Emotion And Motivation II: Sex Differences In Picture Processing
Emotion
Copyright 2001 by the American Psychological Association, Inc.
2001, Vol. 1, No. 3, 300–319
1528-3542/01/$5.00
DOI: 10.1037//1528-3542.1.3.300
Emotion and Motivation II: Sex Differences in
Picture Processing
Margaret M. Bradley, Maurizio Codispoti, Dean Sabatinelli, and Peter J. Lang
University of Florida
Adhering to the view that emotional reactivity is organized in part by underlying
motivational states—defensive and appetitive—we investigated sex differences in
motivational activation. Men’s and women’s affective reactions were measured
while participants viewed pictures with varied emotional and neutral content. As
expected, highly arousing contents of threat, mutilation, and erotica prompted the
largest affective reactions in both men and women. Nonetheless, women showed a
broad disposition to respond with greater defensive reactivity to aversive pictures,
regardless of specific content, whereas increased appetitive activation was apparent
for men only when viewing erotica. Biological and sociocultural factors in shaping
sex differences in emotional reactivity are considered as possible mediators of sex
differences in emotional response.
In “Emotion and Motivation I” (Bradley, Codispoti,
tive disorders, such as anxiety and depression (Nolen-
Cuthbert, & Lang, 2001) we proposed that emotion is
Hoeksema, 1987; Sachs-Ericsson & Ciarlo, 2000),
fundamentally organized by two systems—one appe-
suggesting a higher incidence of negative affect, held
titive and one defensive—that have evolved to medi-
to involve heightened reactions to aversive stimula-
ate transactions in the environment that promote or
tion (Watson, Clark, & Mineka, 1994). Attributing
threaten the survival of individuals and species. In
specifically aversive hyperreactivity to women has
“Emotion and Motivation II,” we explore whether
been questioned, however (Alexander & Wood,
there are sex differences in defensive or appetitive
2000), based on evidence that women also report ex-
activation. In Western culture at least, a persisting
periencing more happiness and joy (e.g., Brody,
stereotype is that women are more emotional than
1996). In the current study, we investigated men’s and
men (Fischer & Manstead, 2000) and, in particular,
women’s reactions to a variety of different pleasant
more reactive to unpleasant events, especially those
and unpleasant picture contents that differentially ac-
that are threatening or traumatic (Allen & Haccoun,
tivate defensive or appetitive motivation.
1976; Kring & Gordon, 1998). Epidemiological stud-
From the motivational perspective advocated here,
ies indicate that women are at a higher risk for affec-
cues that activate appetitive and defensive systems
should be potent for both men and women, who share
survival risks. Nonetheless, differential activation of
these motivational systems could arise. For example,
differences between the sexes in size or strength may
Margaret M. Bradley, Maurizio Codispoti, Dean Sabati-
nelli, and Peter J. Lang, Center for the Study of Emotion and
mediate differential responses to threat, and sociocul-
Attention, University of Florida.
tural learning could selectively reinforce differences
Maurizio Codispoti is now at the Department of Psychol-
in the expression of emotionality in men and women.
ogy, University of Padova, Padova, Italy.
Most theories regarding sex differences in emotional-
This work was supported in part by National Institute of
ity acknowledge the possibility of both biological and
Mental Health Grants P50 MH52384, MH37757, and
sociological contributors to differential emotional ex-
MH43975. We thank Diana Drobes and Jana Axelrad for
perience and expression.
assistance in data acquisition and scoring, and José Soler
Consistent with a biological view, women consis-
Baillo for assistance in preparing this article.
tently report more fear in threatening situations, judg-
Correspondence concerning this article should be ad-
ing that they would be unable to physically protect
dressed to Margaret M. Bradley, Center for the Study of
themselves if attacked (Gordon & Riger, 1991). How-
Emotion and Attention, P.O. Box 100165, Health Sciences
Center, University of Florida, Gainesville, Florida 32610-
ever, the fact that men and women differ particularly
0165.
in facial expressivity and evaluative judgments
300
SEX DIFFERENCES IN EMOTION
301
(LaFrance & Banaji, 1992)—two measures that are
to emotional cues, which is at least partly attributed to
particularly vulnerable to social learning and “volun-
a lack of systematic research in this area (Manstead,
tary” control—suggests that women’s enhanced re-
1992). The data presented in “Emotion and Motiva-
sponse may be due to cultural shaping and reinforce-
tion I” (2001) provided support for the idea that dif-
ment. Numerous studies have found that women are
ferent physiological systems react at different levels
more facially expressive than men (Kring & Gordon,
of defensive and appetitive activation. Thus, for in-
1998; O’Gorman, 1983; Schwartz, Brown & Ahern,
stance, augmented skin conductance responses, a sig-
1980), including when viewing pictures (Lang,
nal of heightened sympathetic nervous system in-
Greenwald, Bradley, & Hamm, 1993), and women
volvement, depend on a relatively high threshold of
also rate emotional pictures more extreme in hedonic
defensive activation. If unpleasant pictures prompt
valence than do men. Conversely, research on sexu-
greater defensive activation for women, we expect
ality has consistently found that, in the context of
increased electrodermal responses when women view
erotic stimuli, men routinely report more emotional
these pictures, compared with men. If women are spe-
arousal than do women (see Murnen & Stockton,
cifically more reactive to contexts that involve threat
1997, for a recent overview). Thus, rather than
and elicit fear, we expect that differential reactivity
women being more emotionally expressive in all af-
should be confined to pictures of attack and mutila-
fective contexts, it is possible that different cues
tion. Similarly, for appetitive stimuli, augmented skin
arouse emotion in men and women.
conductance was obtained only for erotic pictures. If
men are specifically more reactive to these types of
The Research Problem
affective cues, we expect greater reactivity for men,
compared with women.
In the current study, men’s and women’s responses
The prototypical heart rate response when viewing
to a variety of different picture stimuli were explored
unpleasant pictures consists of a sustained decelera-
in multiple response systems, including skin conduc-
tion, interpreted here as indicative of orienting and
tance, heart rate, reflex modulation, facial electro-
attention. If aversive stimuli prompt greater defensive
myographic (EMG) activity, and evaluative judg-
activation in women, we expected greater cardiac de-
ments. This multisystem measurement allowed us to
celeration when women view these pictures, com-
determine whether sex differences are more marked in
pared with men, and particularly in contexts involving
response systems that are less readily influenced by
threat, if women’s hyperreactivity is confined to fear-
social and cultural learning—for example, autonomic
ful cues. Similarly, erotic stimuli prompt an enhanced
responses (i.e., cardiac and electrodermal) and reflex
initial deceleration, compared with other pleasant pic-
modulation—or more readily influenced in systems
ture contents, again attributed to increased orienting
under voluntary control (e.g., facial expression and
and attention, and we expected greater deceleratory
evaluative judgments). The range of affective picture
activity when men view these pictures if erotic stimuli
contents studied here varied from stimuli presumed to
are more activating for male participants.
strongly activate defensive and appetitive systems—
The startle reflex is inhibited for the least arousing,
including threat, mutilation, and erotica—to those that
unpleasant pictures but then rises linearly with in-
are less engaging and are typically rated lower in
creases in the intensity of defensive activation (Brad-
arousal, including, for unpleasant contents, pictures of
ley et al., 2001; Cuthbert et al., 1996). The greatest
pollution, loss, illness, contamination, and accidents
potentiation is found for pictures of threat. For pleas-
and, for pleasant contents, pictures of nature, families,
ant pictures, however, startle magnitude is reduced as
food, adventure, and sports. This design allowed us to
arousal increases, with the greatest inhibition for pic-
determine whether men and women differed generally
tures involving erotica. Reflex modulation is a good
in response to all contexts of a specific valence (e.g.,
index of underlying differences in motivational acti-
unpleasant) or whether specific contexts (e.g., erotic,
vation for men and women, not only because this
threat) may vary in affective meaning and prompt
involuntary reflex is not susceptible to voluntary con-
differential responding on the basis of the partici-
trol, but because it appears to vary systematically with
pant’s sex.
the degree of appetitive or defensive activation. If
erotic stimuli prompt more appetitive activation for
Theoretical Considerations and Hypotheses
men, we expected evidence of greater reflex inhibi-
Past research has not uncovered stable differences
tion when viewing these materials, compared with
between men and women in physiological responses
women; conversely, if women are more reactive in
302
BRADLEY, CODISPOTI, SABATINELLI, AND LANG
general to unpleasant stimuli, or specifically more re-
mutilated bodies). Each picture was presented for 6 s
active to those involving threat, we expected evidence
on a 19-in. (48.3 cm) monitor.
of greater potentiation.
Stimulus control and physiological data acquisition
In both evaluative judgments and facial EMG re-
were accomplished by using an IBM-compatible com-
activity, we expected to replicate previous findings
puter running VPM data acquisition and reduction
that women are more facially expressive than men
software (Cook, 1997). Physiological signals (heart
and, for contexts other than erotica, to report more
rate, skin conductance, corrugator EMG, zygomatic
intense affective experience. Inclusion of multiple af-
EMG, orbicularis oculi EMG) were sampled at 20 Hz
fective contexts allowed us to determine whether the
for 3 s before picture onset, for 6 s during picture
pattern of emotional modulation for different mea-
presentation, and for 2 s after picture offset. Electrode
sures is similar for both men and women, with differ-
placement and data acquisition were as described in
ences mainly in degree of responding, or whether men
Bradley et al. (2001). An acoustic startle stimulus,
and women react differently to specific contexts in-
consisting of a 50-ms presentation, 95 dB (A) burst of
volving, for instance, fear and erotica. Evaluative
white noise, was presented on half of the trials in each
judgments of pleasure and arousal can also be plotted
picture category (i.e., 36 of the 72 pictures), between
in a two-dimensional affective space, which we have
3 and 5 s after picture onset. Pleasure, arousal, and
argued reflects the underlying organization of appeti-
dominance ratings were obtained using the Self-
tive and defensive motivational systems, and these
Assessment Manikin (SAM; Bradley & Lang, 1994;
were assessed separately for male and female partici-
Lang, 1980).
pants. In Experiment 2, we asked men and women to
Procedure.
The procedure was as described in
label their affective experiences when viewing differ-
Bradley et al. (2001). The participant was instructed
ent picture contexts, with the goal of determining
to view each picture as it appeared on the screen, and
whether men and women differ in the emotional la-
to make ratings of pleasure, arousal, and dominance
bels they apply to their emotional experience and how
after picture presentation by using SAM. The partici-
this illuminates differential emotional reactivity.
pant was instructed that brief noises heard over the
headphones could simply be ignored.
Data reduction and analysis.
Reactions in corru-
Experiment 1
gator, zygomatic, and orbicularis oculi EMG, skin
conductance, and heart rate were determined by sub-
Method
tracting activity in the 1 s before picture presentation
Participants.
Participants were 50 women and 45
from that occurring at each half-second after picture
men from a University of Florida introductory psy-
onset. For facial EMG activity, the average changes
chology class who received course credit. Because of
over the 6-s picture period were used to summarize
computer or experimenter error, data for some par-
reactions to picture presentation. For skin conduc-
ticipants and measures were unavailable. Final Ns
tance, the maximum change occurring between 1 and
were as follows: corrugator EMG, n
95; zygomatic
4 s after picture onset was scored, and a log transfor-
EMG, n
95; heart rate, n
94; skin conductance
mation (log [SCR + 1]) was performed to normalize
responses, n
95; startle blink reflexes, n
85 (44
the data. Heart rate waveform scores were computed
women); orbicularis oculi EMG, n
92; and evalu-
by determining, for each participant and each trial, the
ative judgments, n
94.
maximum deceleration from baseline in the first 3 s of
Materials and design.
The materials and design
picture viewing and the peak acceleration from base-
were the same as that described in Part I of Bradley et
line in the last 3 s of picture viewing (c.f. Hodes,
al. (2001). Briefly, 72 pictures were selected from the
Cook, & Lang, 1985). The blink data were reduced
International Affective Picture System1 (Center for the
Study of Emotion and Attention [CSEA–NIH], 1999;
Lang, Bradley, & Cuthbert, 1999) that included four
1 The International Affective Picture System (IAPS;
exemplars in each of 18 different picture contents,
CSEA, 1999) is available on CD-ROM and as photographic
including 8 pleasant (nature, families, food, sports,
slides. The stimulus sets and technical manual (Lang, Brad-
adventure, attractive men, attractive women, erotic
ley, & Cuthbert, 1999) can be obtained on request from the
couples), 2 neutral (household objects, mushrooms)
authors at the Center for the Study of Emotion and Atten-
and 8 unpleasant (pollution, illness, loss, accidents,
tion, P.O. Box 100165 HSC, University of Florida, 32610-
contamination, attacking animals, attacking humans,
0165.
SEX DIFFERENCES IN EMOTION
303
off-line using a program that scored each trial for
unpleasant, arousing quadrant than for men, as illus-
magnitude in analog-to-digital units and onset latency
trated in Figure 1. Thus, women showed a stronger
in milliseconds, using an algorithm devised by Glo-
coupling between ratings of unpleasantness and
bisch, Hamm, Schneider, & Vaitl (1993).
arousal than did men and rated the most unpleasant
pictures as more arousing than did men.
Results
For pleasant pictures (mean pleasure rating >10),
Affective space.
Figure 1 presents each of the 72
however, a different pattern emerged. Now, men
pictures used in this study (solid circles) in the two-
showed a significantly stronger positive correlation (r
dimensional space formed by plotting each picture by
.68) between ratings of pleasure and arousal, com-
its mean pleasantness and arousal rating separately for
pared with women (r
.20; z
3.66, p < .01), and
the men (left panel) and women (right panel). The
for men, pictures tended to project further into the
quadratic correlation between ratings of pleasure and
pleasant, arousing quadrant than for women, as illus-
arousal was significant for both men (r
.59) and
trated in Figure 1. Thus, men showed a stronger cou-
women (r
.49), indicating that as pictures were
pling between pleasure and arousal for pleasant pic-
rated as increasingly more pleasant or more unpleas-
tures than did women, and also tended to find the most
ant, arousal ratings tended to increase as well. The
pleasant pictures more arousing than did women.
vectors describing appetitive and defensive motiva-
Hedonic valence.
Table 1 lists the means for each
tion were determined by computing the correlation
measure as a function of picture valence separately for
between pleasure and arousal ratings separately for
men and women. Significant effects of picture va-
pleasant pictures (appetitive motivation) and unpleas-
lence were obtained for men and women in all mea-
ant pictures (defensive motivation).
sures, except for zygomatic EMG activity, in which
For unpleasant pictures (mean pleasure rating <10),
men did not show differential activity, and orbicularis
women showed a significantly larger (r
−.77; z
oculi EMG activity, in which neither men nor women
2.34, p < .05) negative correlation between ratings of
showed significant differences.
pleasure and arousal, compared with men (r
−.55),
Men and women did not react identically to pleas-
and for women, pictures projected further into the
ant, neutral, and unpleasant pictures, however, as evi-
Figure 1.
Each of the 72 pictures presented in this experiment (solid symbols) is plotted in
the affective space formed by its mean pleasure and arousal rating scale, separately for men
(left panel) and women (right panel) in this study. Open symbols depict International Affec-
tive Picture System (IAPS) pictures that were not presented in this experiment. Compared
with women, men showed a stronger coupling between increased arousal and pleasantness
(i.e., r
.68 vs. r
.20), whereas compared with men, women showed a stronger coupling
between increased arousal and ratings of unpleasantness (i.e., r
−.77 vs. r
−.55).
304
BRADLEY, CODISPOTI, SABATINELLI, AND LANG
Table 1
Mean Ratings of Pleasure, Arousal, and Dominance and Physiological Reactions for Women and Men When Viewing
Pleasant, Neutral, and Unpleasant Pictures
Dependent variable
Pleasant
Neutral
Unpleasant
Main effect
Women
Pleasure ratings (1–20)
15.2
10.0*
3.9*
F(2, 47)
143.0abc
Arousal ratings (1–20)
11.5
4.7
13.8*
F(2, 47)
131.0abc
Dominance ratings (1–20)
12.8*
13.3
6.5
F(2, 47)
60.5bc
Corrugator EMG
( V)
0.54
0.97*
1.45*
F(2, 48)
22.5abc
Orbicularis oculi EMG
( V)
0.81*
0.50
0.63*
F(2, 48)
2.5 ns
Zygomatic EMG
( V)
0.49*
0.14
0.11
F(2, 48)
9.0ab
Skin conductance
(log
S + 1)
0.02
0.02
0.03
F(2, 48)
4.6bc
Initial HR deceleration (bpm)
−4.34
−3.65*
−4.92
F(2, 48)
9.9bc
Peak HR acceleration (bpm)
3.05
3.53
1.83
F(2, 48)
11.5bc
Startle reflex magnitude (T score)
49.2
49.1
51.4
F(2, 47)
10.5bc
Men
Pleasure ratings (1–20)
16.1
10.7*
6.1*
F(2, 43)
150.0abc
Arousal ratings (1–20)
12.5
5.4
12.1*
F(2, 43)
90.0ac
Dominance ratings (1–20)
14.2*
13.4
7.5
F(2, 43)
72.0bc
Corrugator EMG
( V)
0.28
0.39*
0.49*
F(2, 43)
9.5b
Orbicularis oculi EMG
( V)
0.26*
0.38
0.34*
F(2, 43)
1.1 ns
Zygomatic EMG
( V)
0.22*
0.24
0.15
F(2, 43)
0.75 ns
Skin conductance
(log
S + 1)
0.03
0.02
0.03
F(2, 43)
5.4ac
Initial HR deceleration (bpm)
−4.27
−3.37*
−4.87
F(2, 42)
19.9ac
Peak HR acceleration (bpm)
2.43
3.74
1.40
F(2, 43)
13.1abc
Startle reflex magnitude (T score)
49.7
52.4
50.4
F(2, 42)
5.8a
Note.
Initial heart rate is scored as the maximum deceleration in the first 3 s of picture viewing. Peak heart rate is scored as the maximum
acceleration in the last 3 s of picture viewing. EMG
electromyographic; bpm
beats per minute; HR
heart rate.
a Comparison of pleasant versus neutral conditions is significant at p < .05.
b Comparison of pleasant versus unpleasant conditions is significant at p < .05.
c Comparison of unpleasant versus neutral conditions is significant at p < .05.
* p < .05.
denced by significant interactions of sex and picture
.07, and more arousing, F(1, 92)
3.38, p
.069;
valence in pleasure ratings, F(2, 91)
3.45, p
.036;
and reacting with more skin conductance activity,
arousal ratings, F(2, 91)
9.36, p < .001; skin conduc-
F(1, 93)
3.29, p
.07, though these effects were
tance response, F(2, 92)
4.59, p
.013; corrugator
only marginal when averaged across all categories of
EMG activity, F(2, 92)
10.39, p < .001; and zy-
pleasant pictures.
gomatic EMG activity, F(2, 92)
4.28, p
.017.
Second, the pattern of responding to pleasant, neu-
First, these interactions were pursued by simple
tral, and unpleasant pictures was assessed separately
main effects tests comparing the responses of men and
for men and women (see Table 1, right column). As
women to pleasant, neutral, or unpleasant pictures.
expected, both men and women rated pleasant, neu-
Compared with men, women were more reactive to
tral, and unpleasant pictures as significantly different
unpleasant materials, rating these pictures as more
in valence. Whereas women rated unpleasant pictures
arousing, F(1, 92)
7.53, p
.007, and more un-
as more arousing than pleasant pictures, F(1, 48)
pleasant, F(1, 92)
16.72, p < .001; and also react-
28.39, p < .001, men rated these pictures as equiva-
ing with larger changes in corrugator EMG activity,
lently arousing, F < 1. Consistent with the idea that
F(1, 93)
11.67, p
.001. Compared with men,
women found unpleasant pictures more arousing,
women also rated neutral pictures as slightly less
women responded with greater skin conductance
pleasant, F(1, 92)
4.71, p
.03, and reacted with
change when viewing unpleasant, compared with
slightly more corrugator EMG activity, F(1, 93)
pleasant, F(1, 49)
7.49, p
.009, or neutral pic-
7.10, p
.009. Men tended to be more reactive than
tures, F(1, 49)
5.73, p
.02; whereas men re-
women when viewing pleasant pictures, rating these
sponded with equivalent skin conductance changes
pictures slightly more pleasant, F(1, 92)
3.41, p
when viewing pleasant or unpleasant pictures, which
SEX DIFFERENCES IN EMOTION
305
were significantly larger than when viewing neutral
Table 2). However, although women reported greater
pictures (pleasant vs. neutral, F[1, 44]
11.00, p
arousal when viewing unpleasant pictures, the mag-
.002; unpleasant vs. neutral, F[1, 44]
4.46, p
nitude of skin conductance change across different
.04). Although men reacted with greater corrugator
unpleasant picture contents was similar for men and
EMG activity when viewing unpleasant, compared
women, with the largest changes elicited when view-
with pleasant, pictures, F(1, 44)
4.53, p
.01,
ing pictures depicting attack and mutilation for both
women showed more differentiation, with corrugator
men and women (see Figure 3).
EMG activity increasing when viewing unpleasant,
To some extent modulation of the startle reflex was
compared with neutral, pictures, F(1, 49)
8.89, p
also similar for men and women, as the lack of a
.004; and decreasing when viewing pleasant, com-
significant Sex × Stimulus Content interaction and
pared with neutral, pictures, F(1, 49)
12.73, p
Figure 4 (bottom panel) illustrate. When unpleasant
.001. For zygomatic EMG activity, whereas men did
pictures were ordered by arousal ratings, significant
not differentially react to picture valence, greater zy-
linear trends were found for both men, F(1, 40)
gomatic EMG activity was elicited when women
5.57, p
.02, and women, F(1, 43)
19.77, p <
viewed pleasant, compared with neutral, F(1, 49)
.001, indicating that larger reflexes were elicited as
6.45, p
0.14, or unpleasant pictures, F(1, 49)
unpleasant picture contents were rated more arousing.
18.02, p < .001.
For both men and women, the largest blinks were
For the startle reflex, men and women varied in the
elicited when viewing pictures depicting threat (i.e.,
pattern of startle modulation, F(2, 82)
8.57, p <
animal or human attack). In addition, for both men
.001. Significant effects of picture valence on blink
and women, blinks elicited when viewing the most
magnitude were obtained for both men, F(2, 39)
arousing unpleasant contents (i.e., animal and human
5.54, p
.008, and women, F(2, 42)
8.87, p
attack) were significantly larger than when viewing
.001, but the pattern varied. Consistent with other evi-
the most arousing pleasant contents (i.e., erotic
dence of their greater reactivity to unpleasant pictures,
couples and opposite-sex erotica), consistent with past
women showed larger reflexes when viewing unpleas-
research indicating that the largest differences in re-
ant, compared with neutral, F(1, 43)
8.72, p
flex modulation occur for the most highly arousing
.005, or pleasant pictures, F(1, 43)
17.04, p < .001,
content, F(1, 43)
36.32, p < .001, for women; F(1,
whereas men showed smaller reflexes when viewing
40)
18.84, p < .001, for men.
pleasant, compared with neutral, pictures, F(1, 40)
Nevertheless, a main effect of sex in analysis of the
11.15, p
.002. When averaged across all stimulus
reflex data, F(1, 83)
7.78, p
.007, suggested
contents, men tended to show smaller reflexes when
differences between men and women in the relative
viewing unpleasant, compared with neutral, pictures,
size of reflexes elicited when viewing unpleasant pic-
F(1, 40)
5.82, p
.02.
tures, relative to when viewing other pictures. When
Defensive activation.
To assess differential de-
compared with blinks elicited during neutral pictures
fensive reactivity in men and women as it varies with
(M
51.7 and 54.0 for household objects and mush-
specific picture content, we conducted analyses using
rooms, respectively), for instance, men showed sig-
sex and picture content (eight contents: human attack,
nificantly smaller blinks when viewing many of the
animal attack, mutilation, accidents, contamination,
unpleasant contents, including pollution, loss, illness,
illness, loss, and pollution) in a mixed analysis of
disgust, accidents, and mutilations. For women, how-
variance (ANOVA). Main effects of sex, but no in-
ever, blinks elicited when viewing contamination,
teractions involving sex and stimulus content, were
mutilation, animal attack, and human attack were sig-
obtained in both pleasure, F(1, 92)
16.68, p < .001,
nificantly larger than those elicited when viewing
and arousal ratings, F(1, 92)
7.53, p
.007, in-
neutral pictures (M
49.4 and 49.1 for household
dicating that women rated all types of unpleasant pic-
objects and mushrooms, respectively). Thus, com-
tures as more unpleasant and more arousing than did
pared with neutral pictures, men tended to show blink
men, regardless of specific content (see Figure 2).
inhibition when viewing moderately arousing un-
Consistent with reports of greater unpleasantness, a
pleasant pictures, whereas women showed potentia-
main effect of sex was also obtained in initial cardiac
tion when viewing highly arousing unpleasant pic-
deceleration, F(1, 92)
3.97, p
.049, which in-
tures.
dicated that women responded with more cardiac de-
As expected, a main effect of sex was found in both
celeration when viewing unpleasant pictures, com-
corrugator EMG activity, F(1, 93)
11.62, p
pared with men, regardless of specific content (see
.001, and orbicularis oculi EMG activity, with women
306
BRADLEY, CODISPOTI, SABATINELLI, AND LANG
showing more activity over both muscle sites than
men (see Figure 5), and in none of these analyses was
the interaction of sex and specific content2 significant.
Thus, although women responded with larger changes
in facial EMG activity (see Figure 5) when viewing
unpleasant pictures, the pattern of differential activity
as a function of specific content was similar for men
and women.
Appetitive activation.
To assess differences in ap-
petitive reactions for men and women, we conducted
analyses using sex and picture content (seven types:
erotic couples, opposite-sex erotica, adventure, sports,
food, families, nature) in a mixed ANOVA. Men and
women reacted differently to specific pleasant con-
tents, as evidenced by significant interactions between
sex and stimulus content in analyses of pleasure rat-
ings, F(6, 87)
8.75, p < .001; and arousal ratings,
F(6, 87)
5.69, p < .001. A number of these effects
involved erotic materials. Planned comparisons indi-
cated that men rated erotic materials (either couples or
opposite-sex erotica) as more pleasant, F(1, 92)
28.72, p < .001, for erotic couples, F(1, 92)
44.42,
p < .001, for opposite-sex erotica, and more arousing,
F(1, 92)
6.58, p
.01, for erotic couples, F(1, 92)
32.99, p < .001, for opposite-sex erotica, than did
women (see Figure 2).
For skin conductance, a marginal main effect for
sex, F(1, 93)
3.23, p
.07, suggested that men
were more reactive to pleasant stimuli than women,
and consistent with men’s ratings, planned compari-
sons indicated that larger skin conductance changes
were obtained when men viewed opposite-sex erotica,
F(1, 93)
6.09, p
.015, compared with women, as
well as a trend in the same direction when men
viewed erotic couples, F(1, 93)
3.44, p
.07, (see
Figure 3, top panel). When averaged across both
erotic contents, men reacted with larger changes than
did women, F(1, 93)
5.41, p
.02. As illustrated
in Figure 3, men and women did not differ in elec-
trodermal reactions to other pleasant contexts.
The interaction of sex and stimulus content was
marginal in analyses of initial deceleration, F(6, 87)
1.98, p
.08, peak acceleration, F(6, 87)
2.05,
p
.07, corrugator EMG activity, F(6, 88)
2.09,
p
.06, and dominance ratings, F(6, 87)
1.89, p
.09, and these effects were, for the most part, due
Figure 2.
Evaluative judgments of pleasure (top), arousal
(middle), and dominance (bottom) for specific pleasant
2 Differences among specific picture contents within
(open symbols), neutral (shaded symbols), and unpleasant
pleasant and unpleasant valence categories (i.e., main ef-
(solid symbols) picture contents for men (square symbols)
fects of stimulus content) are discussed in Bradley et al.
and women (circular symbols). Asterisks indicate signifi-
(2001).
cant difference in ratings for that content when women were
compared with men (p < .05). SAM
self-assessment
manikin; Opp
opposite; obj
object.
SEX DIFFERENCES IN EMOTION
307
Table 2
Measures of Cardiac Reactivity During Picture Viewing for Men and Women for Specific
Unpleasant, Neutral, or Pleasant Contents
Initial decelerationa
Peak accelerationb
Content
Women
Men
Women
Men
IAPS nos.
Unpleasant pictures
Human attack
−5.82
−5.04
1.03
1.09
3530, 6260, 6350, 6510
Animal attack
−4.86
−3.99
2.80
1.27
1050, 1120 1300, 1930
Mutilation
−5.58
−4.02*
1.56
0.85
3060, 3080, 3110, 3130
Accident
−4.47
−4.46
2.31
1.60
9050, 9600, 9910, 9920
Contamination
−4.88
−4.88
1.88
1.56
7360, 7380, 9300, 9320
Illness
−5.99
−4.86
1.48
1.45
2710, 3180, 3230, 3330
Loss
−5.58
−4.11
1.40
1.50
2205, 2900 9220, 9421
Pollution
−4.91
−4.37
2.21
1.98
9110, 9120, 9330, 9830
Neutral pictures
Mushroom
−4.43
−2.86*
3.14
3.86
7010, 7030, 7040, 7080
Household Object
−3.97
−2.64*
3.91
3.66
5500, 5510, 5520, 5530
Pleasant pictures
Nature
−3.54
−3.45
4.13
3.61
5000, 5760, 5780, 5891
Families
−3.75
−4.75
3.28
1.46*
2070, 2080, 2340, 2360
Food
−4.63
−4.36
4.46
2.43*
7330, 7350, 7400, 7470
Sports
−4.38
−4.30
3.14
2.36
8190, 8200, 8210, 8470
Adventure
−4.28
−3.57
3.09
3.15
8170, 8180, 8370, 8490
Erotic couples
−5.37
−4.21
1.80
1.53
4650, 4660, 4680, 4690
Opposite sex erotica
−5.06
−3.50*
1.63
2.50
4210, 4220, 4250, 4290c
4470, 4490, 4510, 4520d
Same-sex erotica
−4.70
−4.64
1.47
1.15
4470, 4490, 4510, 4520c
4210, 4220, 4250, 4290d
Note.
IAPS
International Affective Picture System (Lang et al., 1999).
a Scored as the maximum heart rate deceleration in the first 3 s of picture viewing. bScored as the
maximum heart rate acceleration in the last 3 s of picture viewing. cData for men. dData for women.
* p < .05.
to differences in men’s and women’s reactions to
activity over these muscles were found primarily
erotic stimuli. When viewing erotic pictures, women
when viewing pleasant pictures that did not involve
responded with significantly larger corrugator EMG
erotic materials, as Figure 5 illustrates. Women re-
changes (opposite-sex erotica, F[1, 93]
6.46, p <
acted with significantly more zygomatic EMG activ-
.013; erotic couples, F[1, 93]
4.49, p
.04; see
ity (see Figure 5, middle panel) when viewing pic-
Figure 5, top panel), greater initial deceleration (op-
tures of families than did men, F(1, 93)
10.88, p
posite-sex erotica, F[1, 92]
5.68, p
.019; erotic
.001; and with more activity in the orbicularis oculi
couples, F[1, 92]
3.52, p
.06; see Table 2), and
EMG for all nonerotic pleasant contents, including
lower ratings of dominance (opposite-sex erotica, F[1,
nature scenes, F(1, 93)
10.90, p
.001; families,
92]
4.65, p
.03; erotic couples, F[1, 92]
F(1, 93)
13.25, p < .001; food, F(1, 93)
6.40, p
11.23, p
.001; see Figure 2), compared with men.
.013; sports, F(1, 93)
5.31, p
.023; and
The single exception was for peak acceleration, in
adventure, F(1, 93)
11.51, p
.001.
which women responded with greater acceleration
Across pleasant picture contents, there were no sig-
than men when viewing pictures of families, F(1, 92)
nificant effects involving sex in startle reflex modu-
4.15, p
.04; or food, F(1, 92)
6.32, p
.014
lation, as Figure 4 illustrates. Both men and women
(see Table 2).
showed the smallest startle responses when viewing
Significant interactions of sex and stimulus content
the most arousing pleasant pictures (i.e., erotic
were obtained for orbicularis oculi EMG activity, F(6,
couples and opposite-sex couples), and when pictures
88)
2.53, p
.026; and zygomatic EMG activity,
were ordered by rated arousal, a comparable linear
F(6, 88)
2.28, p
.04. Differences in facial EMG
relationship was obtained for both men and women.
308
BRADLEY, CODISPOTI, SABATINELLI, AND LANG
Figure 3.
Skin conductance change for women (left panels) and men (right panels) when
viewing specific pleasant (open symbols) and unpleasant (solid symbols) picture contents.
Opp
opposite.
Neutral picture contents.
Women showed signifi-
main effect of sex, F(1,83)
10.71, p
.002, indi-
cantly more initial deceleration to neutral pictures
cated that for men, neutral pictures elicited among the
than did men, F(1, 92)
11.08, p
.001, and this
largest blinks in their distributions (M
51.7 and
was true when viewing either household objects or
54.0 for household objects and mushrooms, respec-
mushrooms (see Table 2). For the startle reflex, a
tively), whereas this was not true for women (M
SEX DIFFERENCES IN EMOTION
309
Figure 4.
Mean startle blink magnitude for women (left panels) and men (right panels) when
viewing specific pleasant (open symbols) and unpleasant (solid symbols) picture contents.
Opp
opposite.
310
BRADLEY, CODISPOTI, SABATINELLI, AND LANG
Figure 5.
Mean changes in corrugator electromyographic (EMG) activity (top), zygomatic
EMG activity (middle), and orbicularis oculi EMG activity (bottom) when viewing specific
pleasant (open bars), neutral (shaded bars), and unpleasant (solid bars) picture contents.
Asterisks in the left panel indicate significant difference in these facial EMG measures for
content when women were compared with men (p < .05). Opp
opposite; obj.
object.
49.4 and 49.1 for household objects and mushrooms,
women, however, skin conductance changes when
respectively). A significant interaction of sex and
viewing attractively posed same-sex nudes were rela-
stimulus content was also obtained for pleasure rat-
tively large and did not significantly differ between
ings, F(1, 92)
7.72, p
.007, which indicated that
the sexes. Women showed more corrugator EMG ac-
women rated mushrooms as slightly less pleasant than
tivity when viewing same-sex erotica than did men,
did men (see Figure 2).
and the magnitude of corrugator EMG changes for
Same-sex erotica.
Table 3 lists the means for
women was not significantly different when viewing
men’s and women’s responses when viewing same-
same-sex erotica as that evoked when viewing un-
sex erotica, as well as the outcome of statistical tests.
pleasant pictures, with the exception of mutilations,
Women rated same-sex erotica as somewhat more
F(1, 49)
12.76, p
.001.
arousing than did men and tended to rate the erotica as
Men and women did not differ in their cardiac re-
slightly more pleasant as well. For both men and
sponses when viewing same-sex erotica; these re-
SEX DIFFERENCES IN EMOTION
311
Table 3
Mean Ratings of Pleasure, Arousal, and Dominance and Physiological Reactions for Women and Men When Viewing
Erotic Pictures of the Same Sex, and Significance Tests
Dependent variable
Women
Men
Significance test results
Pleasure ratings (1–20)
9.3
8.0
F(1, 92)
3.27, p
.07
Arousal ratings (1–20)
9.2
6.4
F(1, 92)
8.65, p
.004
Dominance ratings (1–20)
11.3
10.2
F(1, 92)
1.81, p
.18
Corrugator EMG
( V)
1.6
0.7
F(1, 93)
5.10, p
.03
Orbicularis oculi EMG
( V)
0.5
0.7
F(1, 93) < 1
Zygomatic EMG
( V)
0.2
0.4
F(1, 93)
1.16, p
.28
Skin conductance
(log
S + 1)
0.04
0.05
F(1, 92)
1.53, p
.22
Initial heart rate deceleration (bpm)
−4.70
−4.64
F(1, 92) < 1
Peak heart rate acceleration (bpm)
1.47
1.22
F(1, 92) < 1
Note.
EMG
electromyographic; bpm
beats per minute.
sponses were most similar in both initial deceleration
the 18 stimulus contents in each set. Each participant
and peak acceleration to those obtained when viewing
viewed one of the two sets of 36 pictures. Seventy-
other erotic pictures (see Table 2). Startle reflexes
one participants (46 women, 25 men) viewed and
also did not differ and were as inhibited in magnitude
rated one set of pictures, and 64 participants (36
as those elicited when viewing the most arousing,
women, 28 men) viewed and rated the second set.
pleasant content (i.e., erotic couples) for both men and
Pictures were ordered in blocks of 18, such that there
women (see Table 3). For women, blinks elicited
was one exemplar from each of the 18 contents in
when viewing same-sex erotica were significantly
each block of 18 (as well as 4 pleasant, 1 neutral, and
smaller than those elicited when viewing opposite-sex
4 unpleasant within each half-block of 9) and with the
erotica, F(1, 43)
6.18, p
.017.
constraint that no more than 3 pictures of the same
hedonic valence were presented consecutively. Four
Experiment 2
different presentation orders were constructed, such
In Experiment 2, we sought to determine whether
that each picture was presented equally often, across
people select consistent emotional labels for their ex-
orders, in each quartile.
perience when viewing the different picture contents
Each participant received a 36-page booklet, in
and, in particular, whether men and women differ in
which each page began with the phrase, “When I
the labels they apply to their emotional experience.
looked at the picture, it made me feel:” and was fol-
The same 72 pictures were shown to a new sample of
lowed by a list of 24 emotions presented in four col-
135 participants, who were asked to circle as many
umns and six rows. Among the emotions listed were
emotions from a list of 24 different emotion words as
the following: angry, afraid, bored, sad, pity, anxious,
were evoked when looking at each picture. On the
irritated, disgusted, impatient, unhappy, confused,
basis of the sex differences found in Experiment 1, we
embarrassed, happy, hungry, loving, sexy, excited, ro-
were most interested in how men and women label
mantic, nurturant, satisfied, comfortable, free, play-
their emotional experience when viewing unpleasant
ful, and amused. Four different page orders were cre-
picture contents, as well as when viewing pictures of
ated in which the appearance of each word in each
erotic stimuli.
column was counterbalanced. Each booklet was ar-
ranged in subblocks of four pages, with one page from
Method
each of the four counterbalanced orders.
Participants.
There were 135 (82 women and 53
Procedure.
The procedure involved a 6-s picture
men) participants from a University of Florida intro-
presentation, followed by the ratings. Five seconds
ductory psychology class who received course credit
before each picture appeared, an instruction slide was
for taking part in this study. Sessions were conducted
presented that indicated the correct number in the rat-
in small groups varying from 6 to 15 participants.
ings booklet for the upcoming picture (e.g., “Please
Materials and design.
The materials were the
rate the next slide on page X.”). The participant was
same as described in Experiment 1. The 72 pictures
told to turn to the specified page. Then, a picture was
(18 categories with four exemplars each) were divided
presented for 6 s, and the participant was instructed to
into two sets of 36, with two exemplars from each of
simply view the picture while it was on the screen.
312
BRADLEY, CODISPOTI, SABATINELLI, AND LANG
After the picture was removed, participants were in-
tions were endorsed more often by women compared
structed to circle any of the listed emotions that they
with men, F(1, 129)
16.21, p < .001. Similarly,
may have felt while viewing the picture; 15 s was
whereas both men and women reported feeling happy
allocated for the rating. This procedure continued un-
and free when viewing nature pictures, women re-
til all the ratings were made, after which the partici-
ported feeling these more often than men, F(1, 129)
pants were debriefed, given credit, and thanked.
16.48, p < .001. For both men and women, feeling
For each participant, the proportion of pictures in
free was more frequently selected as the elicited emo-
each content category endorsed with a specific emo-
tion than feeling happy when viewing pictures of na-
tion descriptor was determined, and these proportions
ture, F(1, 129)
14.58, p < .001.
were averaged across participants for each picture
For unpleasant pictures, both men and women re-
content.
ported feeling afraid and angry when viewing pictures
Results
of human threat. However, a significant interaction
between gender and emotion, F(1, 129)
8.38, p
Table 4 lists the mean proportion of trials on which
.004, indicated that although men and women did not
men and women endorsed a specific emotion as de-
differ in their reports of feeling angry, women were
scriptive of their emotional response when viewing
significantly more likely to endorse feeling afraid
pictures in each of the 18 stimulus contents investi-
when viewing these threatening pictures, F(1, 129)
gated in this study. Emotions endorsed on less than
22.89, p < .001. Similarly, for pictures of animal
10% of the trials by participants are not listed. As
threat, both men and women reported feeling afraid
Table 4 illustrates, a variety of different emotion de-
and anxious, but a significant interaction of gender
scriptors were typically selected when viewing each
and emotion, F(1, 129)
4.22, p
.04, indicated
content. For analysis purposes, the two emotions se-
that women were more likely to report feelings of fear
lected most frequently were included in an ANOVA
compared with men, F(1, 129)
16.67, p < .001,
that included sex (men, women) and specific emotion
whereas men and women did not differ in their reports
(two levels) and are highlighted in Table 4.
of anxiety. Both men and women, however, were
In general, there was good agreement between men
more likely to report feeling fear than anxiety when
and women concerning which emotions were se-
viewing pictures of threatening animals, F(1, 77)
lected, although there were several notable differ-
40.73, p < .001, for women; F(1, 52)
10.64, p <
ences. When viewing pictures of erotic couples, both
.005, for men.
men and women reported feeling sexy and romantic
Pictures of mutilation elicited feelings of disgust in
and did not differ significantly in the proportion of
men and women equally often and secondarily, feel-
trials these feelings were reported. Conversely, men
and women diverged in their reports of emotional
ings of pity and sadness, which did not differ between
experience when viewing erotic pictures of the oppo-
men and women. Pictures of contamination and pol-
site sex, with a higher proportion of men endorsing
lution elicited reports of disgust and irritation from
sexy and excited, whereas women reported feeling
both men and women. Women tended to be more
amused and embarrassed, and with less agreement
likely to report these feelings when viewing pictures
among themselves. For adventure pictures, men and
of pollution, however, F(1,22)
4.12, p
.06. Pic-
women both reported feeling excited and free. Al-
tures of accidents, illness, and loss all elicited feelings
though these emotions were also selected to describe
primarily of pity and sadness from both men and
emotional reactions to sports pictures, men reported
women, but for pictures of illness, these emotions
them significantly more often than did women,
were reported significantly more often by women
F(1,129)
5.28, p
.02.
compared with men, F(1, 129)
6.98, p
.009.
Men and women did not differ in their reports of
There were no differences between men and
feeling primarily happy when viewing pictures of
women in reports of emotions when viewing neutral
food, with a secondary listing of excited for men and
pictures. Both reported feeling primarily bored and
satisfied for women, although each of these was re-
then confused (.56 and .22 for household objects; .34
ported somewhat less frequently than reports of hap-
and .28 for mushrooms), with higher reports of bore-
piness, F(1, 77)
16.17, p < .001, for women;
dom than confusion for household objects, F(1, 129)
F(1,52)
3.42, p
.07, for men. Men and women
49.43, but not for mushrooms. When viewing
agreed in selecting “happy” and “love” when viewing
same-sex erotica, men and women also reported feel-
pictures of families and babies, but both of these emo-
ing bored and confused, again, with more frequent
SEX DIFFERENCES IN EMOTION
313
reports of boredom than confusion, F(1, 129)
unpleasant pictures, compared with other picture con-
14.08, p < .001.
tents, only women showed the most heightened reac-
tivity for unpleasant pictures.
General Discussion
Modulation of the startle reflex—an index of de-
fensive mobilization—was generally similar for men
As expected, affective pictures elicited measurable
and women, with the largest reflexes elicited in the
emotional reactions in both men and women. The pat-
context of cues depicting threat and death. The pattern
tern of modulation in the autonomic, somatic, reflex,
of modulation differed, however, across specific con-
and evaluative systems measured here was quite simi-
tents: Women showed a general increase in blink
lar, with both men and women most reactive to con-
magnitude to unpleasant pictures and clear potentia-
tents depicting threat, mutilation, and erotica—cues
tion for the most highly arousing picture contents
that are held to strongly activate the defensive and
(relative to blinks elicited when viewing neutral pic-
appetitive motivational systems underlying emotional
tures), whereas men showed a relative inhibition for
experience. Nonetheless, the data are consistent with
unpleasant contents rated moderate in arousal (e.g.,
hypotheses that women are more defensively reactive,
pollution, loss, illness, accidents), with “heightened”
whereas men are specifically more aroused by erotic
reflexes only when viewing pictures depicting immi-
pictures. These sex differences influence the shape of
nent attack.
the affective space formed by men’s and women’s
Interpreted from the perspective of the defense cas-
judgments of pleasure and arousal: For women, un-
cade model (Lang, Bradley, & Cuthbert, 1997), these
pleasant pictures adhere closely to a steeper, more
data suggest that symbolic picture cues activate the
strongly linear, defense motivation vector than is
defensive motive system more intensely in women
found for men; conversely, for men, the vector for
than in men, as evidenced by greater attention (as
appetitive motivation is steeper and more strongly lin-
measured by sustained cardiac deceleration) and clear
ear than for women. These differences in the shape of
priming of the defensive startle reflex. Furthermore,
affective space for men and women are also found
these reflexive responses were accompanied by judg-
when people rate sounds or words (Bradley & Lang,
ments of greater displeasure and stronger negative
1999a,b). These evaluative judgments, of course, can
facial displays, indicating a consistent, markedly
be shaped through social learning and reinforcement.
negative reaction to aversive stimulation. For men,
The concordance of evaluative reports with auto-
however, evidence suggests that unpleasant pictures
nomic and reflex measures, however, suggests that
were less defensively activating. Increased orienting
differential cue reactivity in men and women may be,
(as indicated by moderate cardiac deceleration and
in part, biologically determined.
inhibited startle reflexes) was the predominant reac-
Defensive Motivation
tion to all but the most arousing unpleasant contents.
These different patterns of reactivity are broadly
Reports of more intense displeasure and greater fa-
similar to those obtained in a recent study investigat-
cial EMG activity were obtained for women irrespec-
ing children’s (7–10 years old) affective reactions to
tive of specific, “unpleasant” content, supporting hy-
pictures. In this research, girls responded more defen-
potheses that women are generally more emotionally
sively in evaluative judgments, facial expression, and
expressive when processing aversive cues. There
reflex potentiation than did boys (McManis, Bradley,
were also indications of heightened defensive reactiv-
Berg, Cuthbert, & Lang, 2001). Whether these differ-
ity for women in measures less amenable to voluntary
ences reflect biological or sociocultural factors cannot
control. Women responded with greater fear brady-
be determined here; however, finding similar sex dif-
cardia (cardiac deceleration) when viewing unpleas-
ferences early in development is compatible with a
ant pictures, again irrespective of the specific aversive
biologically based view (Manstead, 1992). In any
content. Fear bradycardia is interpreted as indicating
event, it is clear that women are highly concordant in
the heightened attention and sensory intake that oc-
responding to aversive cues—across evaluative, auto-
curs when the defensive system is alerted, but activa-
nomic, and reflex measures—and show this respon-
tion is relatively low. For women, this index of de-
siveness to a broad range of stimuli, consistent with a
fensive activation was greater than men even for
stronger disposition to engage the defensive motiva-
unpleasant contents rated as low in arousal. And al-
tional system.
though men and women were equivalent in absolute
That women may be more reactive to aversive
magnitude of electrodermal responses when viewing
stimulation is consistent with other social and clinical
314
BRADLEY, CODISPOTI, SABATINELLI, AND LANG
M
.56
.17
erotica
Same-sex
F
.14
.12
.56
.26
.13
.41
.12
.11
.15
.29
.43
.56
.17
Nature
.60
.22
.12
.29
.42
.60
.76
.26
.13
.58
.58
.20
.29
.11
.25
.32
.27
Families
.79
.78
.14
.11
.28
.48
.19
.23
.49
.49
.27
.17
.11
.12
.11
Food
.37
.15
.17
.12
.10
.49
.55
.18
.16
.52
.13
.37
.17
Sports
FMFMFMFM
.60
.69
.17
.17
.64
.22
.50
.14
M
.39
.61
.55
.25
.38
.22
Adventure
F
Endorsed
.49
.63
.13
.66
.39
.61
.12
.28
Were
M
.29
.11
.50
.44
.25
.15
.14
.11
.23
.27
.12
.21
Emotions
erotica
Opposite-sex
F
.20
.19
.21
.21
.14
.36
.17
.22
Specific
Which
.24
.34
.44
.36
.47
.11
.14
.17
.29
.21
on
Erotic
couples
FM
Trials
.13
.30
.37
.21
.41
.13
.23
.26
.12
of
label
4
Proportion
Happy
Loving
Sexy
Excited
Romantic
Satisfied
Comfortable
Free
Amused
Playful
Nurturant
Bored
Embarrassed
Confused
Irritated
Afraid
Anxious
Impatient
Table
Mean
Emotion
Pleasant
Unpleasant
SEX DIFFERENCES IN EMOTION
315
.21
.14
.16
.30
.13
.35
.37
.30
.16
Human
attack
.13
.25
.33
.19
.42
.67
.33
.19
.31
.15
.10
.42
.23
Animal
attack
FMFM
.11
.20
.69
.31
M
.42
.27
.20
.40
.15
.10
.75
Mutilation
F
.41
.33
.27
.47
.19
.36
.21
.81
participants.
male
M
.50
.12
.49
.12
.32
.16
.12
M
Accidents
F
.55
.27
.22
.63
.22
.53
.22
.19
participants;
M
.42
.16
.78
female
F
Contamination
F
bold.
.13
.50
.17
.11
.88
.14
in
.58
.13
.26
.51
.18
.12
.41
highlighted
are
Illness
.16
.67
.21
.31
.69
.37
.22
.10
.38
participants
.59
.15
.10
.61
.17
across
Loss
FMFM
.11
.56
.18
.11
.79
.12
.17
frequently
M
.10
.25
.24
.26
.21
.18
.11
.34
most
selected
)
Pollution
F
.13
.19
.22
.43
.24
.22
.15
.12
.56
.11
labels
two
(
Continued
4
The
Loving
Excited
Amused
Nurturant
Bored
Pity
Confused
Irritated
Sad
Angry
Afraid
Anxious
Digust
Impatient
Table
Pleasant
Unpleasant
Note.
316
BRADLEY, CODISPOTI, SABATINELLI, AND LANG
research. As Marks (1969) noted, “women consis-
Again, the relative influence of biological and so-
tently report more fear than do men on fear schedules”
ciocultural factors on men’s and women’s reactions to
(p. 78), and considering epidemiological studies of
erotic stimuli cannot be determined here, particularly
anxiety disorder, “there are hardly any phobias which
in terms of men’s heightened arousal to visually pre-
men manifest consistently more than women . . .” (p.
sented erotica. From a social-learning view, it is pos-
75). Although certainly open to a cultural explanation,
sible to entertain hypotheses that emotional expres-
the ubiquity of greater fearfulness in women suggests
sion in the context of sexual stimuli is accepted and
intrinsic factors, perhaps related to differences in hor-
perhaps reinforced for men, but not women, in West-
monal balance, physical size, and strength. Whether
ern culture. However, the idea that these young men
ultimately attributed to biology or culture, or both, the
are in a culture that encourages them to seek erotic
data presented here are consistent with the notion that
stimulation, for example, through magazines, is in-
picture stimuli prompt somewhat greater defensive
consistent with the predominantly Judeo-Christian
activation in women than men, whereas for men, un-
heritage in Western societies. That men may be spe-
pleasant pictures, particularly those representing low
cifically more reactive to visual depictions of erotic
personal threat, mainly prompt interest and attention.
stimuli is supported by the number of commercial
magazines dedicated to this enterprise, the relative
Appetitive Motivation
absence of these materials for women, and research
Reactions to scenes representing life’s joys and
indicating that women consistently report less arousal
pleasures were remarkably similar for men and
in this erotic context.
women, with erotic scenes prompting the most intense
Sociobiological theorists (e.g., Buss, 1994) hypoth-
reactions in all measures for both sexes. Sexual cues
esize that physical (visible) and nonphysical features
represent primary reinforcers that directly engage the
of potential mates are differently appealing to men
appetitive system. These cues also have consistent
and women because of differences in the relative costs
impact across individuals, unlike the idiosyncratic
entailed by sexual activity. For men, who are required
pleasures inherent in, for example, different sports or
to neither carry nor necessarily care for potential off-
recreational activities. Only pictures of erotica ap-
spring, physical features and relative youth are impor-
peared to reach the threshold necessary for additional
tant criteria in determining sexual action. These are
sympathetic activation over the obligatory orienting
attributes that are easily apprehended with visual
response (see Bradley et al., 2001). Consistent with
cues. For women, however, on whom a significant
previous research (Koukounas & McCabe, 1997;
portion of the burden of childbearing and childrearing
Steinman, 1981), men showed these effects more
falls, nonphysical features such as resourcefulness
strongly than women. Men reported more intense
and commitment may be more important in control-
pleasure and arousal when viewing erotic stimuli and
ling potentially critical sexual activities. Data indicat-
reacted with greater sympathetic arousal (as measured
ing the primacy of physical and nonphysical charac-
by palmar conductance) as well. Furthermore, height-
teristics in men’s and women’s choices of mates
ened arousal was evident only for erotica. Men did not
support this conceptualization (Bereczkei, Voros, Gal,
react more intensely than women to any other arous-
& Bernath, 1997; Stewart, Stinnett, & Rosenfeld,
ing appetitive content.
2000; Surbey & Conohan, 2000; Townsend & Levy,
The pattern of startle modulation when viewing
1990). Regardless of the ultimate explanation, the
pleasant pictures was remarkably similar for the two
data obtained here suggest that although erotic stimu-
sexes, with the most inhibited reflexes elicited in the
lation through visual cues activates appetitive moti-
context of erotic picture viewing, consistent with high
vation in both men and women, this activation is more
appetitive motivation. Responses in other measures,
intense for men.
however, suggested that women responded to erotic
Facial Expressivity
stimuli with mixed motivation. Like reactions to un-
pleasant pictures, erotic pictures evoked measurable
The nonspecific hypothesis that women are gener-
corrugator EMG changes, lower ratings of domi-
ally more facially expressive than men (e.g., Buck,
nance, and increased cardiac deceleration in women,
1984; Schwartz et al., 1980; Zuckerman & Larrance,
compared with men. Moreover, whereas men reported
1979) can be refined on the basis of the data acquired
feeling sexy and excited when viewing erotic pictures
here. Unpleasant pictures elicited measurable corru-
of the opposite sex, women reported feeling primarily
gator EMG activity in men (e.g., relative to viewing
amused and embarrassed.
neutral pictures), suggestive of a frown, and the pat-
SEX DIFFERENCES IN EMOTION
317
tern of EMG changes across different unpleasant con-
expressing fear. As noted above, men and women also
texts was not significantly different for men and
apply very different descriptors to their emotional ex-
women, with, for instance, the greatest activity occur-
perience when viewing erotic members of the oppo-
ring in response to pictures depicting mutilation and
site sex. As with other evaluative judgments, how-
death. Conversely, whereas women reacted with mea-
ever, the selection of appropriate words to describe
surable increases in activity over the zygomatic and
one’s emotional experience is highly influenced by
orbicularis oculi muscles when viewing pleasant pic-
social expectations and norms: Although both men
tures, indicative of smiling, men did not. Taken to-
and women were highly reactive to pictures of mem-
gether, these data indicate that when viewing unpleas-
bers of the same sex in erotic poses, both reported
ant materials, women frown more than men.
feeling primarily bored when viewing these pictures.
Conversely, when viewing pleasant stimuli, men
The aversive content most reliably labeled as evok-
smile less than women.
ing a specific emotion involved pictures of contami-
Although there is no theoretical consensus explain-
nation (e.g., spoiled food, vomit) and mutilated bod-
ing this phenomenon, several investigators have sug-
ies, which on a sizable number of trials were reported
gested that differences in emotional expression are
as evoking disgust. These pictures were also rated as
related to social expectations based on different gen-
relatively high in arousal and prompted sustained car-
der roles. LaFrance and Hecht (2000), for example,
diac deceleration, large electrodermal and reflex re-
proposed that the norms governing the use of facial
actions, and a facial expression consistent with one of
displays, particularly smiling, are different for men
disgust (e.g., concurrent increases in activity over cor-
and women, with greater social expectations for
rugator, zygomatic, and orbicularis oculi muscles).
women to smile, as well as more reinforcement for
Notably, these are some of the first data indicating
smiling behavior. These socially defined rules are
that pictures that elicit reports of disgust (i.e., rather
proposed to arise based on gender roles in which, for
than facial expressions of disgust) are potent in elic-
instance, women are more frequently involved in nur-
iting affective reactions in both men and women.
turing and caregiving activities where smiling is an
effective instrumental behavior. This hypothesis is es-
Summary
pecially interesting in view of the fact that pictures of
families and babies elicited the most smiling behavior
Both men and women react to pictures depicting
in the college women assessed here.
highly arousing contents of threat, mutilation, and
Conversely, there is evidence that expressions of
death, as expected if basic motivational dispositions
displeasure, particularly fear, are discouraged in boys
are similarly organized for mammals at survival risk
and men (Brody, 2000), based on their expected roles
in the natural environment. For pictures depicting less
in positions of power and strength, in which displays
arousing appetitive and defensive contexts, men and
of weakness (e.g., threat) are dysfunctional. Taken
women also showed remarkable congruence in their
together, these social proscriptions would encourage
physiological profiles. Nonetheless, the pattern of re-
women to smile and discourage men from frowning.
sponse across evaluative reports, facial EMG activity,
Socialization notwithstanding, however, the current
cardiac orienting, and reflex modulation when women
data indicate that unpleasant pictures prompted mea-
viewed unpleasant pictures, irrespective of specific
surable facial reactions in men, evidencing perhaps
content, suggests a broad disposition to respond with
society’s ability to attenuate, but not eliminate, reflex-
greater defensive activation for these aversive cues.
ive facial displays to potentially threatening cues.
Conversely, men were particularly reactive to visual
pictures involving erotic stimuli.
Discrete Emotions
Biological and sociocultural factors may both play
A relatively wide range of emotions was typically
roles in modulating the reactions of men and women
endorsed when viewing specific picture contents, with
in affective perception. Cross-cultural studies, espe-
pictures of attack, for instance, prompting reports of
cially those investigating men and women raised in
fear, anxiety, sadness, anger, and pity. In general, la-
nontraditional Western societies, as well as additional
beling of discrete emotions was fairly consistent
developmental research, would assist in determining
across men and women. Replicating previous research
the impact of biological–genetic or sociocultural vari-
(Nolan & Ryan, 2000), women were more likely to
ables in affecting men’s and women’s reactions to
report fear when viewing pictures depicting threat,
affective cues. Moreover, the fact that female rats also
consistent with a view that men are discouraged from
consistently show greater defensive reactivity in
318
BRADLEY, CODISPOTI, SABATINELLI, AND LANG
threatening contexts than do male rats (Blanchard,
ment, and differentiation. In A. H. Fischer (Ed.), Gender
Shepherd, de Padua Carobrez, & Blanchard, 1991;
and emotion: Social psychology perspectives (pp. 24–47).
Crabtree & Moyer, 1973; Shepherd, Rodgers, Blan-
New York: Cambridge University Press.
chard, & Magee, 1993) suggests that animal research
Buck, R. (1984). The communication of emotion. New
may also help to elucidate the causal contributors to
York: Guilford Press.
differential emotional reactivity. The data acquired
Buss, D. M. (1994). The evolution of desire: Strategies of
here, however, also demonstrate that despite specific
human mating. New York: Basic Books.
sex differences, men and women are quite similar in
Center for the Study of Emotion and Attention (1999). The
their affective reactions to a variety of pictures de-
International Affective Picture System: Digitized photo-
picting life’s pleasant and unpleasant events.
graphs. Gainesville, FL: University of Florida, The Cen-
ter for Research in Psychophysiology.
References
Cook, E. W., III. (1997). VPM reference manual. Birming-
ham, AL: Author.
Alexander, M. G., & Wood, W. (2000). Women, men, and
Crabtree, J. M., & Moyer, K. E. (1973). Sex differences in
positive emotions: A social role interpretation. In A. H.
fighting and defense induced in rats by shock and d-
Fischer (Ed.), Gender and emotion: Social psychology
amphetamine during morphine abstinence. Physiology
perspectives (pp. 189–210). New York: Cambridge Uni-
and Behavior, 11, 337–343.
versity Press.
Cuthbert, B. N., Bradley, M. M., & Lang, P. J. (1996). Prob-
Allen, J. G., & Haccoun, D. M. (1976). Sex differences in
ing picture perception: Activation and emotion. Psycho-
emotionality: A multidimensional approach. Human Re-
physiology, 33, 103–111.
lations, 29, 711–722.
Fischer, A. H., & Manstead, A. S. R. (2000). The relation
Bereczkei, T., Voros, S., Gal, A., & Bernath, L. (1997).
between gender and emotions in different cultures. In
Resources, attractiveness, family commitment: Repro-
A. H. Fischer (Ed.), Gender and emotion: Social psychol-
ductive decisions in human mate choice. Ethology, 103,
ogy perspectives (pp. 71–94). New York: Cambridge
681–699.
University Press.
Blanchard, D. C., Shepherd, J. K., de Padua Carobrez, A., &
Globisch, J., Hamm, A., Schneider, R., & Vaitl, D. (1993).
Blanchard, R. J. (1991). Sex effects in defensive behav-
A computer program for scoring reflex eyeblink and elec-
ior: Baseline differences and drug interactions. Neurosci-
trodermal responses written in Pascal. Psychophysiology,
ence & Biobehavioral Reviews, 15, 461–468.
39, S30.
Bradley, M. M., & Lang, P. J. (1994). Measuring emotion:
Gordon, M. T., & Riger, S. (1991). The female fear: The
The self-assessment manikin and the semantic differen-
social costs of rape. Chicago: University of Illinois Press.
tial. Journal of Behavioral Therapy and Experimental
Hodes, R. L., Cook, E. W., & Lang, P. J. (1985). Individual
Psychiatry, 25, 49–59.
differences in autonomic response: Conditioned associa-
Bradley, M. M., & Lang, P. J. (1999a). Affective norms for
tion or conditioned fear? Psychophysiology, 22, 545–560.
English words (ANEW): Instruction manual and affective
Koukounas, E., & McCabe, M. (1997). Sexual and emo-
ratings (Tech. Rep. No. C-1). Gainesville, FL: University
tional variables influencing sexual response to erotica.
of Florida, The Center for Research in Psychophysiology.
Behaviour Research and Therapy, 35, 221–230.
Bradley, M. M., & Lang, P. J. (1999b). International affec-
Kring, A. M., & Gordon, A. H. (1998). Sex differences in
tive digitized sounds (IAD): Instruction manual and af-
emotion: Expression, experience, and physiology. Jour-
fective ratings (Tech. Rep. No. B-1). Gainesville, FL:
nal of Personality and Social Psychology, 74, 686–703.
University of Florida, The Center for Research in Psy-
LaFrance, M., & Banaji, M. (1992). Towards a reconsid-
chophysiology.
eration of the gender emotion relationship. In M. S. Clard
Bradley, M. M., Codispoti, M., Cuthbert, B. N., & Lang,
(Ed.), Emotions and social behavior: Review of person-
P. J. (2001). Emotion and motivation I: Defensive and
ality and social psychology (Vol. 14, pp. 178–201). Bev-
appetitive reactions in picture processing. Emotion, 1,
erly Hills, CA: Sage.
276–299.
LaFrance, M., & Hecht, M. (2000). Gender and smiling: A
Brody, L. R. (1996). Gender, emotional expression, and par-
meta-analysis. In A. H. Fischer (Ed.), Gender and emo-
ent–child boundaries. In R. D. Kavanaugh, B. Zimmer-
tion: Social psychology perspectives (pp. 118–142). New
berg, & S. Fein (Eds.), Emotion: Interdisciplinary per-
York: Cambridge University Press.
spectives (pp. 139–170). Mahwah, NJ: Erlbaum.
Lang, P. J. (1980). Behavioral treatment and bio-behavioral
Brody, L. R. (2000). The socialization of gender differences
assessment: Computer applications. In J. B. Sidowski,
in emotional expression: Display rules, infant tempera-
J. H. Johnson, & T. A. Williams (Eds.), Technology in
SEX DIFFERENCES IN EMOTION
319
mental health care delivery systems (pp. 119–137). Nor-
Sachs-Ericsson, N., & Ciarlo, J. A. (2000). Gender, social
wood, NJ: Ablex.
roles, and mental health: An epidemiological perspective.
Lang, P. J., Bradley, M. M., & Cuthbert, M. M. (1997). Mo-
Sex Roles, 43, 605–628.
tivated attention: Affect, activation, and action. In P. J.
Schwartz, G. E., Brown, S. L., & Ahern, G. L. (1980). Fa-
Lang, R. F. Simons, & M. T. Balaban (Eds.), Attention
cial muscle patterning and subjective experience during
and orienting: Sensory and motivational processes (pp.
affective imagery: Sex differences. Psychophysiology,
97–135). Hillsdale, NJ: Erlbaum.
17, 75–82.
Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1999). In-
Shepherd, J. K., Rodgers, R. J., Blanchard, R. J., & Magee,
ternational Affective Picture System (IAPS): Instruction
L. K. (1993). Ondansetron, gender and antipredator de-
manual and affective ratings (Tech. Rep. No. A-4).
fensive behavior. Journal of Psychopharmacology, 7,
Gainesville, FL: University of Florida, The Center for
72–81.
Research in Psychophysiology.
Steinman, D. L. (1981). A comparison of male and female
Lang, P. J., Greenwald, M. K., Bradley, M. M., & Hamm,
patterns of sexual arousal. Archives of Sexual Behavior,
A. O. (1993). Looking at pictures: Affective, facial, vis-
10, 477–492.
ceral, and behavioral reactions. Psychophysiology, 30,
Stewart, S., Stinnett, H., & Rosenfeld, L. B. (2000). Sex
261–273.
differences in desired characteristics of short-term and
Manstead, A. S. R. (1992). Gender differences in emotion.
long-term relationship partners. Journal of Social and
In M. A. Gale & M. W. Eysenck (Eds.), Handbook of
Personal Relationships, 17, 843–853.
individual differences: Biological perspectives (pp. 355–
Surbey, M. K., & Conohan, C. D. (2000). Willingness to
387). Chichester, England: Wiley.
engage in casual sex: The role of parental qualities and
Marks, I. M. (1969). Fears and phobias. New York: Aca-
perceived risk of aggression. Human Nature, 11, 367–
demic Press.
386.
McManis, M. H., Bradley, M. M., Berg, W. K., Cuthbert,
Townsend, J. M., & Levy, G. D. (1990). Effects of potential
B. N., & Lang, P. J. (2001). Emotional reactivity in chil-
partners’ physical attractiveness and socioeconomic sta-
dren: Verbal, physiological, and behavioral responses to
tus on sexuality and partner selection. Archives of Sexual
affective pictures. Psychophysiology, 38, 222–231.
Behavior, 19, 149–164.
Murnen, S. K., & Stockton, M. (1997). Gender and self-
Watson, D., Clark, L. A., & Mineka, S. (1994). Tempera-
reported sexual arousal in response to sexual stimulation:
ment, personality, and the mood and anxiety disorders.
A meta-analytic review. Sex Roles, 37, 135–153.
Journal of Abnormal Psychology, 103, 103–116.
Nolan, J. M., & Ryan, G. W. (2000). Fear and loathing at
Zuckerman, M., & Larrance, D. T. (1979). Individual dif-
the Cineplex: Gender differences in descriptions and per-
ferences in perceived encoding and decoding abilities. In
ceptions of slasher films. Sex Roles, 42, 39–56.
R. Rosenthal (Ed.), Skill in nonverbal communication:
Nolen-Hoeksema, S. (1987). Sex differences in unipolar
Individual differences (pp. 171–203). Cambridge, MA:
depression: Evidence and theory. Psychological Bulletin,
Oelgeschlager, Gunn & Hain.
101, 259–282.
O’Gorman, J. G. (1983). Individual differences in the ori-
enting response. In D. Siddle (Ed.), Orienting and ha-
Received August 17, 2000
bituation: Perspectives in human research. New York:
Revision received May 18, 2001
Wiley.
Accepted August 1, 2001 s
