary self-confidence constructs, and with a
composite construct, Total Self-Confidence (the
sum of the seven primary constructs). Total Self-
Confidence is strongly correlated with each of the
primary constructs (Table I). In view of this, and
to avoid statistical issues involving multiple testing
if we used each construct separately, analyses
investigating the effects of personal odour manipu-
lation on self-confidence used this composite score.
Raters also completed three independent 5-point
Likert scale questions asking participants to rate
their facial attractiveness, physical attractiveness
and overall appearance [27], and the mean of these
was used as a score of self-rated attractiveness.
Images
During the third test session, participants were
asked to film a short video of themselves in a sce-
nario in which they were instructed to imagine
introducing themselves to an attractive woman.
Participants were shown how to use the digital
ª2009 The Authors. Journal compilation
ª2009 Society of Cosmetic Scientists and the Socie
´te
´Franc¸aise de Cosme
´tologie
International Journal of Cosmetic Science,31, 47–54 49
Manipulation of body odour alters men’s self-confidence S. Craig Roberts et al.

video camera (Sony Handycam DCR-SR52E), and
then the researcher left the room. The camera was
positioned at a distance of 2 m from a chair in
which participants sat, immediately behind which
was a plain white background, in a windowless
room with standardized overhead lighting (fluores-
cent tube). A still face photograph was also taken
from a distance of 2 m, asking participants to
adopt a neutral expression and to look straight
into the camera. Video clips were subsequently
processed (Adobe After Effects 7.0, cropped to
400 ·480 pixels) and edited to a duration of 15 s
(the first 15 s after the subject was judged to be
both seated and talking; videos of 15 s or less pro-
vide sufficient time to make accurate social and
personality judgements [28–30] and attractiveness
judgements in particular are typically made in less
than a second [31, 32]) and encoded as 25 fps
QuickTime movies using the MPEG-4 codec. Pho-
tographs were normalized on interpupillary dis-
tance, cropped just above the top of the head and
just below the chin, and resampled to 400 ·480
pixels (resolution 72 dpi); videos showed seated
men from the waist to just above the head. Seven
men wore glasses (four in the D+ group, three in
the D)group) and three had facial hair (one D+,
two D)); these were constant in both photographs
and videos.
A panel of eight independent female raters (aged
19–26) assessed the facial photographs of the par-
ticipants on a 7-point Likert scale (1 = very unat-
tractive, 7 = very attractive). When watching the
video clips, the same raters judged the participants
with respect to confidence (1 = not at all confi-
dent, 7 = very confident) and attractiveness
(1 = very unattractive, 7 = very attractive). Pho-
tographs were rated before the videos. Video rat-
ings were made twice (with/without sound), in
counter-balanced order. Images of individual males
were presented in randomized order for each par-
ticipant within each of the three types of presenta-
tion, using either a java applet (for photographs)
or Powerpoint (for videos). Both kinds of image
were presented with onscreen dimensions of
27 ·20 cm. Mean scores for each male image
(photograph or video) were calculated and used in
the analyses (Cronbach’s acoefficients were within
acceptable limits for psychological constructs [33],
ranging between 0.65 and 0.77).
Analysis
Data were analysed using repeated-measures
MANOVA, with total self-confidence and self-rated
attractiveness scores as dependent variables, Ses-
sion as the within-subjects measure (with three
levels) and Group (D+ or D)) as a fixed factor. If
the experiment was successful in influencing
self-confidence, we would expect a significant
interaction between Group and Session. Analyses
of judgements of static and dynamic stimuli used
two-tailed independent samples t-tests or Pearson
correlations (distribution of the data fulfilled statis-
tical assumptions).
Results
Mean scores of self-confidence and self-rated
attractiveness across the three test sessions are
shown in Fig. 1. Repeated-measures MANOVA on
the two dependent variables (total self-confidence,
self-rated overall attractiveness) revealed, as pre-
dicted, a significant Session x Group interaction
[F(2,66) = 5.30, P= 0.007]. There were no main
effects of Session or Group, indicating that these
variables did not vary systematically across the
three sessions or between the D+ and D)groups.
The significant interaction was attributed more
strongly to changes in self-confidence, [F(2,66) =
3.83, P= 0.027] than to changes in attractiveness
Table I Matrix of correlations between constructs relat-
ing to psychological self-confidence. Data are Spearman
rank correlation coefficients (r
s
) (above) and exact two-
tailed Pvalues (below). Significant correlations are high-
lighted in bold
SEs Ass Com Dom Ext SEf TSC
PA .24 .41 .14 .25 .53 ).15 .56
.158 .015 .436 .142 .001 .395 .000
SEs .21 .16 .16 .42 .25 .57
.232 .368 .354 .012 .144 .000
Ass .36 .63 .55 .43 .74
.036 .000 .001 .010 .000
Com .25 .39 .48 .61
.156 .021 .003 .000
Dom .38 .21 .64
.026 .220 .000
Ext .20 .79
.243 .000
SEf .45
.006
PA, physical attractiveness; SEs, self-esteem; Ass, assertive-
ness; Com, competence; Dom, dominance; Ext, extraversion;
SEf, self-efficacy; TSC, total self-confidence (composite score).
ª2009 The Authors. Journal compilation
ª2009 Society of Cosmetic Scientists and the Socie
´te
´Franc¸aise de Cosme
´tologie
International Journal of Cosmetic Science,31,47–5450
Manipulation of body odour alters men’s self-confidence S. Craig Roberts et al.

[F(2,66) = 2.56, P= 0.085]. Planned contrasts
across sessions revealed that self-confidence dif-
fered between Sessions 1 and 2 [F(1,33) = 8.96,
P= 0.005], whereas attractiveness rating did not
[F(1,33) = 1.84, P= 0.18]. This indicates that
self-confidence was sensitive to even a 15-min
exposure to the deodorants, whereas there was no
comparable short-term effect on self-rated attrac-
tiveness. In contrast, differences between Sessions
1 and 3 approached statistical significance for both
self-confidence [F(1,33) = 3.65, P= 0.065] and
attractiveness [F(1,33) = 4.07, P= 0.052].
We then examined differences in the perception
of individuals in the D+ and D)groups by a panel
of independent female judges. We found no differ-
ence in the attractiveness ratings of facial photo-
graphs of the participants in the two groups
[t(33) = 0.20, P= 0.84; see Fig. 2a]. There was
also no difference in the ratings when video clips
were played with sound, [t(33) = 1.57, P= 0.13],
nor in ratings of participant confidence, either
with or without sound, despite attractiveness and
confidence judgements being positively correlated
[e.g. without sound, r=0.703, n= 35,
P< 0.001; Fig.3]. However, video-rated (without
sound) attractiveness of the male participants was
higher for the D+ group [t(33) = 2.14, P= 0.040;
Fig. 2a].
Furthermore, we also tested these effects while
controlling for facial attractiveness by calculating
standardized residuals from regression with video-
rated attractiveness as the dependent variable and
photograph-rated attractiveness as the indepen-
dent variable, in order to obtain a measure of the
extent to which individual male participants
235
240
245
250
255
123
Self-confidence score
Session
123
Session
3.8
3.6
3.4
3.2
4
Attractiveness
(a)
(b)
Figure 1 Means and standard errors of scores collected
from participants in the D+ (open circles) and D)(closed
circles) groups across three sessions spanning 72 h. (a)
self-confidence, (b) self-rated attractiveness.
–3
–2
–1
0
1
2
3
4
1234567
Residual
Ratin
g
of deodorant
3
3.4
3.8
4.2
4.6
D+ D– D+ D–
Attractiveness
Photo Video
(a)
(b)
Figure 2 Ratings of male facial attractiveness by female
judges. (a) Means (+SE) based on either digital images or
video clips. (b) Standardized residuals of video-rated over
photograph-rated attractiveness and male-assessed odour
pleasantness, for the D+ group (open circles) and
D)group (closed circles). Men who expressed liking for
the deodorant were more likely judged attractive in
video-ratings than expected based on photograph-rated
attractiveness.
ª2009 The Authors. Journal compilation
ª2009 Society of Cosmetic Scientists and the Socie
´te
´Franc¸aise de Cosme
´tologie
International Journal of Cosmetic Science,31, 47–54 51
Manipulation of body odour alters men’s self-confidence S. Craig Roberts et al.

appeared more attractive in the video clips than
predicted based on their static image rating. This
measure thus parses attractiveness of an individ-
ual’s movement and non-verbal behaviour from
attractiveness of structural facial components. The
regression was significant [F(1,34) = 19.1,
P< 0.001, r
2
= .37]. Residuals were higher
among the D+ group [t(33) = 2.59, P= 0.014],
and were predicted by the self-reported pleasant-
ness of the allocated deodorant [r= 0.352,
n= 35, P= 0.038; Fig. 2b].
Discussion
An increasing number of studies demonstrate a
pervasive, and hitherto underestimated, influence
of olfaction on human behaviour. Our results
emphasize the particular importance of combined
personal odour and fragrance in modulating both
self-perception, especially self-confidence, and con-
sequent perception by others. This is emphasized
by the widespread use of exogenous fragrances to
mask or augment body odour in many human
societies [11] and apparent genetic underpinning
of individual fragrance choice [34].
Male participants using active deodorant showed
increases in measures of self-confidence compared
with those who did not, over a period of only
48 h of body odour manipulation. In fact, a
detectable increase was detected in the predicted
direction only 15 min post-application, although
change in self-rated attractiveness was less rapid.
The immediacy of this effect on self-confidence is
perhaps surprising, but consistent with many
other studies in psychological responses to fra-
grances and indeed forms the basis for a growing
industry in provision of ambient fragrances, for
example in marketing psychology [35]. Figure 2
indicates that the significant interaction we
detected between changes in self-rated attractive-
ness and deodorant use was driven mainly by neg-
ative effects in the D)group, with a relatively
homoeostatic effect being evident in the D+ group.
It thus appears that increased intensity of body
odour, probably associated with changes in the
axillary microbial flora, together with absence of
masking fragrance, has a detrimental effect on
self-perception of attractiveness.
Our design allowed us to test any self- or other-
rated effects resulting from the use of the full
deodorant vs. the placebo, but not whether effects
are attributable to the fragrance or the antimicro-
bial agent. This was because we were primarily
interested in investigating whether, rather than
how, such a manipulation would elicit a measur-
able behavioural effect. However, the rapidity
(within 15 min) of the change in self-confidence is
suggestive of an effect of the fragrance in the
deodorant treatment, rather than the presence of
the antimicrobial agents. In contrast, the absence
of these may well have played a role in the declin-
ing self-confidence and self-rated attractiveness
scores in the D)group over the following 48 h,
through a likely increase in body odour intensity
associated with growth of axillary bacterial popu-
lations. Although this conclusion must remain a
preliminary one, the results suggest an interesting
further study, which could be conducted to tease
apart the effects of the positive, pleasant fragrance
contained within the treatment used and the nega-
tive effects of increasing body odour associated
with absence of antimicrobial action. It would also
be of interest to investigate the efficacy of different
fragrances on the reported effects.
The reported changes in men’s self-assessment
could have arisen directly through their own per-
ception of their personal odour and the fragrance,
or alternatively, they could have been influenced
indirectly through positive or negative reactions of
others in interactions during the experimental per-
iod. During debriefing, it was found that eight of
10 men among the D+ group reported that com-
ments had been made about their new deodorant,
of which seven were positive and one was nega-
tive. In contrast, only one (negative) comment
2.5
3
3.5
4
4.5
5
5.5
2.5 3 3.5 4 4.5 5 5.5
Attractiveness
Confidence
Figure 3 Relationship between ratings of male confi-
dence and attractiveness in video clips. Open circles
indicate males in the D+ group, closed circles denote the
D)group.
ª2009 The Authors. Journal compilation
ª2009 Society of Cosmetic Scientists and the Socie
´te
´Franc¸aise de Cosme
´tologie
International Journal of Cosmetic Science,31,47–5452
Manipulation of body odour alters men’s self-confidence S. Craig Roberts et al.

was reported by men in the D)group. These
responses could also have contributed to the
behavioural differences observed in the videos.
However, we think it more likely that the
responses were attributable to direct effects on the
participants themselves, for two reasons; first,
because the increase in self-confidence was higher
after 15 min of exposure than after 48 h; second,
because (within the D+ group) mean video-rated
attractiveness was in fact lower for those men
who had received positive feedback compared with
those who had received none (mean residual ± SE,
0.42 ± 0.39 vs. 0.44 ± 0.34).
The analysis of the video ratings of self-confi-
dence, even though only 15 s of exposure was
shown to the female panel, demonstrated signifi-
cantly higher rated attractiveness of D+ males
compared with males in the D)group. The
absence of a similar effect when raters were asked
to specifically focus on and rate confidence is more
indicative, perhaps, of the perceptual processing of
information in the rater rather than of effects on
the males, because attractiveness and confidence
ratings were strongly correlated. A similar effect
has been found in facial attractiveness research
where raters are found to be relatively poor at dis-
criminating symmetry but good at distinguishing
attractiveness between images that differ in sym-
metry [36]. Further, it has also been shown that
the ability to detect symmetry is dissociated from
preference for it [37], suggesting that the ability to
judge a trait like confidence may also not neces-
sarily be related to how attractive it is found.
Video-recording was carried out only once because
of potential practice effects, and the corresponding
analysis is therefore between-subjects whereas the
other analyses include a within-subjects design. It
is therefore possible that the video-rated attractive-
ness difference between the two groups was coinci-
dental and unrelated to a change in response to
the manipulation. However, we think that this is
unlikely on the basis that (i) critically, there was
no difference in facial attractiveness between the
two groups (the null result for photograph ratings
thus acts as a control for the difference in video
ratings – the between-group difference seen in vid-
eos is not because of the differences in structural
facial features visible only in photographs), (ii)
even though ratings from static and dynamic
images were highly related, the difference between
the D+ and D)groups was larger when we con-
trolled for facial attractiveness than when we did
not, and (iii) the extent to which participants
appeared more or less attractive than expected
based on their static image was predicted by the
degree to which they expressed a liking for the
deodorant they had been given.
Conclusion
Our results show that changes in self-confidence
and self-perceived attractiveness of young men can
be induced by the use of a spray containing fra-
grance and antimicrobial agents compared with a
spray lacking these active ingredients. Further-
more, these changes are associated with an effect
on the attributions made by others on the basis of
visible non-verbal behaviour, even in the absence
of any olfactory cues (see also [20]). This effect
highlights the flexible nature of self-esteem to
respond to rapid changes in one’s own physical
traits through the use of artificial cosmetic prod-
ucts. An individual’s personal odour and the
perfume product chosen may thus influence both
self-perception and impressions formed by others.
Acknowledgements
We thank Unilever Research & Development for
supplying the fragrances, all our participants for
their patience, Camille Ferdenzi for comments and
translation of the abstract, and two anonymous
reviewers for their helpful comments. Anthony Lit-
tle is supported by the Royal Society.
References
1. Gosling, L.M. and Roberts, S.C. Scent-marking by
male mammals: cheat-proof signals to competitors
and mates. Adv. Study Behav. 30, 169–217 (2001).
2. Johansson, B.G. and Jones, T.M. The role of chemical
communication in mate choice. Biol. Rev. 82, 265–
289 (2007).
3. Herrick, C. Neurological Foundations of Behavior. Holt,
New York (1924).
4. Schaal, B. and Porter, R.H. Microsmatic humans
revisited - the generation and perception of chemical
signals. Adv. Study Behav. 20, 135–199 (1991).
5. Mallet, P. and Schaal, B. Rating and recognition of
peers’ personal odors by 9-year-old children: an
exploratory study. J. Gen. Psychol. 125, 47–64
(1998).
6. Singh, D. and Bronstad, P.M. Female body odour is a
potential cue to ovulation. Proc. R. Soc. Lond., B
268, 797–801 (2001).
ª2009 The Authors. Journal compilation
ª2009 Society of Cosmetic Scientists and the Socie
´te
´Franc¸aise de Cosme
´tologie
International Journal of Cosmetic Science,31, 47–54 53
Manipulation of body odour alters men’s self-confidence S. Craig Roberts et al.

7. Havlicek, J., Roberts, S.C. and Flegr, J. Women’s pref-
erence for dominant male odour: effects of menstrual
cycle and relationship status. Biol. Lett. 1, 256–259
(2005).
8. Roberts, S.C., Gosling, L.M., Spector, T.D., Miller, P.,
Penn, D.J. and Petrie, M. Body odor similarity in
non-cohabiting twins. Chem. Senses 30, 651–656
(2005).
9. Wedekind, C., Seebeck, T., Bettens, F. and Paepke,
A.J. MHC-dependent mate preferences in humans.
Proc. R. Soc. Lond., B 260, 245–249 (1995).
10. Roberts, S.C., Gosling, L.M., Carter, V. and Petrie, M.
MHC-correlated odour preferences in humans and
the use of oral contraceptives. Proc. R. Soc. Lond., B
275, 2715–2722 (2008).
11. Stoddart, M. The Scented ape. Cambridge University
Press, Cambridge (1990).
12. Schleidt, M., Hold, B. and Attili, G. A cross-cultural
study on the attitude towards personal odors.
J. Chem. Ecol. 7, 19–31 (1981).
13. Hyde, A. Bodies of Law. Princeton University Press,
Princeton (1997).
14. Largey, G. and Watson, R. The sociology of odors.
Am. J. Sociol. 77, 1021–1034 (1972).
15. Eli, I., Koriat, H., Baht, R. and Rosenberg, M. Self-
perception of breath odor: role of body image and
psychopathologic traits. Percept. Mot. Skills 91,
1193–1201 (2000).
16. Herz, R.S. Influences of odors on mood and affective
cognition. In: Olfaction, Taste and Cognition (Rouby,
C., Schaal, B., Dubois, D., Gervais, R. and Holley, A.,
eds), pp. 160–177. Cambridge University Press, Cam-
bridge (2002).
17. Abriat, A., Barkat, S., Bensafi, M., Rouby, C. and
Fanchon, C. Psychological and physiological evalua-
tion of emotional effects of a perfume in menopausal
women. Int. J. Cosmet. Sci. 29, 399–408 (2007).
18. Baron, R.A. Olfaction and human social behavior -
effects of a pleasant scent on attraction and social
perception. Pers. Soc. Psychol. Bull. 7, 611–616
(1981).
19. Fiore, A.M. Effect of composition of olfactory cues on
impressions of personality. Soc. Behav. Pers. 20, 149–
162 (1992).
20. Baron, R.A. Sweet smell of success - the impact of
pleasant artificial scents on evaluations of job appli-
cants. J. Appl. Psychol. 68, 709–713 (1983).
21. Higuchi, T., Shoji, K., Taguchi, S. and Hatayama, T.
Improvement of nonverbal behaviour in Japanese
female perfume-wearers. Int. J. Psychol. 40, 90–99
(2005).
22. Saxton, T.K., Lyndon, A., Little, A.C. and Roberts,
S.C. Evidence that androstadienone, a putative
human chemosignal, modulates women’s attribu-
tions of men’s attractiveness. Horm. Behav., doi:
10.1016/j.yhbehav.2008.06.001 (2008).
23. Jacob, S. and McClintock, M.K. Psychological state
and mood effects of steroidal chemosignals in women
and men. Horm. Behav. 37, 57–78 (2000).
24. Shelley, W.B., Hurley, H.J. and Nicols, A.C. Axillary
odor: experimental study of the role of bacteria, apo-
crine sweat and deodorants. Arch. Derm. Syphiol. 68,
430–446 (1953).
25. Goldberg, L.R. A broad-bandwidth, public domain,
personality inventory measuring the lower-level fac-
ets of several five-factor models. In: Personality Psy-
chology in Europe (Mervielde, I., Deary, I., Fruyt, F.D.
and Ostendorf, F. eds), pp. 7–28. Tilburg University
Press, Tilburg, The Netherlands (1999).
26. Rosenberg, M. Society and the Adolescent Self-Image.
Princeton University Press, Princeton, NJ (1965).
27. Little, A.C., Burt, D.M., Penton-Voak, I.S. and Per-
rett, D.I. Self-perceived attractiveness influences
human female preferences for sexual dimorphism
and symmetry in male faces. Proc. R. Soc. Lond., B
268, 39–44 (2001).
28. Ambady, N. and Rosenthal, R. Half a minute: pre-
dicting teacher evaluations from thin slices of non-
verbal behavior and physical attractiveness. J. Pers.
Soc. Psychol. 64, 431–441 (1993).
29. Babad, E. Guessing teachers’ differential treatment of
high- and low-achievers from thin slices of their pub-
lic lecturing behavior. J. Nonverbal Behav. 29, 125–
134 (2005).
30. Dabbs, J.M., Bernieri, F.J., Strong, R.K., Campo, R.
and Milun, R. Going on stage: testosterone in greet-
ings and meetings. J. Res. Pers. 35, 27–40 (2001).
31. Olsen, I.R. and Marshuetz, C. Facial attractiveness is
appraised in a glance. Emotion 5, 498–502 (2005).
32. Willis, J. and Todorov, A. First impressions: making
up your mind after a 100-ms exposure to a face.
Psychol. Sci. 17, 592–598 (2006).
33. Kline, P. The Handbook of Psychological Testing (2nd
edition). Routledge, London (1999).
34. Milinski, M. and Wedekind, C. Evidence for MHC-cor-
related perfume preferences in humans. Behav. Ecol.
12, 140–149 (2001).
35. Mitchell, D.J., Kahn, B.E. and Knasko, S.C. There’s
something in the air - effects of congruent or incon-
gruent ambient odor on consumer decision-making.
J. Consum. Res. 22, 229–238 (1995).
36. Scheib, J.E., Gangestad, S.W. and Thornhill, R. Facial
attractiveness, symmetry, and cues to good genes.
Proc. R. Soc. B 266, 1913–1917 (1999).
37. Little, A.C. and Jones, B.C. Attraction independent of
detection suggests special mechanisms for symmetry
preferences in human face perception. Proc. R. Soc.
Lond., B 273, 3093–3099 (2006).
ª2009 The Authors. Journal compilation
ª2009 Society of Cosmetic Scientists and the Socie
´te
´Franc¸aise de Cosme
´tologie
International Journal of Cosmetic Science,31,47–5454
Manipulation of body odour alters men’s self-confidence S. Craig Roberts et al.