Journal of Experimental Psychology:
Human Perception and Performance
2008, Vol. 34, No. 4, 884 – 893
Copyright 2008 by the American Psychological Association
0096-1523/08/$12.00 DOI: 10.1037/0096-1523.34.4.884
Are Attractive Men’s Faces Masculine or Feminine? The Importance of
Type of Facial Stimuli
Jennifer L. Rennels
P. Matthew Bronstad and Judith H. Langlois
University of Nevada, Las Vegas
The University of Texas at Austin
The authors investigated whether differences in facial stimuli could explain the inconsistencies in the
facial attractiveness literature regarding whether adults prefer more masculine- or more feminine-looking
male faces. Their results demonstrated that use of a female average to dimorphically transform a male
facial average produced stimuli that did not accurately reflect the relationship between masculinity and
attractiveness. In contrast, use of averages of masculine males and averages of feminine males produced
stimuli that did accurately reflect the relationship between masculinity and attractiveness. Their findings
suggest that masculinity contributes more to male facial attractiveness than does femininity, but future
research should investigate how various combinations of facial cues contribute to male facial attractiveness.
Keywords: facial attractiveness, masculinity, femininity, face perception, averaged faces
O’Toole et al., 1998; Penton-Voak et al., 2001; Scheib, Gangestad,
& Thornhill, 1999), but other studies found that adults perceive
feminine male faces as attractive (Dunkle & Francis, 1990; Little,
Burt, Penton-Voak, & Perrett, 2001; Little & Hancock, 2002;
Penton-Voak et al., 2003; Perrett et al., 1998; Rhodes, Hickford, &
Jeffery, 2000). Our interest lies in exploring why these inconsistencies occur, so that we can better understand the contribution of
masculinity and/or femininity to male facial attractiveness.
A considerable amount of research has been conducted to define
what makes a face attractive. Most researchers agree that averageness, symmetry, and sexually dimorphic cues (masculine and
feminine facial characteristics) all contribute to attractiveness (e.g.,
Fink & Penton-Voak, 2002; Grammer & Thornhill, 1994; Little,
Penton-Voak, Burt, & Perrett, 2002; Rhodes, 2006; Thornhill &
Gangestad, 1993). It is interesting to note that there is greater
agreement among researchers about female than male facial attractiveness (e.g., Thornhill & Gangestad, 1999; Townsend &
Wasserman, 1997; Wiederman & Dubois, 1998), particularly in
regard to the contribution of sexually dimorphic cues. Researchers
agree that femininity, rather than masculinity, contributes to female facial attractiveness (e.g., Fink & Penton-Voak, 2002; Grammer & Thornhill, 1994; Little et al., 2002; Rhodes, 2006; Thornhill
& Gangestad, 1993), and these cues are desired regardless of the
short- or long-term nature of the relationship (e.g., Buss & Barnes,
1986; Buss & Schmitt, 1993).
In contrast, researchers disagree about how masculinity and
femininity cues contribute to male facial attractiveness (Fink &
Penton-Voak, 2002), and the desired cues appear dependent on the
short- or long-term nature of the relationship and the phase of a
woman’s ovulatory cycle (e.g., Johnston, Hagel, Franklin, Fink, &
Grammer, 2001). Specifically, several studies found that adults
perceive masculine male faces as attractive (Brown, Cash, &
Noles, 1986; Cunningham, Barbee, & Pike, 1990; Dunkle &
Francis, 1996; Grammer & Thornhill, 1994; Johnston et al., 2001;
Beauty Is Good
It is important to determine what makes a male face attractive
because cues that differentiate high and low attractive individuals
elicit discrepant judgments and treatments from perceivers that
favor high attractive individuals, a phenomenon known as the
“beauty is good” stereotype (Dion, Berscheid, & Walster, 1972).
For example, perceivers judge attractive male and female adults as
more occupationally and interpersonally competent, better adjusted, and more socially appealing than their less attractive counterparts. In regard to differential treatment, attractive adults receive
more attention, cooperation, and help and tend to have more
positive interactions than less attractive peers (Langlois et al.,
2000). The responses evoked by an individual’s appearance may
serve to either create a self-fulfilling prophecy (Snyder, Tanke, &
Berscheid, 1977) or reinforce traits already characteristic of that
individual (Scarr & McCartney, 1983) because there are behavioral differences seen in high and low attractive males and females
as well: Attractive adults date more, are more extraverted, are
higher in self-confidence and self-esteem, and enjoy more occupational success and higher wages (Langlois et al., 2000). These
differences are large enough to be observed during actual social
interactions and occur regardless of the perceiver or target’s gender and their familiarity with one another (Langlois et al., 2000).
So, it is clearly important to understand what makes some faces
more attractive than others and subsequently causes attractive
individuals to evoke positive responses from others.
Jennifer L. Rennels, Department of Psychology, University of Nevada,
Las Vegas; P. Matthew Bronstad and Judith H. Langlois, Department of
Psychology, The University of Texas at Austin.
P. Matthew Bronstad is now at Schepens Eye Research Institute, Boston, Massachusetts.
Correspondence concerning this article should be addressed to Jennifer
L. Rennels, Department of Psychology, University of Nevada, Las Vegas,
4505 Maryland Parkway, Box 455030, Las Vegas, NV 89154-5030.
E-mail: [email protected]
884
MALE FACIAL ATTRACTIVENESS
Theories of Attractiveness and the Stimuli Used to Test
the Theory
Over the past 2 decades, various theories of attractiveness have
been developed and empirically tested. In testing these theories,
researchers have devised innovative ways of digitally manipulating facial stimuli to support or refute the various theories. The
manipulations allowed researchers to carefully control stimulus
variation. An unintentional outcome, however, was that certain
manipulations produced results different from unaltered faces. By
unaltered, we mean images of real or natural faces—images of
faces that have not been manipulated by digital or other means.
Averageness
One theory that has received a great deal of attention and
empirical support is that “averaged” faces are attractive (Langlois
& Roggman, 1990). Adults rate faces whose configuration is close
to the population average as attractive, and infants look longer at
these faces than at faces adults rate as low attractive (e.g., Jones &
Hill, 1993; Little & Hancock, 2002; Rhodes & Tremewan, 1996;
Rubenstein, Kalakanis, & Langlois, 1999). Preferences for averageness may signal a preference for health or may be a by-product
of the information-processing system in that people prefer the
central tendency (prototype) of a category (Langlois & Roggman,
1990; Rhodes, 2006).
Stimuli used in studies investigating the effects of averageness
on attractiveness preferences were unaltered facial images, mathematically averaged facial images, and/or images altered to be
more or less average (i.e., anticaricatures and caricatures). Averaged facial images are created by digitizing two images and
computing the mean pixel values of the corresponding facial
features on the two images to produce an average of the two
images. Adults typically rate averaged faces comprised of at least
16 faces as attractive (e.g., Langlois & Roggman, 1990). Individual faces can be altered to be more or less average (anticaricatured
or caricatured) by digitally warping the image toward or away
from an averaged face (Rhodes & Tremewan, 1996). A metaanalysis found that the effect size of averageness on attractiveness
is large and positive, although it is larger for digitally averaged and
anticaricatured– caricatured images than for unaltered facial images (Rhodes, 2006). The discrepancy in effect size suggests that
the type of facial stimuli used is an important consideration in
facial attractiveness research.
Symmetry
Some researchers have suggested that it is the symmetry of
average faces, rather than averageness, that makes the face attractive because symmetry advertises the quality of a potential mate
(e.g., Alley & Cunningham, 1991; Grammer & Thornhill, 1994).
Direct tests of this hypothesis, however, dispel this notion and
support the idea that averageness is necessary and sufficient for a
face to be attractive (e.g., Baudouin & Tiberghien, 2004; Langlois,
Roggman, & Musselman, 1994; Rhodes, Sumich, & Byatt, 1999).
Although facial symmetry is positively correlated with attractiveness (e.g., Grammer & Thornhill, 1994; Jones et al., 2004; PentonVoak et al., 2001; Perrett et al., 1999; Scheib et al., 1999), it
contributes to facial attractiveness independent of averageness, and
885
the importance of averageness does not disappear even when
symmetry is held constant (Rhodes et al., 1999).
Stimuli used in studies investigating the effects of symmetry on
attractiveness preferences were unaltered facial images and images
altered to be more symmetrical in one of the following ways: by
creating chimeras by splitting faces at the midline and copying and
flipping the right (or left) side of the face onto the other side (e.g.,
Langlois et al., 1994) or by blending a facial image with its mirror
image by averaging the two images together (e.g., Rhodes et al.,
1999). Whereas the former technique tends to exaggerate abnormalities in the face, the latter technique confounds averaging with
making the face more symmetrical. In line with these technical
concerns, Rhodes’s (2006) meta-analysis demonstrates that when
chimeras are used, the effect size of symmetry on attractiveness is
large and negative; when blending is used, however, the effect size
is moderate and positive; and when unaltered facial images are
used, the effect size is small and positive. Once again, type of
facial stimuli used appears important.
Masculinity–Femininity Cues
Other researchers have focused on how facial masculinity and
femininity contribute to facial attractiveness more than, or in
addition to, averageness (e.g., Grammer & Thornhill, 1994;
Johnston et al., 2001; Perrett et al., 1998; Rhodes et al., 2000;
Scheib et al., 1999). Masculinity cues in male faces and femininity
cues in female faces should make it easier to identify their gender,
an important factor in mating and attractiveness decisions (e.g.,
Enquist, Ghirlanda, Lundqvist, & Wachtmeister, 2002). This proposed positive correlation between male–female facial attractiveness and masculinity–femininity seems intuitive, but a number of
studies have suggested the correlation between male facial attractiveness and masculinity is negative (Dunkle & Francis, 1990;
Little et al., 2001; Little & Hancock, 2002; Penton-Voak et al.,
2003; Perrett et al., 1998; Rhodes et al., 2000).
Given the importance of type of facial stimuli on the results of
attractiveness studies, we compared the stimuli used in studies
assessing preferences for masculinity and femininity in male faces
to examine potential reasons for the discrepancies in results. Most
studies finding preferences for masculine-looking male faces used
unaltered faces (Brown et al., 1986; Cunningham et al., 1990;
Dunkle & Francis, 1996; Grammer & Thornhill, 1994; O’Toole et
al., 1998; Penton-Voak et al., 2001; Scheib et al., 1999), although
there were some exceptions (e.g., Johnston et al., 2001). In contrast, most studies finding preferences for feminine-looking male
faces used averaged faces that were dimorphically transformed to
look more masculine or more feminine (Little et al. 2001; Little &
Hancock, 2002; Penton-Voak et al., 2003; Perrett et al., 1998;
Rhodes et al., 2000). Dimorphically transformed faces are created
by digitally increasing or decreasing the appearance differences
between an averaged male and averaged female face (Little et al.,
2001; Little & Hancock, 2002; Penton-Voak et al., 2003; Perrett et
al., 1998; Rhodes et al., 2000). Henceforth, we refer to male faces
created via this procedure as dimorphically female-transformed
(with the “female” referring to the gender of face used to make the
transformation) or we specifically refer them to as masculinized or
feminized faces.
Rhodes’s (2006) meta-analysis confirmed that the effect size for
the relationship between masculinity and attractiveness is moder-
886
RENNELS, BRONSTAD, AND LANGLOIS
ate and negative when dimorphically female-transformed faces are
used, whereas the relationship between masculinity and attractiveness is moderate and positive when unaltered faces are used.
Rhodes (2006) concluded that “reported preferences for feminized
male faces appear to be an artifact of using sex continua that do not
adequately capture sexual dimorphism” (p. 211). She also expressed concerns about the extent to which averaged male faces
appropriately capture masculine traits. No empirical research,
however, has directly investigated these concerns. Our goals were
to (a) understand whether dimorphic female-transformations appropriately represented male facial masculinity–femininity and (b)
determine whether averaged male faces could be used to examine
how sexually dimorphic cues affect male facial attractiveness.
There are reasons why using dimorphically female-transformed
faces may be problematic for concluding that feminine-looking
males are more attractive than masculine-looking males. First, the
masculinization and feminization procedures do different things to
the transformed faces. The feminization procedure warps an averaged male face toward an averaged female face, whereas the
masculinization procedure warps an averaged male face away from
an averaged female face. It is possible these differences in techniques of moving toward or away from a female averaged face
contributed to the averageness (anticaricaturing) or distortion (caricaturing) of the resulting face, respectively. There may be some
debate as to whether moving the average of one gender toward the
average of the other gender increases facial averageness, but it
seems more likely to result in increased attractiveness than moving
a face away from an average, which typically makes faces less
average. Moreover, although dimorphic female-transformations
clearly alter the original averaged face to look more masculine or
more feminine, it is unclear whether or not this technique has
external validity. For example, if the neural representations of
men’s and women’s faces are separated to some degree, as Little,
DeBruine, and Jones (2005) argue, then transforming a male face
with a female average is not consistent with how faces are psychologically represented. Thus, it may be problematic to use
feminized or masculinized male faces created via this technique to
test the effects of sexual dimorphism on attractiveness.
Mathematically averaged and symmetrically blended faces
(which require averaging to create) produce effects similar in
direction to effects of unaltered images when researchers test how
averageness and symmetry contribute to facial attractiveness
(Rhodes, 2006). These similarities suggest averaged faces may
appropriately represent characteristics of unaltered facial images.
Although use of averaged faces may inflate effect sizes (Rhodes,
2006), use of averaged faces may prove valuable for testing the
direction of the relationship between sexual dimorphism and attractiveness.
In two studies, we investigated how different image manipulations affect preferences for male facial stimuli. Specifically, we
assessed adults’ attractiveness preferences for dimorphically
female-transformed male facial images that were similar to images
used in previous studies (e.g., Perrett et al., 1998; Rhodes et al.,
2000) and compared those results with adult preferences for averages of masculine and feminine male faces using both a forcedchoice (Experiment 1) and a nonforced-choice method (Experiment 2). We used both methods because most studies finding
preferences for feminine-looking male faces have used dimorphically female-transformed facial images and a forced-choice pro-
cedure (Little et al., 2001; Penton-Voak et al., 2003; Perrett et al.,
1998; Rhodes et al., 2000), whereas most studies finding preferences for masculine-looking male faces used unaltered images and
nonforced-choice methods (Brown et al., 1986; Cunningham et al.,
1990; Dunkle & Francis, 1996; Grammer & Thornhill, 1994;
O’Toole et al., 1998; Penton-Voak et al., 2001; Scheib et al.,
1999).
Experiment 1
The purpose of Experiment 1 was to compare attractiveness
judgments of dimorphically female-transformed male facial images with attractiveness judgments of averages of masculine and
feminine male facial images. To replicate previous research findings, we included a masculinized, feminized, or averaged male
face within the pairs for the dimorphically female-transformed
condition—we created these stimuli using a technique similar to
that of previous researchers (e.g., Perrett et al., 1998; Rhodes et al.,
2000). The averaged-biased condition included a masculine-biased
averaged face created from high masculine male faces, a nonbiased
averaged face created from nonextreme high and low masculine
male faces, or a feminine-biased averaged face created from low
masculine male faces within the pairs—these types of stimuli have
not been previously used in studies assessing the effects of sexual
dimorphism on attractiveness. We used a forced-choice method for
Experiment 1 so that we could directly compare our results with
the results of previous studies—studies that used dimorphically
female-transformed male faces also tended to use a forced-choice
method (e.g., Little et al., 2001; Penton-Voak et al., 2003; Perrett
et al., 1998; Rhodes et al., 2000).
Stimulus Faces (General Information)
For both studies, we created facial stimuli via computer graphic
techniques using a subset of the facial images described here. The
facial images came from a database containing a large sample of
male and female college students posed with neutral facial expressions and whose clothing was masked with a white sheet. The
faces in the database were rated at different times by different
groups of adults using a standardized procedure: At least 40 adults,
with relatively equal amounts of male and female judges, rated the
faces for attractiveness using a 1 (very unattractive) to 5 (very
attractive) Likert scale with high interrater agreement (␣s ⫽ .90 or
greater). From the database, we randomly selected 150 male and
150 female faces with the constraint that faces from low, medium,
and high levels of attractiveness were equally represented within
each gender (50 faces from each attractiveness level) so as to
represent a wide range in facial attractiveness. Mean attractiveness
ratings ranged from 1.03 to 4.06 (M ⫽ 2.39, SD ⫽ 0.82) for the
male faces and from 1.33 to 4.19 (M ⫽ 2.53, SD ⫽ 0.72) for the
female faces. Different groups of adults rated the faces on
masculinity–femininity using a 1 (very feminine) to 5 (very masculine) Likert scale. Interrater agreement was highly reliable (␣ ⫽
.97 for the male faces and .98 for the female faces). Ratings ranged
from 1.80 to 4.65 (M ⫽ 3.53, SD ⫽ 0.52) for the male faces and
from 1.25 to 4.12 (M ⫽ 2.49, SD ⫽ 0.61) for the female faces.
Like previous studies that used unaltered faces, we found that the
correlation between masculinity and attractiveness ratings for the
male faces was positive and significant (r ⫽ .36, p ⬍ .001).
MALE FACIAL ATTRACTIVENESS
We standardized all facial images to be the same size, have the
same background, and be of similar brightness and contrast using
Adobe Photoshop (Version 6.0, Macintosh). Henceforth, we refer
to these faces as the set of male facial images and the set of female
facial images.
Method
Participants. A sample of 77 undergraduate students (34
male) participated in the dimorphically female-transformed condition: 53% Caucasian, 22% Asian American/Pacific Islander, 21%
Hispanic, 3% African American, and 1% did not indicate race or
ethnicity. An additional participant’s data were collected but deleted due to a data-recording error. Another independent sample of
143 undergraduate students (54 male) participated in the averagedbiased condition: 47% Caucasian, 17% Hispanic or Spanish origin,
15% Asian American/Pacific Islander, 10% mixed race, 8% African American, and 3% indicated other or did not indicate race or
ethnicity. We had more participants for the averaged-biased condition because the dimorphically female-transformed condition
was a replication of previously reported research (e.g., Perrett et
al., 1998; Rhodes et al., 2000), whereas the averaged-biased condition was a new study. Participants received credit for partial
fulfillment of their research requirements for an introductory psychology class.
Facial stimuli. For the dimorphically female-transformed condition, the stimuli consisted of three male facial images created via
Morph 2.5 (1992–1994) software. As in previous studies, we
created images of an averaged male face, a masculinized version of
the averaged male face, and a feminized version of the averaged
male face (e.g., Perrett et al., 1998; Rhodes et al., 2000). We
created the averaged male face by morphing together 32 digitized
images of individual male faces randomly selected from the set of
male facial images. Using keypoints placed around the facial
structure and features of one image, we mapped these same keypoints onto the corresponding locations of another facial image so
that the mathematical (pixel) values of the points on the two
images could be averaged to create a 2-face composite. We then
averaged the 2-face composite with another 2-face composite to
create a 4-face composite. The process continued until the composite comprised 32 faces because adults rate 32-face composites
as attractive, and there is little to be gained from adding more faces
to the composite (Langlois & Roggman, 1990).
887
To masculinize the averaged male face, we then created an
averaged female face by morphing together 32 digitized images of
randomly selected female faces from the set of female facial
images. We created the masculinized version of the averaged male
face by warping the averaged male face 50% away from the
averaged female face, exaggerating the appearance differences
between the averaged male and female faces. To create the feminized version of the averaged male face, we warped the averaged
male face 50% toward the averaged female face, reducing the
appearance differences between the averaged male and female
faces (see Perrett et al., 1998 and Rhodes et al., 2000 for details).
Figure 1 shows the masculinized, averaged, and feminized male
facial images.
For the averaged-biased condition, the stimuli consisted of three
averaged faces created via Morph 2.5 software: A feminine-biased
averaged male; a nonbiased averaged male; and a masculinebiased averaged male. We created each of these averages from 32
individual faces using the same procedure as described in the
dimorphically female-transformed condition for creating the averaged face: The feminine-biased averaged male face consisted of
the 32 male faces rated lowest in masculinity from the set of male
facial images, and the masculine-biased averaged male face consisted of the 32 male faces rated highest in masculinity from the set
of male facial images, with 1 exception. There was 1 highly
masculine male face from the top 32 masculine male faces that
could not be averaged with the other male faces; the location of his
features differed drastically from the other males due to his extremely large forehead, creating a blurry image. Thus, the face
rated 33rd in masculinity was included in the masculine-biased
average. To create the nonbiased averaged face, we wanted to
avoid using extremely masculine or extremely feminine faces.
Plus, we wanted to ensure that the nonbiased averaged male had a
facial masculinity as close as possible to the midpoint of the
masculine-biased and feminine-biased averaged faces. Therefore,
we could not use the averaged face included in the dimorphically
female-transformed condition. Instead, we used the 16 faces in the
higher half of the 32 faces rated low in masculinity and the 16
faces in the lower half of the 32 faces rated high in masculinity to
create the nonbiased averaged face. Having the nonbiased average
and the averaged face located approximately midway between the
masculine-biased and feminine-biased facial images and the masculinized and feminized facial images, respectively, allowed us to
Figure 1. The masculinized, averaged, and feminized versions of the digitally transformed male faces used in
Experiments 1 and 2 (dimorphically female-transformed condition).
RENNELS, BRONSTAD, AND LANGLOIS
888
more easily compare the two conditions. See Figure 2 for pictures
of the masculine-biased, nonbiased, and feminine-biased averaged
male facial images.
Procedure. Participants chose which face was more attractive
within one of the following pairs in the dimorphically femaletransformed condition: (a) the masculinized male face paired with
the averaged male face (n ⫽ 25), (b) the masculinized male face
paired with the feminized male face (n ⫽ 26), or (c) the averaged
male face paired with the feminized male face (n ⫽ 26). Participants rated which face was more attractive within one of the
following pairs in the averaged-biased condition: (a) the
masculine-biased averaged male face paired with the nonbiased
averaged male face (n ⫽ 48), (b) the masculine-biased averaged
male face paired with the feminine-biased averaged male face (n ⫽
48), or (c) the nonbiased averaged male face paired with the
feminine-biased averaged male face (n ⫽ 47). Across participants,
we altered the sides on which participants saw the paired faces so
as to avoid any side bias.
In line with research using unaltered male faces that have found
preferences for masculinity, we predicted that participants would
prefer the more masculine than feminine averaged faces in the
averaged-biased condition. This prediction was confirmed: Participants chose the more masculine of the two faces as being more
attractive, ␹2(1, N ⫽ 143) ⫽ 18.19, p ⬍ .001. Table 1 shows
participants’ choice of face according to condition and type of face
pair and demonstrates the direction of effects in each condition
held regardless of the particular face pair. When masculine-biased
or feminine-biased averaged male faces were used as stimuli,
adults preferred more masculine male faces. Dimorphically
female-transformed masculinized and feminized averages, however, produced the opposite results. Because the findings for the
masculine-biased and feminine-biased averages are in agreement
with studies using unaltered male faces, we suggest that facial
averages retain masculinity cues representative of the population
of faces from which they were configured. Thus, masculine-biased
and feminine-biased facial averages may serve as useful stimuli for
testing effects of sexual dimorphism.
Results and Discussion
To determine how the type of face seen in the two conditions
affected participants’ choice of the most attractive face, we conducted a multinomial logistic regression analysis. Our dependent
variable was choice of face, coded as the more masculine or more
feminine of the two faces within the pair. For example, if participants in the dimorphically female-transformed condition chose
the averaged face rather than the masculinized face in Pair 1, we
coded their choice as feminine because it was the more feminine of
the two faces. The independent variables were condition and
gender of participant. As predicted, we found a significant main
effect for condition, ␹2(1, N ⫽ 220) ⫽ 20.32, p ⬍ .001.
To better understand the direction of the condition effect, we
conducted chi-square analyses for choice of face in each condition
across each type of pair. Because the faces in the dimorphically
female-transformed condition were modeled after faces created in
previous studies, we predicted that we would replicate previous
findings—participants would prefer the more feminine than more
masculine versions of the faces (e.g., Perrett et al., 1998; Rhodes
et al., 2000). This prediction was confirmed: Participants chose the
more feminine of the two faces as being more attractive, ␹2(1, N ⫽
77) ⫽ 5.73, p ⫽ .009.
Experiment 2
Most studies using dimorphically female-transformed faces
have also used a forced-choice procedure like the one we used in
Experiment 1 (Little et al., 2001; Penton-Voak et al., 2003; Perrett
et al., 1998; Rhodes et al., 2000). To overcome the general confounding of using forced-choice procedures with dimorphically
female-transformed faces, we reran the study using a nonforcedchoice measure. We expected the nonforced choice procedure to
better represent participants’ preferences. For example, participants may find the faces from which to choose to be equally
attractive or equally unattractive, but the forced-choice method
does not permit this response. The literature is replete with evidence of problems with the reliability of forced-choice methods
used in assessments of personality, achievement motivation, personnel decisions, knowledge, and vocational choices (Grosse &
Wright, 1985; Jackson, Neill, & Bevan, 1973; Nederhof, 1985;
Ray, 1980, 1982; Travers, 1951). The problems of reliability result
from forcing individuals to make decisions that were not necessarily accurate characterizations of the person being rated (i.e.,
overestimating or underestimating the quality of the individual
being rated; Travers, 1951). This problem can be extended to
Figure 2. The masculine-biased, nonbiased, and feminine-biased averaged male faces used in Experiments 1
and 2 (average-biased condition).
MALE FACIAL ATTRACTIVENESS
889
Table 1
Percentage of Participants in Each Condition Choosing the More Masculine or More Feminine Face for Each Type of Pair in
Experiment 1
Dimorphically female-transformed face pairs
Average-biased face pairs
Choice
Masculinizedaveraged
Masculinizedfeminized
Averagedfeminized
Masculinebiased-nonbiased
Masculine-biasedfeminine-biased
Nonbiasedfeminine-biased
Masculine
Feminine
36
64
38
62
35
65
71
29
73
27
60
40
Note.
Significant results are indicated in bold.
studies of attractiveness in that researchers may conclude that the
most commonly chosen face is more attractive than it actually is.
Examining whether the preferences participants exhibited in Experiment 1 maintain or change when a nonforced-choice procedure
is used helps us to determine whether the robustness of the results
depends on method of preference assessment.
Method
Participants. A new sample of 77 undergraduate students (40
female) participated in the dimorphically female-transformed condition: 65% Caucasian, 17% Asian American/Pacific Islander, 9%
Hispanic, 5% African American, and 4% indicated other. An
additional participant’s data were collected but deleted because of
a data-recording error. Another independent sample of 144 undergraduate students (96 female, 1 unknown) participated in the
averaged-biased condition: 42% Caucasian, 23% Hispanic or
Spanish origin, 14% Asian American/Pacific Islander, 11% mixed
race, 8% African American, and 3% indicated other or did not
indicate race or ethnicity. An additional participant failed to complete the data form. Sample size differences for the two conditions
mirrored sample size differences for the two conditions in Experiment 1. Participants received credit for partial fulfillment of their
research requirements for an introductory psychology class.
Facial stimuli. Stimulus faces were the same images used in
Experiment 1 (see Figures 1 and 2).
Procedure. The procedure was similar to Experiment 1 except
that participants could choose one face as more attractive than the
other or choose both faces as equally attractive. This nonforcedchoice procedure enabled us to more easily compare results with
Experiment 1 than would a rating measure, which is a common
measure in studies using unaltered facial images (e.g., O’Toole et
al., 1998; Penton-Voak et al., 2001; Scheib et al., 1999).
Results and Discussion
Because participants had three potential responses for each pair
of stimuli in this study, as opposed to two potential responses in
Experiment 1, there was no statistical analysis that we could use to
directly compare the results of the two studies. We did, however,
execute a multinomial logistic regression analysis similar to the
one in Experiment 1 and compared how responses differed when
participants had the choice of stating both faces were equally
attractive or were forced to choose a single face. Choice of face
was the dependent variable (i.e., choosing the more masculine or
more feminine face as more attractive or both faces as equally
attractive), and condition and participant gender were the independent variables. Once again, we found that participants responded
differently depending on whether they participated in the dimorphically female-transformed or averaged-biased condition, ␹2(2,
N ⫽ 221) ⫽ 27.60, p ⬍ .001.
To determine the direction of effects in each of the two conditions, we executed chi-square analyses for choice of face across
each type of comparison. In the dimorphically female-transformed
condition, participants were again more likely to choose the more
feminine face as being more attractive, ␹2(2, N ⫽ 77) ⫽ 11.64,
p ⫽ .002; this finding, however, differed depending upon the type
of comparison. Within both the feminized and masculinized face
pair and the averaged and masculinized face pair, participants
chose the more feminine face as more attractive. Within the
feminized and averaged face pair, however, participants showed
no significant difference in choice (see Table 2).
Table 2
Percentage of Participants in Each Condition Choosing the More Masculine or More Feminine Face or Both Faces as Equally
Attractive for Each Type of Pair in Experiment 2
Dimorphically female-transformed face pairs
Average-biased face pair
Choice
Masculinizedaveraged
Masculinizedfeminized
Averagedfeminized
Masculine-biasednonbiased
Masculine-biasedfeminine-biased
Nonbiased-femininebiased
Masculine
Feminine
Both
20
60
20
19
58
23
19
35
46
59
14
27
64
19
17
33
31
35
Note.
Significant results are indicated in bold.
RENNELS, BRONSTAD, AND LANGLOIS
890
In the averaged-biased condition, participants were more likely
to choose the more masculine face as being more attractive, ␹2(2,
N ⫽ 144) ⫽ 23.29, p ⬍ .001. This finding, however, differed
depending upon the type of comparison. Within both the
masculine-biased and nonbiased averaged face pair and the
masculine-biased and feminine-biased averaged face pair, participants chose the more masculine face as more attractive. Within the
nonbiased and feminine-biased averaged face pair, however, participants showed no significant choice (see Table 2).
The results in the forced-choice and nonforced-choice versions
of the study are similar, but the differences within the averaged and
feminized pairs and the nonbiased and feminine-biased averaged
pairs in the two different types of procedures are informative (cf.
Table 1 with Table 2). Note that when either the masculinized or
masculine-biased averaged face is not included in the comparison
and participants have the opportunity to rate both faces as equally
attractive, they perform differently (they show no clear choice)
than when the masculinized or masculine-biased averaged face is
included in the comparison (they are more likely to demonstrate
little preference for the masculinized male face and a strong
preference for the masculine-biased averaged face). The differences in responding signify that there are important appearance
differences between the masculinized and the masculine-biased
averaged male faces.
General Discussion
The results of male facial attractiveness studies are inconsistent:
Some have found that adults prefer more masculine-looking male
faces, whereas others have found that adults prefer more femininelooking male faces (cf. Brown et al., 1986; Cunningham et al.,
1990; Dunkle & Francis, 1996; Grammer & Thornhill, 1994;
Johnston et al., 2001; O’Toole et al., 1998; Penton-Voak et al.,
2001; Scheib et al., 1999, with Dunkle & Francis, 1990; Little et
al., 2001; Little & Hancock, 2002; Penton-Voak et al., 2003;
Perrett et al., 1998; Rhodes et al., 2000). These differences in
findings have inspired several researchers to provide theoretical
explanations for the discrepancies (e.g., Johnston et al., 2001;
Little et al., 2002; Penton-Voak et al., 1999; Perrett et al., 1998).
In an effort to understand what makes a male face attractive, we
examined whether methodology (i.e., stimulus differences) could
account for these inconsistencies. The results from our studies
strongly suggest that using female facial averages to dimorphically
transform male facial images creates faces that do not accurately
reflect the relationship between attractiveness and masculinity–
femininity.
Differences in Stimuli Cause Differences in Results
In Experiment 1, using a forced-choice procedure, adults chose
the more feminine of the transformed faces as more attractive in
the dimorphically female-transformed condition, preferring the
feminized to masculinized male faces. These results sharply contrasted with those obtained in the average-biased condition,
wherein adults chose the more masculine of the averaged faces as
more attractive, preferring the average of masculine men’s faces to
the average of feminine men’s faces. We obtained similar results
in Experiment 2 using a nonforced-choice procedure, with some
exceptions. Specifically, the results from the nonforced-choice
procedure suggested that participants found the masculinized male
face less attractive than the alternatives and the masculine-biased
averaged male face more attractive than the alternatives. When
those faces were not included in the pairs, there appeared to be no
strong preferences.
The results from the masculine-biased and feminine-biased male
facial averages in the average-biased condition were consistent
with studies that used unaltered facial images (i.e., masculinity was
positively and significantly associated with male facial attractiveness) and were inconsistent with studies that used dimorphically
female-transformed facial images. Whereas the association between masculinity and attractiveness obtained with averaged faces
is congruent with the correlation between attractiveness and masculinity of unaltered faces, the association between masculinity
and attractiveness obtained with dimorphically femaletransformed faces is not. We suggest that averages of a particular
sample of faces may be useful for testing effects of sexual dimorphism on facial attractiveness because averages retain information
about the masculinity–femininity of the face sample.
The results from the dimorphically female-transformed faces in
Experiment 2 suggest participants did not find the masculinized
male attractive, and subsequently defaulted to preferring the feminized male. To better understand why we found this result, refer
to the masculinized male in Figure 1. Recall that this face was
created by warping a male facial average away from a female
facial average, which caused caricaturing of masculine (or nonfeminine) facial characteristics. Caricaturing an averaged male
face away from an averaged female face greatly affects face
width—the masculinized male face in Figure 1 is stretched wide
and is reminiscent of athletes who have used steroids. The distortion from the stretching likely contributed to little preference for
the masculinized male face in Experiments 1 and 2. Moreover,
some researchers have suggested that the neural representations of
men’s and women’s faces are separated to some degree (Little et
al., 2005), so transforming a male facial average with a female
facial average may not be consistent with how faces are psychologically represented. We therefore caution against using othergender images to make facial transformations, at least for studies
examining facial attractiveness and masculinity–femininity.
Dimorphically female-transformed male faces represent the relationship between attractiveness and masculinity–femininity differently from how it is represented in unaltered male and averaged
faces, suggesting that dimorphically female-transformed male
faces are not ecologically valid. These findings point toward
problems with studies that used such images to conclude that
feminine-looking males are attractive (Little et al., 2001; Little &
Hancock, 2002; Penton-Voak et al., 2003; Perrett et al., 1998;
Rhodes et al., 2000). We therefore have caveats about using
dimorphically female-transformed masculinized and feminized
male facial stimuli for testing hypotheses about male facial attractiveness, and we recommend using masculine- or feminine-biased
averages instead.
Implications for Male Facial Attractiveness
Masculinity significantly contributes to male facial attractiveness. Studies using unaltered male faces find a positive relationship between masculinity and attractiveness (Brown et al., 1986;
Cunningham et al., 1990; Dunkle & Francis, 1996; Grammer &
MALE FACIAL ATTRACTIVENESS
Thornhill, 1994; O’Toole et al., 1998; Penton-Voak et al., 2001;
Scheib et al., 1999). The correlation between the attractiveness and
masculinity ratings from our set of 150 male facial images was
significant and positive. In addition, adults rated a masculinebiased averaged male face as more attractive than nonbiased or
feminine-biased averaged males. Thus, a summary representation
of masculine males is considered more attractive than a summary
representation of males in general.
Although there should be no disputing the importance of masculinity, it is clear that because of the moderate correlation between masculinity and attractiveness, masculinity in and of itself
does not make a male face attractive. Participants’ lack of preference for the dimorphically female-transformed, masculinized male
in Experiments 1 and 2 suggests that extremes in masculinity are
not attractive. This finding replicates previous research that found
adults seem to prefer a male face with both mature and babyish
features (Cunningham et al., 1990). Such findings suggest that
different combinations of cues may define male facial attractiveness. For example, because of the established importance of masculinity cues for male facial attractiveness, a greater amount of
masculine than feminine features may make averageness and symmetry cues secondary to the masculinity cues. In contrast, an
extreme amount of masculine or feminine features may make the
presence of averageness and symmetry cues necessary (primary)
for the male face to be considered attractive. Researchers have
hinted at the importance of these different combinations of cues
(e.g., Fink & Penton-Voak, 2002), but little research has specifically manipulated different weightings of averageness, symmetry,
and masculinity–femininity cues to determine their combined role
in contributing to male facial attractiveness. Others have proposed,
in terms of stimulus control theory, that preferences for averageness and sexual dimorphism are solutions to the problem of finding
a mate of the appropriate species and sex (Enquist et al., 2002), but
again the ideal combination of these cues has not been assessed.
Testing such ideas could require digital manipulations of facial
images. Before using digitally manipulated facial images to test a
particular hypothesis regarding male facial attractiveness, we
highly encourage face perception researchers to first test the ecological validity of their facial stimuli.
Implications for Face Perception Research
Given the concerns about the ecological validity of dimorphically female-transformed faces, our findings have implications for
other studies that have used these images (e.g., Penton-Voak,
Jacobson, & Trivers, 2004; Penton-Voak & Perrett, 2000; PentonVoak et al., 1999). For example, researchers have used such faces
to make conclusions about how female preferences for certain
types of male faces shift during the ovulatory cycle (Little et al.,
2002; Penton-Voak & Perrett, 2000; Penton-Voak et al., 1999). If
women’s preferences do shift during their ovulatory period, it is
now unclear what the direction of this shift is. Until studies
assessing how ovulatory phase affects female’s preferences for
male faces are replicated with unaltered facial images, the field is
ill-equipped to make strong conclusions about the nature of these
shifts.
Researchers have also used dimorphically female-transformed
faces to make claims about the relationship between self and
others’ ratings of attractiveness (Penton-Voak et al., 2003) and to
891
explore cross-cultural differences in female preferences for males
(Penton-Voak et al., 2004). In addition, researchers have used
variations of the dimorphic transformation techniques to create
“juvenilized” faces by warping an averaged adult male or averaged
adult female face toward an averaged child male or averaged child
female face, respectively (Hanae, Gyoba, Kamachi, Mukaida, &
Akamatsu, 2004; Ishi & Gyoba, 2001). Because of the caveats
associated with using such transformations for creating stimulus
faces, the results from these studies should be interpreted with
caution until the findings are replicated with unaltered facial
images.
Note that we are not claiming it is never appropriate to use
transformed faces as stimulus materials in research. If the goal is
to gain a better understanding of perceptual processes (e.g., categorical perception), then using transformed facial images may be
appropriate (e.g., Beale & Keil, 1995). If the goal, however, is to
model natural facial variation, then researchers should ensure the
ecological validity of their faces before generalizing conclusions to
actual faces. Indeed, Johnston et al. (2001) used a different procedure for manipulating the masculinity–femininity of male faces
than the procedure we examined in our studies. They found results
more congruent with those using unaltered real faces, so it would
be beneficial to further examine the ecological validity of facial
stimuli created via this technique.
Limitations
Because our sample was restricted in age and race, we do not
claim that these results will necessarily generalize beyond the
population sampled. The sample of faces we used, however, were
very typical of samples used in the published literature (see
Rhodes, 2006, review), which enabled us to compare our results
with those of other published studies. Plus, strong cross-cultural
agreement about who is and is not attractive (Langlois et al., 2000)
suggests that our findings may be generalizable to other face
populations.
Moreover, factors other than stimulus faces may impact the
results of male facial attractiveness studies. Understanding male
facial attractiveness requires an investigation of how stimulus
faces, rater characteristics, context, and the question being asked
(e.g., attractiveness preferences in general vs. mating preferences)
interact with one another to affect preferences.
Conclusions
By exploring the problems associated with certain stimuli used
in male facial attractiveness studies, we hope to have resolved the
debate about whether masculine or feminine male faces are more
attractive. Our results clearly point in favor of masculinity contributing more than femininity to attractiveness in men. With this
issue hopefully resolved, we encourage researchers to assess how
various weightings of cues may contribute to male facial attractiveness rather than to define individual components of or one
universal form of male facial attractiveness. When conducting
such studies, researchers need to use facial stimuli and methods
that are ecologically valid, so that other researchers and reviewers
can feel confident about the generalizability of results to real-life
faces.
RENNELS, BRONSTAD, AND LANGLOIS
892
References
Alley, T. R., & Cunningham, M. R. (1991). Averaged faces are attractive,
but very attractive faces are not average. Psychological Science, 2,
123–125.
Baudouin, J., & Tiberghien, G. (2004). Symmetry, averageness, and feature size in the facial attractiveness of women. Acta Psychologica, 117,
313–332.
Beale, J. M., & Keil, F. C. (1995). Categorical effects in the perception of
faces. Cognition, 57, 217–239.
Brown, T. A., Cash, T. F., & Noles, S. W. (1986). Perceptions of physical
attractiveness among college students: Selected determinants and methodological matters. Journal of Social Psychology, 126, 305–316.
Buss, D. M., & Barnes, M. (1986). Preferences in human mate selection.
Journal of Personality and Social Psychology, 50, 559 –570.
Buss, D. M., & Schmitt, D. P. (1993). Sexual strategies theory: An
evolutionary perspective on human mating. Psychological Review, 100,
204 –232.
Cunningham, M. R., Barbee, A. P., & Pike, C. L. (1990). What do women
want? Facialmetric assessment of multiple motives in the perception of
male facial physical attractiveness. Journal of Personality and Social
Psychology, 59, 61–72.
Dion, K. K., Berscheid, E., & Walster, E. (1972). What is beautiful is good.
Journal of Personality and Social Psychology, 24, 285–290.
Dunkle, J. H., & Francis, P. L. (1990). The role of facial masculinity/
femininity in the attribution of homosexuality. Sex Roles, 23, 157–167.
Dunkle, J. H., & Francis, P. L. (1996). “Physical attractiveness stereotype”
and the attribution of homosexuality revisited. Journal of Homosexuality, 30, 13–29.
Enquist, M., Ghirlanda, S., Lundqvist, D., & Wachtmeister, C. (2002). An
ethological theory of attractiveness. In G. Rhodes & L. A. Zebrowitz
(Eds.), Facial attractiveness: Evolutionary, cognitive, and social perspectives (pp. 127–151). Westport, CN: Ablex.
Fink, B., & Penton-Voak, I. (2002). Evolutionary psychology of facial
attractiveness. Current Directions in Psychological Science, 11, 154 –
158.
Grammer, K., & Thornhill, R. (1994). Human (Homo sapiens) facial
attractiveness and sexual selection: The role of symmetry and averageness. Journal of Comparative Psychology, 108, 233–242.
Grosse, M. E., & Wright, B. D. (1985). Validity and reliability of true-false
tests. Educational and Psychological Measurement, 45, 1–13.
Hanae, I., Gyoba, J., & Kamachi, M., Mukaida, S., & Akamatsu, S. (2004).
Analyses of facial attractiveness on feminised and juvenilised faces.
Perception, 33, 135–145.
Ishi, H., & Gyoba, J. (2001). Analyses of psychological aspects of attractiveness in feminized and juvenilized Japanese faces. Tohoku Psychologica Folia, 60, 29 –34.
Jackson, D. N., Neill, J. A., & Bevan, A. R. (1973). An evaluation of
forced-choice and true-false item formats in personality assessment.
Journal of Research in Personality, 7, 21–30.
Johnston, V. S., Hagel, R., Franklin, M., Fink, B., & Grammer, K. (2001).
Male facial attractiveness: Evidence for hormone-mediated adaptive
design. Evolution & Human Behavior, 22, 251–267.
Jones, B. C., Little, A. C., Feinberg, D. R., Penton-Voak, I. S., Tiddeman,
B. P., & Perrett, D. I. (2004). The relationship between shape symmetry
and perceived skin condition in male facial attractiveness. Evolution and
Human Behavior, 25, 24 –30.
Jones, D., & Hill, K. (1993). Criteria of facial attractiveness in five
populations. Human Nature, 4, 271–296.
Langlois, J. H., Kalakanis, L., Rubenstein, A. J., Larson, A., Hallam, M.,
& Smoot, M. (2000). Maxims or myths of beauty?: A meta-analytic and
theoretical review. Psychological Bulletin, 126, 390 – 423.
Langlois, J. H., & Roggman, L. A. (1990). Attractive faces are only
average. Psychological Science, 1, 115–121.
Langlois, J. H., Roggman, L. A., & Musselman, L. (1994). What is average
and what is not average about attractive faces? Psychological Science, 5,
214 –220.
Little, A. C., Burt, D. M., Penton-Voak, I. S., & Perrett, D. I. (2001).
Self-perceived attractiveness influences human female preferences for
sexual dimorphism and symmetry in male faces. Proceedings of the
Royal Society of London, Series B, 268, 39 – 44.
Little, A. C., DeBruine, L. M., & Jones, B. C. (2005). Sex-contingent face
aftereffects suggest distinct neural populations code male and female
faces. Proceedings of the Royal Society of London, Series B, 272,
2283–2287.
Little, A. C., & Hancock, P. J. B. (2002). The role of masculinity and
distinctiveness in judgments of human male facial attractiveness. British
Journal of Psychology, 93, 451– 464.
Little, A. C., Penton-Voak, I. S., Burt, D. M., & Perrett, D. I. (2002).
Evolution and individual differences in the perception of attractiveness:
How cyclic hormonal changes and self-perceived attractiveness influence female preferences for male faces. In G. Rhodes & L. A. Zebrowitz
(Eds.), Facial attractiveness: Evolutionary, cognitive, and social perspectives (pp. 59 –90). Westport, CT: Ablex Publishing.
Morph. (1992–1994). (Version 2.5) [Computer software]. San Diego, CA:
Gryphon.
Nederhof, A. J. (1985). Methods of coping with social desirability bias: A
review. European Journal of Social Psychology, 15, 263–280.
O’Toole, A. J., Deffenbacher, K. A., Valentin, D., McKee, K., Huff, D., &
Abdi, H. (1998). The perception of face gender: The role of stimulus
structure in recognition and classification. Memory & Cognition, 26,
146 –160.
Penton-Voak, I. S., Jacobson, A., & Trivers, R. (2004). Populational
differences in attractiveness judgements of male and female faces:
Comparing British and Jamaican samples. Evolution and Human Behavior, 25, 355–370.
Penton-Voak, I. S., Jones, B. C., Little, A. C., Baker, S., Tiddeman, B.,
Burt, D. M., & Perrett, D. I. (2001). Symmetry, sexual dimorphism in
facial proportions and male facial attractiveness. Proceedings of the
Royal Society of London, Series B, 1617–1623.
Penton-Voak, I. S., Little, A. C., Jones, B. C., Burt, D. M., Tiddeman,
B. P., & Perrett, D. I. (2003). Female condition influences preferences
for sexual dimorphism in faces of male humans (Homo sapiens). Journal
of Comparative Psychology, 117, 268, 264 –271.
Penton-Voak, I. S., & Perrett, D. I. (2000). Female preference for male
faces changes cyclically: Further evidence. Evolution & Human Behavior, 21, 39 – 48.
Penton-Voak, I. S., Perrett, D. I., Castles, D. L., Kobayashi, T., Burt, D. M.,
Murray, L. K., & Minamisawa, R. (1999, June 24). Menstrual cycle
alters face preference. Nature, 399, 741–742.
Perrett, D. I., Burt, D. M., Penton-Voak, I. S., Lee, K. J., Rowland, D. A.,
& Edwards, R. (1999). Symmetry and human facial attractiveness.
Evolution and Human Behavior, 20, 295–307.
Perrett, D. I., Lee, K. J., Penton-Voak, I., Rowland, D., Yoshikawa, S.,
Burt, D. M., et al. (1998, August 27). Effects of sexual dimorphism on
facial attractiveness. Nature, 394, 884 – 887.
Ray, J. J. (1980). The comparative validity of Likert, projective, and
forced-choice indices of achievement motivation. Journal of Social
Psychology, 111, 63–72.
Ray, J. J. (1982). Machiavellianism, forced-choice formats and the validity
of the F scale: A rejoinder to Bloom. Journal of Clinical Psychology, 38,
779 –782.
Rhodes, G. (2006). Evolutionary psychology of facial beauty. Annual
Review of Psychology, 57, 199 –226.
Rhodes, G., Hickford, C., & Jeffery, L. (2000). Sex-typicality and attractiveness: Are supermale and superfemale faces super-attractive. British
Journal of Psychology, 91, 125–140.
Rhodes, G., Sumich, A., & Byatt, G. (1999). Are average facial configu-
MALE FACIAL ATTRACTIVENESS
rations attractive only because of their symmetry? Psychological Science, 10, 52–58.
Rhodes, G., & Tremewan, T. (1996). Averageness, exaggeration, and facial
attractiveness. Psychological Science, 7, 105–110.
Rubenstein, A. J., Kalakanis, L., & Langlois, J. H. (1999). Infant preferences for attractive faces: A cognitive explanation. Developmental Psychology, 35, 848 – 855.
Scarr, S., & McCartney, K. (1983). How people make their own environments: A theory of genotype3environment effects. Child Development,
54, 424 – 435.
Scheib, J. E., Gangestad, S. W., & Thornhill, R. (1999). Facial attractiveness, symmetry and cues of good genes. Proceedings of the Royal
Society of London, Series B, 266, 1913–1917.
Snyder, M., Tanke, E. D., & Berscheid, E. (1977). Social perception and
interpersonal behavior: On the self-fulfilling nature of social stereotypes.
Journal of Personality and Social Psychology, 35. 656 – 666.
893
Thornhill, R., & Gangestad, S. W. (1993). Human facial beauty: Averageness, symmetry, and parasite resistance. Human Nature, 4, 237–269.
Thornhill, R., & Gangestad, S. W. (1999). Facial attractiveness. Trends in
Cognitive Sciences, 3, 452– 460.
Townsend, J., & Wasserman, T. (1997). The perception of sexual attractiveness: Sex differences in variability. Archives of Sexual Behavior, 26,
243–268.
Travers, R. M. W. (1951). A critical review of the validity and rationale of
the forced-choice technique. Psychological Bulletin, 48, 62–70.
Wiederman, M. W., & Dubois, S. L. (1998). Evolution and sex differences
in preferences for short-term mates: Results from a policy capturing
study. Evolution and Human Behavior, 19, 153–170.
Received April 26, 2006
Revision received July 2, 2007
Accepted December 13, 2007 䡲
E-Mail Notification of Your Latest Issue Online!
Would you like to know when the next issue of your favorite APA journal will be available
online? This service is now available to you. Sign up at http://notify.apa.org/ and you will be
notified by e-mail when issues of interest to you become available!