Arch Sex Behav
DOI 10.1007/s10508-012-9945-3
O R I G I N A L P A P E R
Eye Fixations Indicate Men's Preference for Female Breasts
or Buttocks
Bruno Dagnino * Joaquin Navajas * Mariano Sigman
Received: 7 May 2010 / Revised: 15 December 2011 / Accepted: 30 December 2011
O Springer Science+Business Media, LLC 2012
Abstract
Evolutionary psychologists have been interested in
Keywords
Breasts A Buttocks A Eye tracking A
male preferences for particular female traits that are thought to
Sexual preference A Sexual choice
signal health and reproductive potential. While the majority of
studies have focused on what makes specific body traits attrac-
tive--such as the waist-to-hip ratio, the body mass index, and
Introduction
breasts shape and size--there is little empirical research that has
examined individual differences in male preferences for specific
Evolutionary psychologists have been interested in male pref-
traits (e.g., favoring breasts over buttocks). The current study
erences for particular female traits that are thought to signal health
begins to fill this empirical gap. In the first experiment (Study 1),
and reproductive potential (Buss, 2003; Grammer, Fink, Moller,
184 male participants were asked to report their preference
& Thornhill, 2003; Jasienska, Ziomkiewicz, Ellison, Lipson, &
between breasts and buttocks on a continuous scale. We found
Thune, 2004; Symons, 1979; Zaadstra et al., 1993). Most of this
that (1) the distribution of preference was bimodal, indicating
work focuses on traits in isolation, such as the waist-to-hip ratio
that Argentinean males tended to define themselves as favor-
(WHR), the body mass index (BMI) (Singh, 1993a, 1993b;
ing breasts or buttocks but rarely thinking that these traits con-
Tovee, Maisey, Emery, & Cornelissen, 1999), and breast shape
tributed equally to their choice and (2) the distribution was biased
and size (Furnham & Swami, 2007; Furnham, Swami, & Shah,
towards buttocks. In a second experiment (Study 2), 19 male partic-
2006).
ipants were asked to rate pictures of female breasts and buttocks.
Markers of preference can be obtained from questionnaires
This study was necessary to generate three categories of pictures
and also from direct behavioral measures, such as eye movements,
with statistically different ratings (high, medium, and low). In a
which provide additional information of the temporal course of
third experiment (Study 3), we recorded eye-movements of 25
attention to different traits (Dixson, Grimshaw, Linklater, &
male participants while they chose the more attractive between
Dixson, 2010). For example, men look more often at faces they
two women, only seeing their breasts and buttock. We found that
judge to be more attractive (Fink et al., 2008; Shimojo, Simion,
the first and last fixations were systematically directed towards
Shimojo, & Scheier, 2003) and direct the first fixation to the
the self-reported preferred trait.
breasts or the waist more often than the face or the lower body of
a front-posed woman (Dixson, Grimshaw, Linklater, & Dixson,
2009).
While the majority of studies have focused on what makes a
specific body trait attractive, there is little empirical research that
has examined individual differences in male preferences (e.g.,
favoring breasts over buttocks). The current study begins to fill
thisempiricalgap.Weusedself-reportandeye-movementstoexam-
ine whether there were individual differences among Argen-
B. Dagnino A J. Navajas A M. Sigman (&)
tinean males in their preference for buttocks versus breasts and
Integrative Neuroscience Laboratory, Department of Physics,
assessed whether self-reported preferences correlated with non-
University of Buenos Aires, PC 1428 Buenos Aires, Argentina
e-mail: sigman@df.uba.ar
verbal behavioral markers.
123

---

Arch Sex Behav
Study 1
with a breast preference, MPI = (0.45 0.21) while for partic-
ipants with a buttocks preference, MPI = (0.53 0.26). This
Method
effect was significant, t(181) = 2.22, p\.05.
Participants
Study 2
A total of 184 men participated in the first study. All men were
self-reported heterosexual and none of them had participated
Method
in any previous study on female attractiveness judgment.
Participants
Procedure and Measures
The participants in Study 2 were male undergraduate and grad-
Participants were asked to judge the relative contributions
uate students (N = 19, M age, 25.7 years, SD 4.2). None of the
of breasts and buttocks in their female attractiveness judgments.
participants from Study 1 were included in Study 2. All partic-
They were asked to click on the leftmost point of a bar if they con-
ipants were self-reported heterosexual and none of them had par-
sidered that, between these two traits, they solely relied on but-
ticipated in any previous study on female attractiveness judg-
tocks to determine whether a woman was attractive and on the
ment.
rightmost point if they considered that they relied solely on
breasts. Participants were asked to click in the center if they
Procedure
thought that breasts and buttocks equally contributed to their
determination of a woman's attractiveness and in intermediate
A total of 180 publicly available color photographs were down-
positions between these extremes to indicate the relative con-
loaded from different sites on the Internet. Images showing stand-
tributions of breasts and buttocks in their female attractiveness
ing women in underwear were downloaded from different web-
judgments. We refer to this measure as subjective preference
sites, including sites with erotic and pornographic pictures (http://
report (SPR). The range of possible values of SPR was [-1 1]
supertangas.com/). Ninety images corresponded to the breasts of
(Fig. 1a). A SPR of 1 indicates that the participant self-report of
a standing woman seen from the front in underwear (breast
preference was completely biased towards breasts and a SPR of
images). The downloaded photographs were cropped to the min-
-1 towards buttocks.
imal square window in which the breasts were visible. The nip-
ples were not visible in any of the images used in this study. None
Results
of the pictures contained any portion of the face (the chin) or
the belly button. Ninety images corresponded to a standing
Reports of Preferences Between Breasts or Buttocks
woman seen from behind in underwear (buttocks images). Ima-
ges were cropped to the minimal square window in which the but-
We first investigated the distribution of the relative contribution
tocks were fully visible, including the lower back and the upper
of breasts and buttocks in determining female attractiveness as
portion of the legs. Figure 1a shows schematic drawings, but real
measured by the SPR. The distribution of SPR (Fig. 1a) was
pictures were used in this study.
clearly bimodal. Of 184 participants, 109 reported a preference
Pictures were used only if, after cropping to the relevant
for buttocks and 71 a preference for breasts. The probability
portion containing the breasts or the buttocks, the resolution was
that this asymmetry resulted from chance, estimated from a
larger than 250 9 250 pixels. All images were downscaled to
two-tailed test of the binomial distribution was p\.01. Only four
250 9 250 pixels using the bicubic spline interpolation of Matlab
participants responded with an SPR of 0, indicating that they
(see http://www.cs.ait.ac.th/*mdailey/matlab/ for a tutorial and
estimated that breasts and buttocks had equal weights in their
code for image downsampling using Matlab).
determination of female attractiveness. We then defined the
The images were embedded as a patch in the center of a
preference intensity (PI) as the absolute value of SPR and
1024 9 768 image of uniform gray background. (i.e., the pic-
investigated whether this measure was different for men pre-
tures covered about one-third of the vertical extension and about
ferring breasts and buttocks (Fig. 1b). The distribution of PI
one-fourth of the horizontal extension of the screen). Images
indicates that, at higher levels of preference, men preferring
were presented in the center of the screen of a 19-inch View-
buttocks report a more marked preference (9 participants had a
Sonic CRT screen. The screen resolution was always 1024 9
buttocks preference with a PI[0.8 and, in sharp contrast, only
768 pixels. The viewing distance was always 57 cm. Patches
1 participant had a breast preference with a PI[0.8). The mean
containing the breasts or the buttocks had an overall size of
of PI (MPI) confirmed this observation (Fig. 1c). For participants
15(H) 9 15(V) degrees of visual angle.
123

---

Arch Sex Behav
Fig. 1 Biases in the relative contribution of breasts and buttocks derived
and breasts (SPR[0). At higher levels the distribution for buttocks is
from self-reports of preference: a Histogram of participants' SPR shows a
shifted towards higher values than the distribution for breasts. c Difference
clear bimodal distribution. The brightness of the breast and buttock images
in the mean value of PI for participants with a preference biased towards
simply sketches the contribution of each trait at different SPRs. b Histogram
breasts and buttocks
of PI for participants with a preference biased towards buttocks (SPR\0)
Measures
could respond with no time pressure. Participants were asked to
rate the attractiveness of the breasts or buttocks on a continuous
Participants were asked to rate breasts and buttocks images in
scale from 0 to 10, by freely moving the mouse on a horizontal
separate blocks of trials. The order of blocks was varied ran-
line with 20 vertical tick-marks for reference.
domly across participants. Within each block, the order of pre-
sentation of the images also varied randomly. Before starting the
Results
study, participants were shown 15 images (chosen randomly
and independently for each participant from the 90 images),
The mean and SE of response times were 4.8 2.2 s. We grouped
each one for a period of 2.4 s, in order to provide an idea of the
the images of our database in three categories (low, medium, and
dispersion in the pool of images. During the study, participants
high rated images) using a 33% percentile split of the average
123

---

Arch Sex Behav
attractiveness distribution. This was done independently for the
Procedure
breasts and buttocks images. Naturally, there was variability in
participants' judgment of attractiveness. To assure that, despite
Participants performed a total of 40 trials. In each trial, partici-
this variability, the distributions of attractiveness of the three
pants saw breast images in the top-left and top-right quadrants
categories were significantly different, we submitted the ratings
and buttock images in the lower-left and the lower-right quad-
of attractiveness to an ANOVA with category (low, medium, and
rants of the visual field. The images used in Study 3 were the
high) and image type (breasts or buttocks) as main factors. This
same as the ones used in Study 2. Participants were asked to indi-
analysis showed a significant effect of category, F(2, 3306) =
cate whether they judged that a woman whose buttocks and
8.23, p\.01. The breasts-buttocks factor, F(1, 3306)\1, and the
breasts were those presented on the left was more or less attrac-
interaction, F(36, 3306) = 1.05, were not significant. The mean
tive than a woman whose buttocks and breasts were those pre-
and SE for each category were: ``breasts high'' (7.48 0.08),
sented on the right (see Fig. 2). Participants indicated their pref-
``breasts medium'' (6.65 0.04), ``breasts low'' (5.37 0.13),
erence clicking the left (right) button of the mouse if they thought
``buttocks high''(7.43 0.10),``buttocks medium''(6.24 0.05),
that the woman in the left (right) was more attractive.
``buttocks low''(5.07 0.10).
After completing all trials, participants were asked to judge the
relative contributions of breasts and buttocks in their female attrac-
Study 3
tiveness judgments. The way in which we asked participants to
report their preference was exactly the same as in Study 1.
Method
Symmetrical and Asymmetrical Trials
Study 3 was designed
Participants
in a factorial manner. On asymmetrical trials participants had to
choose between a woman with high-rated breasts and low-rated
A third set of 25 male participants (M age 25.6 years, SD 2.9
buttocks and a woman with low-rated breasts and high-rated
years) took part in Study 3. None of these participants had been
buttocks (ratings and categories were derived from Study 2). On
included in the previous experiments or any previous study on
symmetrical trials, participants chose between women with breasts
female attractiveness judgment. All participants were self-reported
and buttocksinthesamecategory.Symmetricaltrialswereincluded
heterosexual.
only for gaze data analysis.
Decision
Factor 1
Factor 2
Choice 1
Choice 2
Presentation
Decision
Fig. 2 Sketch of the experimental design of Study 3, aimed to inves-
movements were recorded during the course of the trial. The central panel
tigate relative weights of gaze towards breasts and buttocks and their
of the figure illustrates the sequence of fixations of a typical trial overlaid on
contribution to female choice. In each trial participants saw breasts
top of the stimulus screen. Each fixation is color-coded to illustrate the
images in the top-left and top-right quadrants and buttocks images in the
temporal sequence in which these fixations occurred. The color bar
lower-left and the lower-right quadrants of the visual field. Participants
indicates, for this specific trial, the time from the presentation of the
were asked to indicate whether they judged that a woman whose buttocks
stimuli (dark gray) until the moment of decision-making (light gray).
and breasts were those presented in the left was more or less attractive than a
Participants indicated their choice clicking with the mouse button to
woman whose buttocks and breasts were those presented on the right. Eye
indicate if they preferred the woman on the left or on the right
123

---

Arch Sex Behav
Behavioral Analysis
For each asymmetrical trial, we assigned
measured the FDC, which has a value of 1 for participants who
a value of 1 when participants chose the high-rated breasts woman
always chose the woman with the higher rated breasts, a value of
and -1 when they chose the high-rated buttocks woman. We then
-1 when the woman with the higher rated buttocks was chosen
averaged this value across trials to obtain, for each participant, the
in all trials and a value of 0 when choices were made evenly. The
factor determining choice (FDC). A FDC of 1 (-1) signifies that a
average FDC across all participants and SE were -0.196
participant always chose the option with a high-rated breasts
0.064. In agreement with participants preference reports, the
(buttocks) image. A FDC of 0 signified that the participant chose
distribution of FDC was significantly biased towards buttocks
the high-rated breasts woman and the high-rated buttocks woman
choice, t(24) = -3.03, p = .0058. Although the correspondence
with equal probability.
between the two measures was not exact, there was a significant
correlation between FDC and SPR. A linear correlation analysis
Eye Tracking
revealed that these two measures were correlated by a slope
significantly different than zero (FDC = B*SPR ? A; B =
Movements of the observers' left eye were recorded with a video-
(0.42 0.11), p\.05; A = (-0.14 0.05), p\.05) (Fig. 3a).
based eye tracker (SR Research EyeLink 2K, http://www.
sr-research.com/) at a sample rate of 1,000 samples/s. All eye
Correlations Between Explicit Preferences for Breasts
movements were labeled as fixations, saccades, and blinks accord-
or Buttocks and the Temporal Course of Gaze
ing to the default parameters of the eye tracker's software.1 The
thresholds for saccade detection were 0.15 for motion, 30/s for
We investigated the sequence of eye movements from the onset
velocity and 9500/s2 for acceleration.
of the trial to the moment of choice. Previous studies had shown
that the last fixations before choice are indicative of preference
Gaze Data Analysis
To quantify whether participants gaze
(Fink et al., 2008; Shimojo et al., 2003). We reasoned that this
was directed towards breasts or buttocks we assigned, to each
could be extended to analyze the contribution of different fac-
fixation, a value of 1 if it was directed towards breasts and -1 if it
tors. We hypothesized that if a participant has a marked pref-
was directed towards buttocks. Fixations directed outside of the
erence for breasts, the last fixations would be directed towards
image were excluded from the average. We refer to this number
the upper quadrants, with the decision based on the breasts of
as the preference value of a fixation. We were interested in
both women. Conversely, if a participant had a marked prefer-
quantifying gaze direction throughout the course of the trial.
ence for buttocks, last fixations should be directed to the lower
Thus, we indexed all fixations locked to the onset of the trial (for
quadrants. In addition, since in our study the location of breasts
example F1 corresponds to the first fixation of the trial). We then
and buttocks was constant throughout the, we expected that the
averaged for each index i (with 1 B i B 9), and for each partic-
first fixation of the trial might also be indicative of preference for
ipant, the preference value of the corresponding fixations across
breasts or buttocks.
all trials. We refer to this measure as fraction of gaze to factor,
Figure 3b, c illustrate the course of gaze during the decision
FGFi. If for a given participant, FGFi was close to 1, it indicates
for all trials in the study for representative participants with high
that the fixation i after the beginning of the trial was typically
and low SPR, respectively. The total number of fixations varied
directed to breasts. Since fixation times were not very variable,
widely in different trials but they suggest a consistent pattern:
an analysis based on time (after the onset of the trial and before
the locations of the first and last fixations of each trial were highly
the response respectively) instead of number of fixations yielded
indicative of their individual subjective reports while intermedi-
virtually identical results.
ate fixations seemed less correlated with their preference.
This is only an illustration of a single (but representative)
Results
participant. To quantify this observation, we calculated for each
fixation i (with 1 B i B 9) the likelihood of a given participant to
Correlations Between Explicit Preferences for Breasts
observe a breasts image (FGFi). For each value of i, we then
or Buttocks and the Factor Determining Choice
measured the linear correlation across all participants between
FGFi and SPR (Fig. 3d). For each index, a high correlation indi-
For all asymmetrical trials (high-rated breasts and low-rated
cates that, across participants, those reporting a preference for
buttocks versus low-rated breasts and high-rated buttocks), we
breasts (buttocks) also tended to look to breasts (buttocks). A cor-
relation close to zero indicates that there was no relation between
1
reported preference and gaze direction. We observed that FGF
We collected only left eye because all subjects had normal vision, with no
1 and
reported amblyopia or dysfunction of convergence which could make eye
SPR were highly correlated (Fig. 3d-i), (q = 0.71, p\.001). In
measurements from both eyes significantly different. Moreover, we only
addition, the y-intercept of this regression (FGF1 = B*SPR ?
use eye-movements to know which image is being viewed (and not the
A; B = (0.74 0.15), p\.001; A = (0.55 0.07), p\.001) was
precise pixel being looked at). Hence, discrepancies of a few pixels do not
positive, indicating that participants by default looked at the
change the results. As a control, we recorded eye-movements binocularly
from 11 (out of 25) participants and the results were identical.
upper-portion of the screen. This reflects an initial location bias
123

---

Arch Sex Behav
(a)

Better
Worst
Prob
Worst
Better
Choice (FDC)
Factor Determining
SPR
Participant with SPR = -0.53
Participant with SPR = 0.43
(b)
(c)
Decision
Fixation
Fixation
(d)
Correlation
Fixation Number (F )
i
i)
ii)
iii)
1
5
9
FGF
FGF
FGF
SPR
SPR
SPR
123

---

Arch Sex Behav
b Fig. 3 Relation between self-reports of preference, gaze and the FDC. a
to a participant's preferred trait. Intermediate fixations were
Linear correlation between the FDC and SPR. While there was not a strict
directed homogeneously to breasts and buttocks, regardless of
one to one correspondence these two measures were significantly corre-
preference. The last fixations of a trial, close to the moment of
lated. b, c Raster plots showing qualitatively the dynamics of gaze fixations
for a participant with marked buttocks (b) and breasts (c) preferences. Each
choice, are redirect to a participant's preferred trait.
row corresponds to a trial and each column to a fixation, indexed from the
beginning of the trial. The color code (gray scale intensity) indicates which
Implications for Human Decision Making
quadrant was observed at each fixation. d Correlation between FGFi and
SPR. For some values of i (i.e., for some fixations indices) the explicit
correlations with the values for each individual participant of FGF
First, at a purely methodological level, our results showed that
i and SPR
are shown in a scatter plot. Correlation values close to 1 indicate a very tight
analysis of eye-movements to determine preference (Fink et al.,
correspondence between both measures. Correlation values close to zero
2008; Shimojo et al., 2003) can be extended to disentangle con-
indicate no relation between these two measures
tributions of individual factors in complex choices. Second, our
results argue that metacognition (the ability of a person to know
(in our studies, breasts were always in the upper quadrants and
which elements govern their choices) for this specific aspect of
buttocks in the lower quadrants) which was nevertheless mod-
decision making is relatively accurate. This observation is not
ulated, as revealed by the linear correlation, by self-reports of
self-evident, since considerable research has shown marked dis-
preference.
sociations between one's meta-understanding of one's own pref-
This correlation decreased throughout the course of the trial
erences and the preferences themselves (Ariely & Wertenbroch,
(e.g., FGF5 = B*SPR ? A; B = (0.01 0.11); A = (-0.15
2002). Such dissociations include sexual behavior and choice
0.05), p\.01; q = 0.01, see Fig. 3d-ii) and became again posi-
in different states of arousal (Ariely & Loewenstein, 2006).
tive during the last fixations (e.g., FGF9 = B*SPR ? A; B =
Our results showed that subjects' awareness of their prefer-
(0.20 0.09), p\.05; A = (0.10 0.04), p\.05; q = 0.40, p\
ences regarding the relative contribution of breasts and buttocks
.05, see Fig. 3d-iii). We emphasize that in our study participants
in female choice were consistent with behavioral and gaze
do not choose between breasts or buttocks images, but rather
makers of preference.
between pairs containing both factors. Thus, gaze to breasts or
Third, our results have implications in complex decision
buttocks images is not indicative of choice or motor response,
making involving many simultaneous factors. A broad literature
but rather of a bias in which of the two factors had a greater
has examined the following question: When we face a complex
influence determining choice.
important decision, how should we deal with it in order for the
outcome to be highly rewarding or for it to maximize happiness
and to minimize negative emotions such as regret (Dijksterhuis,
Discussion
Bos, Nordgren, & Van Baaren, 2006)? Dijksterhuis et al. have
shown that in decisions involving many parameters, conscious
While there are numerous investigations focused on what makes
deliberation of choice may not optimize happiness, while uncon-
a specific body trait attractive, understanding how these traits
scious choices (hunches) may be more efficient. Part of the ratio-
combine on a global judgment of attractiveness remains largely
nale for this argument is the impossibility of the cognitive system to
unexplored. The main objective of this study was to fill this
consciously weigh all the factors due to its limited capacity. Our
empirical gap. Specifically, we asked a narrow and concrete
results showed that different factors are segregated in time, allo-
question: Do Argentinean males rely more on breasts or but-
cating the initial and last moments of the decision according to the
tocks when deciding if a woman is attractive?
specific preferences.
Our results showed that: (1) When Argentinean males were
asked whether they rely more on breasts, on buttocks, or equally
Sociological, Cultural, and Individual Variability
among them, when judging attractiveness, responses showed a
in Sexual Preference for Breasts or Buttocks
bimodal distribution with a significant tendency towards but-
tocks. (2) Participants who responded with a preference for but-
The selection of a sexual partner involves a complex compari-
tocks responded, on average, with a more marked preference
son among possible choices. The result of this selection, as
than participants who reported a preference for breasts. (3) When
happens with other social judgments, is highly dependent on
choosing between highly rated-breasts and low rated-buttocks
previous knowledge retrieved during the comparison process
and low rated breasts and high rated buttocks women, participants
(Mussweiler, 2003). At the same time, this previous knowledge
tended to rely on the factor that they reported as preferred. The
depends strongly on each particular culture, leading to different
correlation between this objective measure of choice and a report
sexual preferences based on sociological, geographical, and
of confidence was not perfect but was significant, showing that
temporal determinants. For instance, the optimal WHR has been
non-verbal reports are informative of male preferences for spe-
shown to be 0.7 for some industrialized societies (Dixson,
cific traits. (4) The dynamics of gaze also correlated with indi-
Dixson, Li, & Anderson, 2007; Henss, 2000) while it is higher
vidual preference. The first fixations of a trial tended to be directed
for less industrialized cultures (e.g., 0.9 among the Matsigenka
123

---

Arch Sex Behav
of Peru; 0.8 in Bakossiland in Cameroon (Dixson, Dixson,
Acknowledgments
Bruno Dagnino and Joaquin Navajas contributed
Morgan, & Anderson, 2007)). There is also evidence showing
equally to this work.
that men from the same culture may prefer different features of a
trait depending on their civil status. A study by Dixson et al.
(2010) reported that married men in New Zealand preferred
References
larger breasts than unmarried when asked to rate the attrac-
tiveness of pictures of breasts. In the same line, a study con-
Ariely, D., & Loewenstein, G. (2006). The heat of the moment: The effect of
ducted in the United Kingdom (Furnham & Swami, 2007)
sexual arousal on sexual decision making. Journal of Behavioral Deci-
showed that breast but not buttock size played an important role
sion Making, 19, 87-98.
in attractiveness judgments made by British undergraduate
Ariely, D., & Wertenbroch, K. (2002). Procrastination, deadlines, and
performance: Self-control by precommitment. Psychological Sci-
students.
ence, 13, 219-224.
These observations raise questions which can only be respon-
Buss, D. M. (2003). The evolution of desire (rev. ed.). New York: Basic
ded in future studies. What is the origin of the bias towards buttock
Books.
preference in Argentinean males? Would this tendency persist in a
Dijksterhuis, A., Bos, M. W., Nordgren, L. F., & Van Baaren, R. B. (2006).
On making the right choice: The deliberation-without-attention effect.
cross-cultural study? While the methodology of the studies dis-
Science, 311, 1005-1007.
cussed above is not directly comparable, they suggest that there
Dixson, B. J., Dixson, A. F., Li, B., & Anderson, M. J. (2007a). Studies of
may be significant variability in the relative contributions of breasts
human physique and sexual attractiveness: Sexual preferences of men
and buttocks in different cultural settings.
and women in China. American Journal of Human Biology, 19, 88-
95.
Moreover, trends in the elements which make a female
Dixson, B. J., Dixson, A. F., Morgan, B., & Anderson, M. J. (2007b).
attractive can change, as revealed by studies measuring Playboy
Human physique and sexual attractiveness: Sexual preferences of men
Playmates (Pettijohn & Jungeberg, 2004), Miss America con-
and women in Bakossiland, Cameroon. Archives of Sexual Behav-
test winners (Wiseman, Gray, Mosimann, & Ahrens, 1992), and
ior, 36, 369-375.
Dixson, B. J., Grimshaw, G. M., Linklater, W. L., & Dixson, A. F. (2009).
models in popular women's magazines (e.g., Vogue, Ladies
Eye-tracking of men's preferences for waist-to-hip ratio and breast
Home Journal) and advertisements across time (Garner, Gar-
size of women. Archives of Sexual Behavior, 40, 43-50.
finkel, Schwartz, & Thompson, 1980; Owen & Laurel-Seller,
Dixson, B. J., Grimshaw, G. M., Linklater, W. L., & Dixson, A. F. (2010).
2000; Voracek & Fisher, 2002). Consistent trends have been
Eye tracking of men's preferences for female breast size and areola
pigmentation. Archives of Sexual Behavior, 40, 51-58.
found and related to popular culture and dieting or clothing fads.
Fink, B., Matts, P. J., Klingenberg, H., Kuntze, S., Weege, B., & Grammer,
In Argentina, one of the most popular men magazines (Revista
K. (2008). Visual attention to variation in female facial skin color
Hombre, http://www.revista-hombre.com.ar) has progressively
distribution. Journal of Cosmetic Dermatology, 7, 155-161.
shifted its covers to images of buttocks (66% of its 18 publica-
Furnham, A., & Swami, V. (2007). Perception of female buttocks and breast
size in profile. Social Behavior and Personality, 35, 1-8.
tions in 2009 were buttocks images). This individual measure by
Furnham, A., Swami, V., & Shah, K. (2006). Body weight, waist-to-hip
itself does not resolve the fundamental issue of causality, which
ratio and breast size correlates of ratings of attractiveness and health.
remains for further investigation. However, it remains possible
Personality and Individual Differences, 41, 443-454.
that popular women's magazines and commercials do not simply
Garner, D. M., Garfinkel, P. E., Schwartz, D., & Thompson, M. (1980).
Cultural expectations of thinness in women. Psychological Reports,
reflect current notions of attractiveness; they also contribute to
47, 483-491.
shape them.
Grammer, K., Fink, B., Moller, A. P., & Thornhill, R. (2003). Darwinian
We emphasize that the choice of our study by no means
aesthetics: Sexual selection and the biology of beauty. Biological
implies that these two traits are the most relevant or especially
Reviews, 78, 385-407.
Henss, R. (2000). Waist-to-hip ratio and female attractiveness: Evidence
distinctive female traits. We simply sought to determine whe-
from photographic stimuli and methodological considerations. Per-
ther they contribute differentially to male choice because we
sonality and Individual Differences, 28, 501-513.
identified that this was--in Argentina--a matter of folkloric
Jasienska, G., Ziomkiewicz, A., Ellison, P. T., Lipson, S. F., & Thune, I.
debate which had not been the subject of rigorous investigation.
(2004). Large breasts and narrow waists indicate high reproductive
potential in women. Proceedings of the Royal Society B: Biological
Indeed, the relevance of the shape of breasts and buttocks in
Sciences, 271, 1213-1217.
modern culture is evident by the number of cosmetic surgical
Mussweiler, T. (2003). Comparison processes in social judgment: Mech-
procedures, more than 1,500,000 in 2008 in the United States.
anisms and consequences. Psychological Review, 110, 472-488.
The most popular procedure is breasts augmentation, with more
Owen, P. R., & Laurel-Seller, E. (2000). Weight and shape ideals: Thin is
dangerously in. Journal of Applied Social Psychology, 30, 979-990.
than 300,000 procedures in 2008. This corresponds to an increase
Pettijohn, T. F., & Jungeberg, B. J. (2004). Playboy playmate curves:
of 46% in the last decade and more than a billion dollars of expen
Changes in facial and body feature preferences across social and
diture (American Society of Plastic Surgeons, http://plasticsurgery.
economic conditions. Personality and Social Psychology Bulletin, 30,
org/). Our study thus addresses in a quantitative and formal manner
1186-1197.
Shimojo, S., Simion, C., Shimojo, E., & Scheier, C. (2003). Gaze bias both
a question of natural science, with broad economical and health
reflects and influences preference. Nature Neuroscience, 6, 1317-
relevance, which has been mostly left to folklore and anecdote.
1322.
123

---

Arch Sex Behav
Singh, D. (1993a). Adaptive significance of female physical attractiveness:
Voracek, M., & Fisher, M. L. (2002). Shapely centrefolds? Temporal
Role of waist-to-hip ratio. Journal of Personality and Social Psychol-
change in body measures: Trend analysis. British Medical Journal,
ogy, 65, 293-307.
325, 1447-1448.
Singh, D. (1993b). Body shape and women's attractiveness. Human
Wiseman, C. V., Gray, J. J., Mosimann, J. E., & Ahrens, A. H. (1992).
Nature, 4, 297-321.
Cultural expectations of thinness in women: An update. International
Symons, D. (1979). The evolution of human sexuality. New York:
Journal of Eating Disorders, 11, 85-89.
Oxford University Press.
Zaadstra, B. M., Seidell, J. C., Van Noord, P. A., te Velde, E. R., Habbema,
Tovee, M. J., Maisey, D. S., Emery, J. L., & Cornelissen, P. L. (1999).
J. D., Vrieswijk, B., et al. (1993). Fat and female fecundity: Pro-
Visual cues to female physical attractiveness. Proceedings of the Royal
spective study of effect of body fat distribution on conception rates.
Society of London Series B: Biological Sciences, 266, 211-218.
British Medical Journal, 306, 484-487.
123

---

Document Outline

• Eye Fixations Indicate Mens Preference for Female Breasts or Buttocks
  • Abstract
  • Introduction
  • Study 1
    • Method
      • Participants
      • Procedure and Measures
    • Results
      • Reports of Preferences Between Breasts or Buttocks
  • Study 2
    • Method
      • Participants
      • Procedure
      • Measures
    • Results
  • Study 3
    • Method
      • Participants
      • Procedure
        • Symmetrical and Asymmetrical Trials
        • Behavioral Analysis
      • Eye Tracking
        • Gaze Data Analysis
    • Results
      • Correlations Between Explicit Preferences for Breasts or Buttocks and the Factor Determining Choice
      • Correlations Between Explicit Preferences for Breasts or Buttocks and the Temporal Course of Gaze
  • Discussion
    • Implications for Human Decision Making
    • Sociological, Cultural, and Individual Variability in Sexual Preference for Breasts or Buttocks
  • Acknowledgments
  • References

---