Content uploaded by David L Vogel
Author content
All content in this area was uploaded by David L Vogel
Content may be subject to copyright.
VOGEL ET AL.OVERESTIMATING GENDER DIFFERENCES
GENDER DIFFERENCES IN EMOTIONAL
EXPRESSION: DO MENTAL HEALTH TRAINEES
OVERESTIMATE THE MAGNITUDE?
DAVID L. VOGEL
Iowa State University
STEPHEN R. WESTER
University of Wisconsin – Milwaukee
MARTIN HEESACKER
University of Florida
GUY A. BOYSEN AND JOANN SEEMAN
Iowa State University
Three studies examined whether mental health trainees overestimate emotional dif-
ferences between the sexes. In Studies 1 and 2, two samples of mental health train-
ees, reporting about dating women and men (Study 1), and female and male clients
(Study 2), were found to overestimate sex differences in emotional expressiveness.
Mental health trainees reported sex differences in women’s and men’s willingness to
express specific emotions and in their comfort and perceptions of risk involved with
talking about emotions that were greater than the differences found in women’s and
men’s self–reports. In fact, in contrast to the mental health trainees’ reports, Study 3
revealed no significant sex differences in the observed emotional expression of
women and men or female and male clients. Across studies, mental health trainees
were found to overestimate sex differences 50 to 67% of the time.
Research on the emotional expressions, defined as the verbal communi
-
cation of one’s feelings to another person, of women and men illustrate
that, while women and men seem to exhibit some situational differences
in emotional expression (Balswick, 1982; Brooks, 1990; Doyle, 1989;
Fujita, Diener, & Sandvik, 1991; Greenwald, Cook, & Lang, 1989; Lang,
Greenwald, Bradley, & Hamm, 1993), the societal stereotype of gen
-
der–based emotional abilities often reflects an overestimation of these
305
Journal of Social and Clinical Psychology, Vol. 25, No. 3, 2006, pp. 305-332
Please address correspondence to: David L. Vogel, W149 Lagomarcino Hall, Depart
-
mentofPsychology, Iowa State University, Ames, IA 50011. E-mail: dvogel@iastate.edu.
emotional differences (Barett, Robin, Pietromonoca, & Eysell, 1998).
Furthermore, women and men are believed to not only differ in their ex
-
ternal expression of emotion (i.e., emotion regulation) but also in their
internal expression of emotions. In other words, their comfort with and
awareness of their emotions (i.e., experience of emotions) is perceived as
different. Indeed, the stereotypical image of a “reactive female” and
“stoic male” is so implicit in the European American definitions of femi
-
ninity and masculinity (Robinson & Clore, 2002; Shields, 1995) that re
-
searchers have warned of the potential dangers associated with
overendorsing these stereotypical beliefs (Deaux, 1976; Heesacker et al.,
1999; Shields, 1995; Wester, Vogel, Pressly, & Heesacker, 2002), includ
-
ing reinforcing the status quo (Stabb, Cox, & Harber, 1997) and
perpetuating power differences in male–female relationships
(Knudson–Martin, 2003).
EMOTIONAL STEREOTYPING IN THERAPY
Over–reliance on an exaggerated emotional stereotype is of particular con-
cern in psychotherapy because “emotion and emotional related processes
lie at the heart of” therapy (Heesacker & Bradley, 1997, p. 201). Fischer et al.
(1993), for example, recommended that clinicians “might be well–advised
toavoidstereotypicaltreatmentof...affectiveissues”(p.442), as the influ-
ence of overendorsing such stereotypes in psychotherapy may be destruc-
tive for clients who do not conform to the perceived norms for their gender.
Similarly, Shields (1995) suggested that reliance on gender stereotypes of
emotion can lead to therapists unintentionally reinforcing gender-consis-
tent emotional expression and thus limiting, rather than expanding, the
range of emotional behavior available to their clients. Giving some cre
-
dence to these concerns, the overreliance on gender stereotypes to under
-
stand client affective issues has also been connected to the misdiagnosis of
clients (Garb, 1997), the overdiagnosing of clients (Robertson & Fitzgerald,
1990), and to choosing treatment modalities that fit gender–stereotypical
needs rather than actual client needs (Wester et al., 2002).
Unfortunately, research exploring the degree to which mental health
professionals actually overestimate the differences in the emotional ex
-
pression of women and men is limited. Indeed, while a number of stud
-
ies have shown that undergraduate samples overestimate the emotional
differences between women and men (see Plant, Hyde, Keltner, &
Devine, 2000), we found only one empirical study that directly assessed
the beliefs of mental health professionals (Heesacker et al., 1999).
Heesacker and colleagues (1999) demonstrated that volunteer mental
health workers, mental health trainees, and practicing clinicians all en
-
dorsed a view of men as hypo–emotional, relative to women. They also
306 VOGEL ET AL.
showed that experience had some effect with practicing clinicians show
-
ing the least endorsement and mental health trainees and volunteers
showing greater endorsement of the gender stereotypical statements.
This single study, while intriguing, leaves several unanswered ques
-
tions regarding the degree to which individuals providing mental
health services overestimate the differences in the emotional expression
of women and men. To begin, there is a need to assess trainees’ beliefs
about the emotions of both women and men, separately. Heesacker et al.
(1999) used a measure that directly compared the emotional expression of
men to women. This made sense, as they were exploring the issue from the
perspective of men’s hypo–emotionality. However, building on this re
-
quires assessing mental health trainees’ independent beliefs regarding the
emotional expression of women and men. Doing so should help to rule out
potential biases based on the demand characteristics of having to compare
men to women. Second, given that Heesacker et al. (1999) evaluated train
-
ees’ general perceptions of the emotional expression of women and men,
there is also a need to examine how mental health trainees view clients.
Mental health trainees may believe that women and men differ in ways that
fulfill traditional gender roles, but abandon those views when thinking
about a client.
Furthermore, the Heesacker et al. (1999) study used a single global
measure to examine the emotional stereotype. Studies have demon-
strated, gender stereotypes are less pronounced when undergraduates
rated discrete emotional traits rather than when they made global cate-
gorizations of emotional expression (Robinson & Clore, 2002), Thus, an-
other useful perspective would involve the examination of trainees’ be-
liefs regarding discrete emotions (e.g., sadness, happiness, anger). For
example, it may be that mental health trainees see women as only more
expressive of sadness, but not anger. Similarly, there is a need to repli
-
cate the previous findings across discrete emotional competencies in ad
-
dition to focusing on the global emotional distinction targeted by
Heesacker et al. (1999). The rationale for this is that the global emotional
stereotype may lead to specific beliefs about women’s and men’s emo
-
tional expression, such as differences in women’s and men’s comfort in
dealing with emotions or perceived risks/benefits associated with dis
-
playing emotion. Examination of trainees’ perceptions of women and
men across such specific dimensions may provide a clearer understand
-
ing of the specific aspects of the gender stereotype that are or are not
overestimated by mental health trainees.
Finally, the Heesacker et al. (1999) study documented that mental
health trainees endorsed beliefs consistent with the gender stereotype of
men as being less emotional than women. However, they did not show
whether the beliefs endorsed were actual overestimations of emotional
OVERESTIMATING GENDER DIFFERENCES 307
differences between the sexes. Heesacker and colleagues argued that the
items they developed were specifically designed, based on the extant lit
-
erature, to be inaccurate statements. However, while the literature is not
consistent, several studies do support the presence of certain situational
sex differences in emotional expression (Balswick, 1982; Brooks, 1990;
Doyle, 1989; Fujita et al., 1991; Greenwald et al., 1989; Lang et al., 1993). As
such, perhaps a better way to assess the degree to which trainees’ beliefs
are overestimations (i.e., accurate), would be to empirically compare a
sample of trainees’ beliefs about the emotional expression of women and
men to the self–reported emotional expression of women and men.
ASSESSING THE ACCURACY OF A STEREOTYPE
In the social psychology literature the most frequently used method to
assess the accuracy of a belief has been to determine a group’s (i.e.,
women and men) actual characteristics by asking a sample of the group
members to rate themselves as individuals on the stereotype–relevant
attributes of interest and compare these responses to the those reported
by participants (e.g., mental health professionals; see Diekamn, Eagly, &
Kulesa, 2002; Judd, Ryan, & Park, 1991; Madon et al., 1998; McCauley &
Stitt, 1978; Nisbett & Kunda, 1985; Ryan, 1996; Ryan & Bogart, 2001). In
following this paradigm, therefore, mental health trainees could be
asked to rate women and men on emotionality, women and men could
rate themselves on emotionality, and these ratings could then be com-
pared. Mental health trainees who endorse sex differences larger than
womenand men self–report are quite likely to be exhibiting a perception
that is an overestimation of reality. Alternatively, if trainees’ evaluations
mirror the actual emotional expression of women and men, then they are
likely to be exhibiting an accurate perception.
Self–reports are the most common and arguably the best way to mea
-
sure emotion, as the measures are generally reliable and valid (Clore,
1994; Diener, 2000; Watson, 2000). However, self–reports of emotion can
be subject to potential biases (see Robinson & Clore, 2002; Robinson, John
-
son, & Shields, 1998). For example, the two groups being compared
(self–report vs. other report) may not be using the same internal scale to
judge the behavior or trait. Ryan (2002) also noted that stereotype studies
based on self–reports of a group’s characteristics are subject to biases
based on (a) sample selection, (b) memory and knowledge gaps, and (c)
social desirability concerns. However, several steps can be taken to mini
-
mize the potential impact of these biases. Problems with the use of differ
-
ing internal scales can be reduced by having the groups make judgments
using a specific comparison group (i.e., clients). Sample selection prob
-
lems can be addressed by using multiple independent samples. Asking
308 VOGEL ET AL.
about specific behaviors across multiple measurements can reduce prob
-
lems associated with memory and knowledge gaps. Further, social desir
-
ability can be dealt with by including both positive (i.e., expression of hap
-
piness) and negative characteristics (i.e., expression of jealousy) and
seeing if the effects differ for women and men reporting about socially de
-
sirable versus socially undesirable characteristics. Finally, perhaps the
best way to understand if potential self–report biases are present in a
study is to include behavioral data. Inclusion of the degree to which the
characteristics occurred in a realistic setting would show whether these
potential biases in the self–reporting were occurring.
THE CURRENT STUDIES
Three studies were conducted to directly address these issues. Study 1 as
-
sessed the beliefs of mental health trainees concerning the willingness of
women and men to express specific emotions (i.e., Depression, Happi-
ness, Jealousy, Anxiety, Anger, and Fear) as well as their comfort in talk-
ing about emotional issues with a dating partner. Study 2 extended the
findings of Study 1 by assessing the perceptions of mental health trainees
regarding the willingness of female and male clients to express specific
emotions (i.e., Depression, Happiness, Jealousy, Anxiety, Anger, and
Fear) to a therapist. Study 2, also for the first time, examined whether
mental health trainees would endorse the part of the stereotype in which
female clients are viewed as seeing emotional expression to a therapist as
less risky and more beneficial than male clients, who in turn are viewed as
seeing emotional expression to a therapist as more risky and less benefi-
cial. In addition, we examined in both Study 1 and Study 2 whether the be
-
liefs held by the mental health trainees were actually overestimations of
the differences in the emotional expressions of women and men by com
-
paring their beliefs to women’s and men’s (i.e., females and males who
were dating in Study 1 and female and male clients in Study 2) self–re
-
ports of their own behavior. Finally, to further rule out potential con
-
founds of the results due to biases in the self–reports of the women and
men, in Study 3, we observed the actual emotional expressions of women
and men during intimate conversations with a dating partner and
between clients and therapists during therapy sessions.
STUDY 1
In Study 1, mental health trainees were asked to report the degree to
which they believe women and men express specific emotions to a dating
partner. Mental health trainees were also asked to assess how comfortable
women and men feel when talking about emotional topics with a dating
OVERESTIMATING GENDER DIFFERENCES 309
partner. Study 1 then examined whether the beliefs endorsed by the men
-
tal health trainees were overestimations of the differences in the emo
-
tional expression of women and men by comparing their beliefs to dating
individuals’ self–reported willingness and comfort to express emotions.
Dating individuals were chosen as the context for emotional expression in
Study 1 for four reasons. First, we wanted to ensure that all participants
were easily able to imagine a situation in which others would be express
-
ing emotions. Because romantic relationships are so common and the ex
-
pression of emotion is a natural part of a relationship, the context of a
dating relationship serves as an ideal situation for assessing emotional ex
-
pression. Second, by having the mental health trainees rate the emotional
expression of a specific group (i.e., women and men in a dating relation
-
ship) it increased the likelihood that they would think of a similar person
to the comparison group when doing the rating task. Third, relationship
issues are among the common topics of focus in therapy with a college
population (Benton, Roberston, Tseng, Newton, & Benton, 2003) and
therefore, the mental health trainees were likely to have knowledge about
this group. Finally, researchers have suggested that the stereotype re-
garding women’s and men’s emotional expression within a relationship
is especially pronounced in this society, as women are seen as more
relationship focused and men as more autonomous.
To examine the degree to which mental health trainees endorse sex dif-
ferences in the emotional expression of women and men, we gathered
three independent samples. Sample 1 was a group of mental health train-
ees. Samples 2 and 3 were comparison groups of women and men report-
ing on their own emotional behavior with a dating partner. We hypothe-
sized, based on the Heesacker et al. (1999) study, that mental health
trainees would endorse differences between women and men across six
discrete emotions (i.e., Depression, Happiness, Jealousy, Anxiety, Anger,
andFear) and that these reported differences would reflect an overestima
-
tion of the true differences between women and men (i.e., the mental
health trainees would report women as more emotionally expressive than
men and more dissimilar than the women and men self–report). We also
hypothesized that mental health trainees would report sex differences in
the difficulty experienced by women and men talking about emotional
topics (i.e., feelings toward partner) and that they would do so to a greater
degree than self–reported by women and men in dating relationships.
METHOD
Participants
Sample 1. We recruited mental health trainees (N = 110; 88 women, &
22 men) from a Counseling and Counselor Education program at a Mid
-
310 VOGEL ET AL.
western university in order to assess their beliefs regarding women’s
and men’s emotional expression. Participants were between the ages of
22 and 59 (M = 30.3, SD = 9.3). Participants self–reported as European
American (80%), African American (16%), Asian American (1%), His
-
panic (2%), and one person did not respond (1%). Clinical experience in
the sample varied from none to 11 years (M = 14.2 months, SD = 27.6).
Sample 2. We recruited this sample of collegians (N = 209; 143 women,
66 men) from a Midwestern university in order to assess the degree to
which they expressed specific emotions to their partners (i.e., Depression,
Happiness, Jealousy, Anxiety, Anger, and Fear). All participants were
currently involved in heterosexual dating relationships. Twenty–one per
-
cent of the participants were freshman, 36% were sophomores, 23% were
juniors, 19% were seniors, and 1% reported other. Participants self–re
-
ported as European American (89%), African American (3%), Asian
American (3%), Hispanic (2%), and six reported as other (3%).
Sample 3. We recruited another sample of collegians (N = 290; 180
women,110 men) from a Midwestern university in order to report on the
degree to which they were comfortable talking to their partners about
emotional topics (i.e., intimacy, feelings about their partners). Again all
participants were currently involved in heterosexual dating relation-
ships. Participants were between the ages of 17 and 47 (M = 21.6, SD =
4.4). Although, racial data were not collected from participants, the ra-
cial composition of the collegiate population at the time the study was
conducted was 68.8% Caucasian, 8.9% Hispanic, 8.9% International,
6.0% African American, 5.6% Asian American, and 1.8% Native
American.
Measures
Emotional Expression. Beliefs in the degree to which women and men
express specific emotions to their partners were measured by the
40–item Emotional Self–Disclosure Scale (ESDS, Snell, Miller, & Belk,
1988). The ESDS consists of eight subscales, each reflective of a different
emotion (i.e., Depression, Happiness, Jealousy, Anxiety, Anger, Calm
-
ness, Apathy, and Fear). Each subscale contains five separate items.
Items are scored from 0 (not at all) to 4 (totally willing). The items are
then summed so that higher scores correspond to greater emotional dis
-
closure for that emotion. For the present study, participants completed
six of the eight subscales; calmness and apathy were not used due their
nonsignificant findings in previous studies (Snell et al., 1998). The men
-
tal health trainees (Sample 1) were asked how willing they thought a
woman would be to express these emotions to her partner. They also
were asked how willing they thought a man would be to express these
emotions to his partner. The women and men in Sample 2 answered the
OVERESTIMATING GENDER DIFFERENCES 311
questions regarding how willing they would be to self–disclose these
emotions to their partners.
The ESDS is designed to detect how much a person discloses to differ
-
ent individuals (Snell, Miller, Belk, Garcia–Falconi, & Hernan
-
dez–Sanchez, 1989). For example, Snell et al. (1988; 1989) found that the
disclosure recipient substantially affected emotional self–disclosure,
with female recipients being entrusted with the most self–disclosure.
Vogel & Wester (2003) also found each of the subscales of the ESDS to be
correlated (r = .20–.34) with one’s general tendency to self–disclose (Dis
-
tress Disclosure Index; Kahn & Hessling, 2001). In the current study, the
mental health trainees had internal consistencies that were generally
good in responding to questions about women [i.e., Depression (.80),
Happiness (.89), Jealousy (.78), Anxiety (.76), Anger, (.80), and Fear (.88)]
and in responding to questions about men [i.e., Depression (.72), Happi
-
ness (.90), Jealousy (.64), Anxiety (.77), Anger (.88), and Fear (.87)]. The
women and men in Sample 2 also had internal consistencies that were
excellent across all six emotions (.93–.95). Finally, the ESDS has been re-
ported to be somewhat “stable across time” (Snell et al., 1988, p. 64).
Test–retest reliabilities over a 12–week period showed that responses
were somewhat stable (.58–.75; Snell et al., 1988).
Difficulty Talking about Emotional Issues. Beliefs in how difficult it
would be for women and men to talk about emotional issues with their
partners were measured by the Difficulty in Relationship Issues Ques-
tionnaire (DRIQ, Vogel, Wester, & Heesacker, 1999; Vogel, Wester,
Heesacker, & Madon, 2003). The DRIQ has 17–items that assess the per-
ceived difficulty of discussing different topics with a partner. Partici-
pants indicate their responses to these items on a nine–point Likert–type
scale with anchors of 1 (very emotionally easy to talk about) to 9 (very
emotionally difficult to talk about). For the current study only the eight
emotional topics were selected. Example items include “Your needs in
the relationship” and “Your feelings about your partner.” We adminis
-
tered the eight emotional DRIQ items to the mental health trainees (Sam
-
ple 1) who filled out responses for how difficult they thought it would be
for a woman and a man to talk about each item with their partner and to
the women and men in Sample 3, who filled out responses for how diffi
-
cult it would be for them to talk about each item with their partners.
The issues making up the DRIQ were originally determined by asking
a group of undergraduate and graduate students to list the common is
-
sues they have faced in relationships. To assess whether the eight emo
-
tional items were all measuring one construct, we conducted a principal
component factor analysis using the college students in Sample 3. This
analysis revealed one factor (eigenvalue = 4.4, 55% of the variance) con
-
taining seven of the eight items. Each retained item had a factor loading
312 VOGEL ET AL.
of .7 or greater. The one dropped item had a factor loading of below .4.
Additional support for construct validity of the DRIQ can be found in
previous studies comparing participants’ self–reports to independent
observers’perceptionsof the level of comfort of theparticipants during a
discussion of DRIQ issues with their partners (Vogel et al., 1999). Fur
-
ther, the seven emotional items from the DRIQ correlate (–.34) with the
degree of emotional self–disclosures an individual tends to express
(Vogel, 2004). In the current study, the internal consistency for the seven
retained items in the mental health training sample was .88 (women)
and .83 (men). The internal consistency for the seven retained items for
the women and men in Sample 3 was .89.
Procedure
Mental Health Trainees. Mental health trainees were recruited from
graduate classes and offered extra credit for their participation. They
were informed that participation was voluntary and anonymous and
would consist of questions involving their beliefs about the emotionality
of men and women who were dating. After completingan informed con-
sent, the ESDS and DRIQ were distributed with participants being ran-
domly assigned to rate either women’s or men’s emotionality first. The
survey also contained some demographic items (i.e., sex, age, race/eth-
nicity, and clinical experience). After finishing the questionnaires, all
participants were debriefed and then dismissed.
Dating Women and Men. Participants in all samples were recruited from
undergraduate classes and received extra credit for their participation.
They were informed that participation was voluntary and anonymous.
They were told that the procedure would involve questions about their be
-
liefs to which they discuss specific feelings and emotions within an intimate
dating relationship. After completing an informed consent, participants re
-
ceived a packet either containing the ESDS or the DRIQ (i.e., Sample 2 re
-
ceived the ESDS; Sample 3 received the DRIQ) and some demographic
questions (e.g., sex, age, and relationship status). After finishing the ques
-
tionnaire, all participants were debriefed and then dismissed.
RESULTS
Beliefs Regarding Expression of Emotion
We hypothesized that mental health trainees would endorse differences
between women and men across six discrete emotions (i.e., Depression,
Happiness, Jealousy, Anxiety, Anger, and Fear) and that these reported
differences would reflect an overestimation of the true differences be
-
tween women and men. To examine this, the ESDS scores of the mental
health trainees (Sample 1) were, first, compared to one another by sex.
OVERESTIMATING GENDER DIFFERENCES 313
Bonferroni-corrected (a=.05/6 = .008) paired t–tests revealed that men
-
tal health trainees believed that women express more emotion than men
(p’s < .003) on all six measured emotions (see Table 1). Next, we exam-
ined the differences in women’s and men’s self–reported emotional ex-
pression (Sample 2). Bonferroni-corrected (a = .05/6 = .008) independent
t–tests comparing women’s and men’s self–reported emotional expres-
sion revealed no significant differences (see Table 2; p = .03 for fear and
greater than .08 for the other five emotions).
These results suggested that mental health trainees endorse beliefs
about women’s and men’s expression of emotion that are not consistent
with dating individuals’ own perceptions. To directly examine whether
themental health trainees beliefs represent overestimations of the differ-
ences in the emotional expression of dating women and men, we directly
compared the degree of difference that mental health trainees believed
existed between women’s and men’s emotional expression to the degree
of difference that women and men in dating relationships actually re
-
ported. To obtain the difference scores for each emotion we subtracted
the six ESDS scores for women from the six ESDS scores for men as re
-
ported by the mental health trainees. In addition, dating women’s ESDS
scores were subtracted from dating men’s ESDS scores to establish dif
-
ference scores for the dating individuals. The difference scores from the
mental health trainees were then compared to the difference scores from
the dating women and men. Bonferroni-corrected (a=.05/6 = .008) inde
-
pendent t–tests revealed that mental health trainees believed that
women expressed more emotion than men than the dating individuals
reported on four of the six emotions measured (Depression, Happiness,
Anxiety, and Fear; see Table 3). Jealousy did not reach Bonferroni-cor
-
rectedsignificance (p < .03) and anger did not reach significance (p >.80).
314 VOGEL ET AL.
TABLE 1. Paired
t
–Test Results for the Beliefs of Mental Health Trainees Regarding
Women’s and Men’s Emotional Expression
Women Men
Emotion
MSD MSD t df p
Depression 18.5 3.3 13.1 3.3 –13.3 109 .001
Happiness 21.0 3.8 18.3 4.7 –6.37 109 .001
Jealousy 15.6 3.8 13.3 3.7 –5.38 109 .001
Anxiety 18.4 3.2 12.8 3.6 –13.0 109 .001
Anger 18.5 3.7 17.0 4.7 –3.11 109 .002
Fear 19.9 3.4 10.3 4.0 –20.1 109 .001
Beliefs Regarding Degree of Difficulty Talking about Emotion
The next goal of Study 1 was to assess if mental health trainees had
overendorsed the idea that men have more difficulty talking about emo-
tions than women. We hypothesized that mental health trainees would
report that women have less difficulty talking about emotional topics
(e.g., needs, intimacy, time together) than men and that they would
overestimate this difference by reporting a greater difference than
women and men self–report. The mental health trainees filled out the
DRIQ regarding the degree to which women and the degree to which
men would have difficulty talking about emotional topics with their in-
timate partners. A paired t–test of the beliefs of mental health trainees
about women (M = 25.6, SD = 11.3) versus men (M = 39.2, SD = 10.6) was
significant, t(103) = –9.8, p < .001, supporting the notion that mental
health trainees endorsed this dimension of the stereotype. Next, we
were interested in the difference between women’s and men’s reports
(Sample 3) of their own difficulty talking about emotional topics with a
partner. An independent t–test comparing the self–reports of women’s
(M = 22.1, SD = 10.9) and men’s (M = 21.4, SD = 10.5) difficulty revealed
no significant differences between the groups, t(276) = –.55, p > .58.
These analyses suggested that mental health trainees might be en
-
dorsing beliefs that reflect an overestimation of women’s and men’s
emotional comfort. To directly assess the accuracy of this we compared
the difference between how much difficulty mental health trainees be
-
lieved women and men have talking about emotional topics to what
women and men actually report. Difference scores were computed by
subtracting the DRIQ score of women from the DRIQ score of men as
reported by the mental health trainees. The same was done for the
women’s and men’s self–reported scores on the DRIQ in Sample 3. The
OVERESTIMATING GENDER DIFFERENCES 315
TABLE 2. Independent
t
–Test Results for Women’s and Men’s
Self–Reported Emotional Expression
Women Men
Emotion
MSD MSD t df p
Depression 21.2 4.3 20.5 5.1 –1.0 206 .30
Happiness 23.7 2.7 23.0 3.5 –1.6 206 .10
Jealousy 19.6 5.1 19.0 5.2 –0.7 206 .46
Anxiety 21.6 4.0 20.6 4.7 –1.5 206 .12
Anger 21.1 4.3 19.9 5.3 –1.7 205 .09
Fear 21.7 4.1 20.1 5.3 –2.4 206 .02
two difference scores were then compared to one another. The inde-
pendent t–test revealed that mental health trainees (M = –13.6, SD =
14.1) believed that women, compared to men, would have significantly
less difficulty talking about emotional topics than the dating couples
reported (M = –.03, SD = 14.3), t(206) = –6.9, p < .001. As such, the
hypothesis was supported.
DISCUSSION
Overall, the mental health trainees in the current study endorsed the
ideas that (a) women are more comfortable and more willing to talk
about emotions, whereas (b) men are less comfortable and less disposed
to talking about emotions. These findings are consistent with previous
work (e.g., Heesacker et al., 1999) that found mental health trainees en
-
dorsing stereotypical beliefs about men’s emotional expression. The re
-
sults of Study 1 also showed that mental health trainees overendorsed
statements saying that women and men are differentially willing to ex
-
press four specific emotions (i.e., Depression, Happiness, Anxiety, and
Fear) and are differently comfortable talking about emotional topics
(i.e., needs, intimacy, time together). Interestingly, however, the mental
health trainees only overendorsed certain emotions (i.e., they did not
overendorse anger or jealousy). These differential findings are consis
-
tent with the literature, which suggests that some sex differences may be
present in emotional expression (Balswick, 1982; Brooks, 1990; Doyle,
1989; Fujita et al., 1991; Greenwald et al., 1989; Lang et al., 1993) but that
the societal stereotype of the emotional abilities of women and men of
-
ten reflect an overestimation of these emotional differences (Barrett et
al., 1998). This is an important finding, as previous studies have not
316 VOGEL ET AL.
TABLE 3. Independent
t
–Test Results for the Comparison of the Reports of Mental Health
Trainees and Dating Individuals Regarding the Degree of Difference in Women’s and
Men’s Emotional Expression
Mental Health Trainees Dating Individuals
Emotion
MSDMSDtdfp
Depression 5.4 4.3 0.5 6.9 6.0 178 .001
Happiness 2.6 4.3 0.8 4.3 2.7 178 .007
Jealousy 2.3 4.4 0.4 7.1 2.2 178 .02
Anxiety 5.5 4.5 0.7 6.5 5.9 178 .001
Anger 1.6 5.2 1.4 7.0 0.2 177 .82
Fear 9.6 5.0 1.4 7.1 9.1 178 .001
shown the degree to which the beliefs endorsed by mental health
trainees were actually overestimations of the emotional differences
between the sexes.
Several strengths of Study 1 should be noted. First, similar results
were found using two independent samples as comparison groups. Sec
-
ond, two independent measures were used with items that asked about
specific behaviors (i.e., talking about a specific topic with a partner) and
that included both positive and negative characteristics (i.e., happiness
and sadness). Third, the mental health trainees made judgments about a
specific group (i.e., dating individuals), which should have increased
the likelihood that they were thinking of a person similar to the compari
-
son group when doing the rating task. Although consistent across multi
-
ple measures and samples, one limitation of these results is that they still
are only based on self–report data. A second limitation is that the mea
-
sures used assessed the behaviors of women and men in dating relation
-
ships and did not specifically focus on client behaviors and mental
health trainees’ beliefs about female and male clients. These limitations
were addressed in Studies 2 and 3.
STUDY 2
It is possible that while mental health trainees may overendorse certain
differences in the expression of emotion for women and men, in general,
they put aside these beliefs when they are thinking about clients’ emo-
tional expression. To explore this, Study 2 extended the findings of
Study 1 and assessed how mental health trainees perceived female and
male clients’ emotional expression. As before, these perceptions were
compared to actual female and male clients’ perceptions of their own
emotional expression. Additionally, Study 2 measured the beliefs of
mental health trainees concerning the risks and benefits that female and
male clients experience expressing emotions to a therapist, and assessed
the degree to which these perceptions reflected female and male clients’
perceptions of their own experiences of the risk and benefits of
emotional expression to a therapist.
We hypothesized, as in Study 1, that mental health trainees would en
-
dorse differences between female and male clients across the six discrete
emotions (i.e., Depression, Happiness, Jealousy, Anxiety, Anger, and
Fear) and that these reported differences would reflect an overestima
-
tion of the true differences between female and male clients. Similarly,
we hypothesized that mental health trainees would endorse sex differ
-
ences in the level of risk and benefits experienced by clients and that they
would do so to a greater degree thanfemale and male clients self–report.
OVERESTIMATING GENDER DIFFERENCES 317
METHOD
Participants
Sample 1. We recruited mental health trainees (N = 56; 45 women & 11
men) from a Counseling and Counselor Education program at a Mid
-
western university in order to assess their beliefs regarding female and
male client’s emotional expressions. Participants self–reported as Euro
-
pean American (79%), African American (11%), Asian American (4%),
Hispanic (2%), Native American (2%), and one person did not respond
(2%). The participants’ clinical experience varied from none to five years
(M = 2.78 months, SD = 9.77).
Sample 2. We recruited this sample of clients who were currently or
had been engaged in therapy (N = 71; 51 women, & 20 men) from a Mid
-
western university. Sample 2 was recruited in order to report on how
comfortable they were talking to a therapist about specific emotions
(e.g., depression and anxiety). Participants self–reported as European
American (89%), African American (2%), Asian American (2%), His-
panic (2%), Native American (2%), and eight chose either not to respond
or responded as other (5%).
Sample 3. We recruited a sample of clients who were currently or had
been engaged in therapy (N = 56; 38 women, 18 men) from a Midwestern
university. Sample 3 was recruited in order to assess how risky and ben-
eficial it was to talk to a therapist about emotional topics. Participants
self–reported as European American (91%), African American (4%), and
Asian American (5%).
Measures
Emotional Expression. As in Study 1, the 40–item Emotional Self–Dis
-
closure Scale (ESDS, Snell et al., 1988) was used as an indication of how
often women and men express six specific emotions (i.e., Depression,
Happiness, Jealousy, Anxiety, Anger, and Fear). For Study 2, mental
health trainees answered the questions concerning how willing they
thought a female and a male client would be to express these emotions to
a therapist. The female and male clients in Sample 2 answered the ques
-
tions regarding how willing they would be to express these emotions to
a therapist. The mental health trainees had internal consistencies that
were generally good for the responses about female clients [i.e., Depres
-
sion (.63), Happiness (.86), Jealousy (.86), Anxiety (.75), Anger (.83), and
Fear (.78)] and for their responses about male clients [i.e., Depression
(.54), Happiness (.84), Jealousy (.70), Anxiety (.72), Anger; (.81), and Fear
(.83)]. The clients in Sample 2 had adequate internal consistencies across
all six emotions (.87–.92).
Anticipated Risks and Benefits of Disclosing Emotions. The anticipated
risks and the anticipated benefits of self–disclosing emotional material
318 VOGEL ET AL.
to a therapist were both measured by the Disclosure Expectations Scale
(DES; Vogel & Wester, 2003). The DES is an eight–item scale consisting
of four items asking how risky it is to self–disclose emotional informa
-
tion to a therapist and four items asking how beneficial it is to self–dis
-
close emotional information to a therapist. Responses are rated on a
Likert–type scale from 1 (strongly disagree) to 5 (strongly agree) and are
summed for each subscale such that higher scores reflect greater antici
-
pated risks and anticipated benefits. A sample item for anticipated risk
is “How risky would it feel to disclose your hidden feelings to a thera
-
pist?” A sample item for anticipated benefits is “Would you feel better if
you disclosed feelings of sadness or anxiety to a therapist?” For Study 2,
mental health trainees answered the questions for how they thought a
female and a male client would anticipate self–disclosing emotional in
-
formation to a therapist. The female and male clients in Sample 3 an
-
swered the questions regarding their own anticipations about
self–disclosing emotional information to a therapist.
The two subscales have been confirmed in a factor analysis and have
been found to correlate with measures of self–disclosure and self–con-
cealment as well as social support and psychological distress (Vogel &
Wester, 2003). The internal consistency was previously found to be .74
for anticipated risk and .83 for anticipated benefits (Vogel & Wester,
2003). For Study 2 the internal consistency for the mental health trainee
was good for responses about women (.83, Risks; .93, Benefits) as well as
for responses about men (.86, Risks; .85, Benefits). The clients in Sample 3
also had acceptable internal consistency: 71 (Risks) and .78 (Benefits).
Procedures
Mental Health Trainees. Mental health trainees were recruited from
graduate classes and offered extra credit for their participation. They
were informed that participation was voluntary and anonymous and
participation would consist of questions involving their beliefs concern
-
ing how much clients discuss specific feelings and emotions with a ther
-
apist. After completing an informed consent, the ESDS and DES were
distributed with participants being randomly assigned to rate either
women’s or men’s emotionality first. The survey also contained some
demographic items (i.e., sex, age, race/ethnicity, and clinical experi
-
ence). After finishing the questionnaires, all participants were debriefed
and then dismissed.
Clients. Clients in Samples 2 and 3 were recruited from undergradu
-
ate classes and received extra credit for their participation. They were in
-
formed that participation was voluntary and anonymous. They were
told that the procedure would involve questions about how much they
discuss specific feelings and emotions with a therapist. After completing
OVERESTIMATING GENDER DIFFERENCES 319
an informed consent, participants received a packet either containing
the ESDS or the DES (i.e., Sample 2 received the ESDS; Sample 3 received
the DES) and some demographic questions (e.g., sex, age, and therapy
status). After finishing the questionnaire, all participants were debriefed
and then dismissed.
RESULTS
Beliefs Regarding Expression of Emotion
We hypothesized that mental health trainees would endorse differences
between female and male clients across the six discrete emotions (i.e.,
Depression, Happiness, Jealousy, Anxiety, Anger, and Fear) and that
these reported differences would reflect an overestimation of the true
differences between female and male clients. Bonferroni-corrected (a =
.05/6 = .008) paired t–tests revealed that the mental health trainees be
-
lieved that female clients express more emotion than male clients (ps <
.001) on four of the six emotions measured (see Table 4). Jealousy (p =
.01) and anger (p = .07) did not reach the Bonferroni-corrected signifi
-
cance levels. Next, we examined the degree to which female and male
clients (Sample 2) differed in their self–reports of their own emotional
-
ity. Bonferroni-corrected (a = .05/6 = .008) independent t–tests compar
-
ing women’s and men’s self–reported emotional expression revealed no
significant differences between the two groups (see Table 5; all ps >.06).
These analyses are consistent with the notion that the mental health
trainees were overestimating the differences in female and male clients’
expression of certain emotions. To directly examine the accuracy of this,
difference scores were again calculated by subtracting the ESDS scores
reported by mental health trainees for female targets from the ESDS
scores reported for male targets. Likewise, female clients’ ESDS scores
320 VOGEL ET AL.
TABLE 4. Paired
t
–Test Results for the Beliefs of Mental Health Trainees about Female
and Male Clients’ Emotional Expression
Women Men
Emotion
MSD MSD t df p
Depression 21.0 2.3 13.9 2.8 –16.0 55 .001
Happiness 22.2 2.5 18.6 3.4 –7.54 55 .001
Jealousy 17.7 4.2 16.1 3.4 –2.68 55 .01
Anxiety 21.0 2.7 14.2 3.5 –12.7 55 .001
Anger 18.3 3.9 19.6 3.8 1.85 55 .07
Fear 21.1 2.7 11.2 3.4 –18.4 55 .001
were subtracted from male clients’ ESDS scores in Sample 2. The differ
-
ence scores for the mental health trainees and the clients were then com-
pared for each of the six emotions measured by the ESDS. Thus, we di-
rectly compared the degree of difference that mental health trainees
placed on female and male clients’ emotions to how much difference cli-
ents actually reported. Bonferroni-corrected (a = .05/6 = .008) independ-
entt–tests revealed that the differences between female and male clients’
emotional expressions as reported by the mental health trainees were
greater than the differences reported by actual clients (p < .001) on three
of the six emotions measured (see Table 6). Jealousy, Happiness, and
Anger did not reach significance (p > .05). Thus, mental health trainees
appeared only to overendorse beliefs regarding three of the emotions
measured (i.e., 50% accurate or inaccurate). Of further interest, however,
is that the differences in the perceptions of the mental health trainees re
-
garding women’s and men’s expression of these three emotions (i.e.,
anxiety, fear, and depression) were between one and two points larger
than was found in Study 1 (see Tables 3 and 6), suggesting that for these
specific emotions mental health trainees may believe that female and
male clients are particularly different.
Beliefs Regarding Risk and Benefits of Talking about Emotion
We hypothesized that mental health trainees would endorse sex differ
-
ences in the level of risk and benefits experienced by clients and that they
would do so to a greater degree than female and male clients self–report.
Bonferroni-corrected (a = .05/2 = .025) paired t–tests of the beliefs of
mental health trainees of women (M = 12.1, SD = 3.4) versus men’s (M =
15.8, SD = 2.5) showed that anticipated risks in emotionally disclosing to
a therapist achieved statistical significance, t(54) = –6.8, p < .001, but an
-
ticipated benefits did not, t(54) = –.59, p > .55). In turn, Bonferroni-cor
-
OVERESTIMATING GENDER DIFFERENCES 321
TABLE 5. Independent
t
–Test Results for Female and Male Clients’
Reported Emotional Expression
Women Men
Emotion
MSD MSD t df p
Depression 20.9 4.4 20.5 5.2 –0.3 69 .74
Happiness 23.8 2.1 22.2 5.2 –1.9 69 .07
Jealousy 18.7 5.6 19.3 5.9 0.4 69 .67
Anxiety 21.0 4.2 20.5 4.8 –0.4 69 .67
Anger 20.8 4.4 19.8 5.5 –0.8 69 .41
Fear 21.0 4.4 20.4 5.5 –0.5 69 .61
rected (a = .05/2 = .025) independent t–tests comparing female to male
clients’ (Sample 3) reported risks and benefits revealed no significant
differences in anticipated risks, t(54) = –2.1, p > .04, or anticipated bene-
fits, t(54) = –1.3, p > .20.
These results support the notion that mental health trainees might be
over–endorsing sex differences in the level of risk and not over–endors-
ing the level of benefits associated with emotional expression. As before,
we wanted to assess the accuracy of this by comparing the beliefs of the
mental health trainees regarding the differences in female and male cli-
ents to the actual reported differences of the female and male clients. To
do so we again computed difference scores between what mental health
trainees reported on the DES regarding women and what they reported
on the DES regarding men. We also computed difference scores between
the DES scores of the female and male clients in Sample 3. We then com
-
pared the difference score from the mental health trainees to the differ
-
ence score of female and male clients. Bonferroni-corrected (a = .05/2 =
.025) independent t–test revealed as significant effect for anticipated
risks, t(71) = –3.5, p < .001, and anticipated benefits, t(71) = –2.8, p < .006.
Consistent with predictions, the mental health trainees believed that fe
-
male clients compared to male clients (M = –3.7, SD = 4.0) would feel less
risk associated with expressing an emotion to a therapist than the clients
themselves did (M = 1.7, SD = 9.1). In fact, while the mental health train
-
ees thought female clients would experience fewer risks, the male clients
actually reported less risks (M = 19.1, SD = 8.3) than the female clients
did (M = 23.2, SD = 6.2), which is consistent with other research
(Kushner & Sher, 1989). Contrary to predictions, however, the mental
health trainees marginally believed that men (M = –.22, SD = 2.7) would
anticipate greater benefits with expressing an emotion to a therapist
322 VOGEL ET AL.
TABLE 6. Independent
t
–test Results for the Comparison of the Reports of Mental Health
Trainees and Clients Regarding the Degree of Difference in Women’s and Men’s
Emotional Expression
Mental Health Trainees Dating Individuals
Emotion
MSDMSDtdfp
Depression 7.1 3.3 0.2 5.8 6.5 74 .001
Happiness 3.5 3.5 1.6 5.9 1.8 74 .08
Jealousy 1.5 4.2 –1.2 8.0 1.9 74 .06
Anxiety 6.4 3.7 –0.3 6.5 5.5 74 .001
Anger –1.2 4.9 0.8 7.6 –1.3 74 .19
Fear 9.8 4.0 0.1 7.5 7.3 74 .001
than women, while the clients reported that women would anticipate
greater benefits (M = 4.4, SD = 11.5).
DISCUSSION
In Study 2, the results showed that these mental health trainees believed
that female and male clients differed in their expressions of sadness,
anxiety, and fear as well as how risky it is to express emotion to a thera
-
pist. These beliefs in the differences between female and male clients
were not consistent with clients’ own reports of their behavior suggest
-
ing that they were reflecting an overestimation of these emotional differ
-
ences. However, the mental health trainees did not think that female and
male clients’ emotional expressions differed for certain other emotions
(i.e., happiness, jealousy, or anger) or that there was any greater benefit
to female than to male clients for talking about emotions with a thera
-
pist. These results, therefore, provide both some promising and discon-
certing news. On a positive note, the mental health trainees appeared to
employ fewer overestimated beliefs regarding female and male client’s
emotions than the mental health trainees did regarding the dating indi-
viduals in Study 1 (i.e., only three instead of four of the six emotions
measured and only one of thetwodiscreteemotional competencieswere
overendorsed). Of concern, though, is that the emotional differences
they did perceive between female and male clients were more pro-
nounced (i.e., larger than was found in Study 1) and, in one case, in the
opposite direction as the clients themselves reported (i.e., perceptions of
risk). As such, these results underscore the importance of not relying on
gender–based stereotypes about clients’ emotional expression as 50% of
the time mental health trainees were overestimating sex differences in
emotional expression and competency.
Several strengths of Study 2 should be noted. First, similar results
were found using two independent samples as comparison groups. Sec
-
ond, two independent measures were used with items that asked about
specific behaviors (i.e., talking about expressing distress to a therapist)
and that included both positive and negative characteristics (i.e., happi
-
ness and sadness). Third, the mental health trainees made judgments
about a specific group (i.e., female and male clients), which should have
increased the likelihood that they were thinking of a person similar to
the comparison group when doing the rating task. Of course, using
self–reports as comparison data can be problematic if there are biases in
their self–reporting. Therefore, to further assess the viability of this con
-
cern, Study 3 examined the actual observed emotional behavior of
women and men during intimate discussions with their partners and
with female and male clients during actual therapy sessions.
OVERESTIMATING GENDER DIFFERENCES 323
STUDY 3
Two new samples were collected for Study 3. First, we gathered a new
sample of dating partners asked talk to each other about an emotional is
-
sue (i.e., feelings toward each other) while being videotaped. Second, we
gathered a new sample of female and male clients videotaped during
their actual therapy sessions. The videotapes of the dating discussions
and therapy sessions allowed us to directly observe women’s and men’s
emotional behavior in a realistic setting. Thus, in Study 3 we directly ex
-
amined whether women and men differed in the number and type of
emotions expressed to their partners or to a therapist. In addition, the dat
-
ing women and men provided self–reports of their current emotional ex
-
periences right before and after their discussions. Specifically, we
hypothesized that observations of participants in these samples will sup
-
port the previous findings of sex similarities in Study 1 and Study 2, and
accordingly there will be no significant differences between the twosexes.
METHOD
Participants
Sample 1. We collected this sample of heterosexual dating couples (N
= 39) from an archived set of couple interactions (see Vogel, Wester,
Heesacker, & Madon, 2003) in order to examine the degree to which
women and men felt and exhibited specific emotions toward their part-
ners while discussing an emotional topic with one another. Female par-
ticipants ranged in age from 17 years to 25 years (M = 19.6, SD = 1.77).
Male participants ranged in age from 17 years to 27 years (M = 20.3, SD =
2.17). The average length of the dating relationship was 18.64 months
(SD = 22.32). Participants self–reported as European American (96%),
African American (2%), Asian American (1%), and one participant
reported as other (1%).
Sample 2. We collected this sample of clients (N = 29; 18 female clients,
11 male clients) from an archived set of therapy sessions (see Day &
Schneider, 2002) in order to assess the degree to which clients actually
express different emotions to a therapist. Clients ranged in age from 19
to 75 (M = 40.6, SD = 15.6). Problems most frequently discussed during
these sessions concerned body image/weight, family, relationships,
self–esteem, and work/school. The clients self–reported as European
American (82%), African American (10%), Asian American (4%), and
other (4%).
Instruments
Emotion Coding. Emotional expressions during the couple and client
videotapes were identified using the “affective exploration” category
324 VOGEL ET AL.
from the Client Behavior System (CBS; Hill et al., 1992; Hill & O’Brien,
1999). The CBS consists of eight mutually exclusive categories. How
-
ever, due to the selective focus of the study on emotion only the “affec
-
tive exploration” category was used. Affective exploration is defined by
Hill & O’Brien as statements that indicate that the individual is “cur
-
rently involved and exploring feelings about significant material; spe
-
cific feeling words must be stated (e.g., happy, sad, anxious), or clearly
visible nonverbal behavior (e.g., audible sighs, clenched fists, lowering
of the head, crying, or shifting body position) must accompany affective
material; the [individual’s] . . . voice must sound as if feelings are being
experienced in the present moment; discussion of past feeling would be
coded as recounting unless the [individual] ...isre–experiencing the
feelings in the present moment” (p. 385). After identifying the emotion,
the raters then recorded the type of emotional expression (i.e., sadness,
happiness, anger).
Raters for the client tapes were two undergraduate research assistants
who were initially trained for eight weeks (70 hours) using this system
with an unrelated set of videotaped clients. Training continued until
they reached at least 90% agreement with the trainer (the fifth author)
and with each other. The raters continued to meet each week with the
trainer to compare their ratings and help prevent rater drift. The sessions
were randomly checked for reliability and ranged from 80 to 100%
agreement across the 23 weeks of coding of the client sample.
While the CBS was originally designed to identify client emotional ex-
pressions, we adapted it for use with the couple population in order to
keep the two samples comparable. This was accomplished by having
two new undergraduate raters receive four weeks of training, practicing
with an unrelated set of videotaped dating couples (36 hours) until they
reached at least 90% percent agreement with the trainer (the fifth author)
and with each other. The raters continued to achieve acceptable reliabil
-
ity (78%) with each other across the three weeks of coding the dating
couples’ videotapes.
Emotional Experiences. Emotional experiences from the dating cou
-
ples’ discussions were also gathered with the self–reported Post–Dis
-
cussion Questionnaire (PDQ; Heavey, Layne, & Christensen, 1993). The
PDQ was developed to assess a person’s emotional experiences during a
discussion with their partner. Partners indicate on a nine–point
Likert–type scale ranging from 1 (not all) to 9 (very much) the extent to
which they experienced specific emotions (e.g., sadness, joy, and anger)
while talking to their partners about an issue in their relationships. For
Study 3, the dating couples (Sample 1) filled out the items right after
talking to their current partners about the emotional topics. This allowed
an examination of their immediate perceptions of their emotional expe
-
OVERESTIMATING GENDER DIFFERENCES 325
riences. For the current study, we used the six emotions that most closely
corresponded to the emotions measured in Study 1 and Study 2 (i.e.,
sadness, joy, jealousy, worry, anger, and fear).
Difficulty Talking about Emotional Issues. As in Study 1, dating individ
-
uals’ beliefs regarding how difficult it would be for them to talk about an
emotional issue with their partners was measured by the DRIQ (Vogel et
al., 1999). For Study 3, the emotional DRIQ items were administered to
the couples in Sample 1 right before they were asked to talk about one of
the topics with their partners. As such, they filled out the items not
knowing which topic they would have to talk about. This made their
self–reports based on their actual feelings at that moment about possibly
having to talk about that topic.
Procedure
The couples in Sample 1 were originally asked to fill out the DRIQ and
some demographic questions. They were then assigned to talk to each
other about one of the emotional issues from the DRIQ for eight minutes.
After the discussion, the partners were separated and asked to fill out
thePDQ. Participants in Sample 1 received extra credit for their and their
partner’s participation. The clients in Sample 2 were originally gathered
from newspaper advertisements and posted flyers. Participants in Sam-
ple 2 received free therapy for their participation.
RESULTS
Behavioral Observations of the Emotional Expression
of Women and Men
In Study 3, we wanted to assess the degree to which women and men ex
-
press emotions in both dating relationships and in actual therapy ses
-
sions. This was done to examine whether self–report biases constituted a
viable alternative to understanding the significant differences between
mentalhealth trainees and other participants reported in Studies 1 and 2.
We predicted that there would be no differences in the observations of
women’s and men’s emotional behaviors. In the dating and therapy
samples, the emotional expressions of women and men were compared
across the total number of emotions expressed as well as individually for
five of the specific emotions examined in the previous studies (e.g., De
-
pression, Happiness, Anxiety, Anger, and Fear). Jealousy was not exam
-
ined in either the dating or the client sample, because only one
individual, across both samples, expressed it.
For the dating couples, six paired t–tests compared the female part
-
ners’ emotional expressions to the male partners’ emotional expres
-
sions. The tests were not significant for total number of emotional ex
-
326 VOGEL ET AL.
pressions (p > .3) or for any of the five specific emotions (ps > .05).
Similarly, the six independent t–tests of the female and male clients’
emotional expressions were not significant for total number of emo
-
tional expressions (p > .5) or the five specific emotions (ps > .2). Consis
-
tent with the hypotheses, these findings suggest that the responses of the
dating individuals in Study 1 reflect an accurate perception of their
emotional expression.
Self–Reports of the Emotional Experiences of Women and Men
Immediately before their discussion, the couples in Sample 1 filled out
the DRIQ concerning how emotionally difficult they felt the upcoming
discussion would be for them. Immediately afterward they filled out the
PDQ regarding the specific feelings they experienced during the discus
-
sion. Similar to the observational findings, paired t–tests revealed that
these dating women and men did not differ in their emotional experi
-
ences of sadness, joy, jealousy, worry, anger, or fear (ps > .19). In addi
-
tion, similar to Study 1’s findings, paired t–tests between dating
partners’ level of comfort talking about the emotional topics revealed
that these dating women (M = 3.0, SD = 2.5) and men (M = 3.0, SD = 2.3)
did not differ in their reported comfort, t(36) = .21, p > .83. These findings
suggest that the responses of the dating individuals in Study 1 reflect an
accurate perception of their emotional competencies.
DISCUSSION
The results of Study 3 confirm our belief that self–report biases did not
contribute to the results of Studies 1 and 2, and that responses of the dat
-
ing individuals in the previous studies were likely accurate perceptions
of their emotional expressions and competencies. Both the self–report
data and the behavioral data of the couples, along with the behavioral
data from actual clients, revealed no differences in the emotions ex
-
pressed by women and men. As a result, the data from this study builds
on that from the prior two studies in supporting the conclusion that
mental health trainees were likely overendorsing certain sex differences
in emotion.
There are some limitations associated with using archival samples as
done in Study 3. However, the consistency of the findings across behav
-
ioral and self–report data, across different samples and times, and con
-
sistency across multiple measures all serve to strengthen the conclusions
and to overshadow the limitations of archival data. Furthermore, the
nonanalogue nature of the archival data lends generalizability to the re
-
sults, which constitutes a significant strength.
OVERESTIMATING GENDER DIFFERENCES 327
GENERAL DISCUSSION
The findings of the current studies are consistent with the notion that
people often overestimate the magnitude of the differences in the emo
-
tional expression of women and men (Barrett et al., 1998). Importantly,
in these studies the people responsible for this overestimation were
mental health trainees responsible for working with the emotional is
-
sues of female and male clients. This is an important finding because the
previous research has not shown the degree to which the beliefs en
-
dorsed by mental health trainees were actually overestimations of the
emotional differences between the sexes. In the current studies the men
-
tal health trainees overendorsed sex difference 50 to 67% of the time. As
such, it is important for mental health trainees and those teaching them
to be aware of these data because they are in a position to influence the
emotion–related behavior of clients (Shields, 1995). Specifically, over
-
emphasizing sex differences in emotional behavior (e.g., instrumental
vs. expressive) may perpetuate limited notions associated with tradi-
tional masculinity and femininity (Shields, 1995), while de–emphasiz-
ing sex differences may facilitate more effective role modeling, greater
gender–role flexibility, decreased gender conflict, and more egalitarian
therapeutic environments.
The current findings dovetail with prior research demonstrating that
mental health trainees hold beliefs regarding the emotional expression
of women and men that are rooted in societal stereotypes (e.g.,
Broverman, Vogel, Broverman, Clarkson, & Rosenkrantz, 1972;
Heesacker et al., 1999; Tredinnick & Fowers, 1997; Vogel et al., 2003). In-
deed in Studies 1 and 2, two samples of mental health trainees reported
significant differences in their beliefs regarding women’s and men’s
comfort, risk, and expression of specific emotions. Women were per
-
ceived by mental health trainees as generally more emotionally expres
-
sive, whereas men were perceived as generally less emotionally expres
-
sive. Interestingly, however, with the current studies’ measurement of
specific emotions and emotional competencies we found that the mental
health trainees only overendorsed certain emotions and emotional com
-
petencies, and not others. These differential findings are consistent with
the literature that suggests that mental health trainees, while affected by
societal stereotypes, are generally less biased in their beliefs than
undergraduate samples (Heesacker et al., 1999).
An important issue to consider in interpreting these findings is whether
the data on stereotyping and the data on emotional experiences are fully
comparable. While most studies on stereotype accuracy use this compara
-
tive approach (see Diekamn et al., 2002; Judd et al., 1991; Madon et al.,
1998; McCauley & Stitt, 1978; Nisbett & Kunda, 1985; Ryan, 1996; Ryan &
328 VOGEL ET AL.
Bogart, 2001), the main concern is whether the different samples are using
the same internal scale to rate the behavior. One reason for the mental
health trainees’ overestimation of the emotional expression of men and
women compared to the self–report measures of the different comparison
groups may have been due to different situational imagination. This po
-
tential problem was reduced in the current set of studies by having the
counselors make judgments about two different and specific comparison
groups (i.e., dating individuals or clients) about which they were likely to
have direct knowledge (i.e., they were likely to have dated and to have
seen clients). Furthermore, the inclusion of the degree to which the emo
-
tional behaviors occurred in a realistic setting (i.e., Study 3) helped to
show that the differences found in Studies 1 and 2 were not likely occur
-
ring just because of differences in internal perceptions as the behavioral
data were consistent with the women’s and men’s self–report data. There
-
fore, we believe that comparing these two sources of data is informative
and becomes even more informative when coupled with observation of
overtbehavior.A strength of this article is the use of all three perspectives.
While our study examined stereotypes by comparing participants’
ratings of females with their ratings of males, future research could as-
sess stereotypes more overtly. For example, participants could be asked
to evaluate the accuracy of various stereotypical belief statements. Such
an approach may yield interesting and different findings. Furthermore,
our study examined global stereotypes, but a future study could mea-
sure perceptions of specific individuals (e.g., a therapist reports his/her
perceptions of specific male and female clients) and then compare thera-
pist perceptions with those of specific female and male clients. In such a
design, one could also compare both therapist and client perceptions
with observations of judges who are assessing overt behavior. This type
of investigation could be an important next step because research on
global perceptions does not fully reveal how stereotypes effect percep
-
tions of specific individuals. Such research may show that therapists are
less influenced by global stereotypes when thinking about a specific cli
-
ent. In this situation, therapists may be more motivated to be accurate
and may possess additional, stereotype–negating information about the
individual. Additionally, research should also explore the role of cul
-
tural diversity in trainees’ understanding of client emotionality. For ex
-
ample, people of color in general, and men of color specifically, often
face double standards with regard to their emotional experience and
emotional expression (Lazur & Majors, 1995). The possibility that such a
pattern exists in the therapeutic process, and the degree to which it
affects outcomes, needs to be fully explored.
A potential limitation of the present research is that practicing mental
health professionals were not sampled and instead mental health train
-
OVERESTIMATING GENDER DIFFERENCES 329
ees were sampled. In addition, the two mental health trainee samples
were collected from the same institution and so may not be representa
-
tive of all mental health trainees. The importance of these limitations,
however, is tempered by a review of the extant literature, which demon
-
strates few differences between the responses of practicing mental
health professionals and mental health trainees at various cites
(Heesacker et al., 1999). Still, future studies should examine the effects of
training and experience on gender–based beliefs and behavior. For ex
-
ample, one important future question is this: Does experience with cli
-
ents in general, or perhaps with specific types of clients, affect the degree
to which mental health professionals endorse sex differences in emo
-
tional expression? Additional research might also focus on how such be
-
liefs can be overcome and on what role graduate training has on this pro
-
cess. Heesacker et al. (1999) and Wester et al. (2002) suggested that most
training programs do not sufficiently teach mental health trainees about
the science of emotion. Future research may, therefore, need to examine
the impact of emotion science training that includes careful review of the
gender–emotion research literature.
REFERENCES
Balswick, J. O. (1982). Male inexpressiveness: Psychological and social aspects. In K.
Solomom & N.B. Levy (Eds.), Men in transition: Theory and therapy (pp. 131–150).
New York: Plenum Press.
Barrett, L. F., Robin, L., Pietromonaco, P. R., & Eyssell, K. M. (1998). Are women the “more
emotional” sex? Evidence from emotional experiences in social context. Cognition &
Emotion, 12, 555–578.
Benton, S. A., Roberston, J. M., Tseng, W., Newton, F. B., & Benton, S. L. (2003). Changes in
counseling center client problems across 13 years. Professional Psychology: Research
and Practice, 34, 66–72.
Brooks, G. R. (1990). The inexpressive male and vulnerability to therapist–patient sexual
exploitation. Psychotherapy, 27, 344–349.
Broverman, I. K., Vogel, S. R., Broverman, D. M., Clarkson, F. E., & Rosenkrantz, P. S.
(1972). Sex–role stereotypes: A current appraisal. Journal of Social Issues, 28, 59–78.
Clore, G. L. (1994). Why emotions are never unconscious. In P. Ekman & R. J. Davidson
(Eds.), The nature of emotion: Fundamental questions (pp. 285–290). New York: Oxford
University Press.
Day, S.X. & Schneider, P. L. (2002). Psychotherapy using distance technology: A compari
-
son of face–to–face, video, and audio treatment. Journal of Counseling Psychology, 49,
499–503.
Deaux, K. (1976). Sex: A perspective on the attitudinal process. In J. H. Harvey, W. J. Ickes,
& R. F. Kidd (Eds.), New directions in attribution research (Vol. 1, pp. 335–352).
Hillsdale, NJ: Erlbaum.
Diekamn, A. B., Eagly, A. H., & Kulesa, P. (2002). Accuracy and bias in stereotypes about
330 VOGEL ET AL.
the social and political attitudes of women and men. Journal of Experimental Social
Psychology, 38, 268–282.
Diener, E. (2000). Subjective well–being: The science of happiness and a proposal for a na
-
tional index. American Psychologist, 55, 34–43.
Doyle, J. A. (1989). The male experience (2nd ed.). Dubuque, IA: Wm. C. Brown.
Fischer,P. C., Smith, R. J., Leonard, E.,Fuqua,D.R.,Campbell, J. L., & Masters, M. A.(1993).
Sex differences on affective dimensions: Continuing examination. Journal of Coun
-
seling and Development, 71, 440–443.
Fujita, F., Diener, E., & Sandvik, E. (1991). Gender differences in negative affect and
well–being: The case for emotional intensity. Journal of Personality and Social Psychol
-
ogy, 61, 427–434.
Garb, H.N. (1997). Race bias, social class bias, and gender bias in clinical judgement. Clini
-
cal Psychology: Science & Practice, 4, 99–120.
Greenwald, M. K., Cook, E. W., III, & Lang, P. J. (1989). Affective judgment and
psychophysiological response: Dimensional covariation in the evaluation of
pictoral stimuli. Journal of Psychophysiology, 3, 51–64.
Heavey, C. L., Layne, C., & Christensen, A. (1993). Gender and conflict structure in marital
interaction:A replication and extension. Journalof Consulting and Clinical Psychology,
61, 16–27.
Heesacker, M., & Bradley, M. M. (1997). Beyond feelings: Psychotherapy and emotion. The
Counseling Psychologist, 25, 201–219.
Heesacker, M., Wester, S.,R.,Vogel, D. L.,Wentzel,J.T., Mejia–Millan, C. M., &Goodholm,
C. R., Jr. (1999).Gender–basedemotional stereotyping. Journal of Counseling Psychol-
ogy, 46, 417–429.
Hill, C. E., Corbett, M. M., Kanitz, B., Rios, P. Lightsey, R., & Gomez, M. (1992). Client be-
havior in counseling and therapy sessions: Development of a pantheoretical mea-
sure. Journal of Counseling Psychology, 39, 539–549.
Hill, C. E., & O’Brien, K. M. (1999). Helping skills: Facilitating exploration, insight, and action.
Washington, DC: American Psychological Association.
Judd, C. M., Ryan, C. S., & Park, B. (1991). Accuracy in the judgments of in–group and
out–group variability. Journal of Personality and Social Psychology, 61, 366–379.
Kahn, J. H., & Hessling, R. M. (2001). Measuring the tendency to conceal versus disclose
psychological distress. Journal of Social and Clinical Psychology, 20, 41–65.
Knudson–Martin, C. (1996). Differentiation and self–development in the relationship con
-
text. The Family Journal, 4, 188–198.
Knudson–Martin, C. (2003). How to avoid gender bias in mental health treatment. Journal
of Family Psychotherapy, 14, 45–66.
Kushner, M. G., & Sher, K. J. (1989). Fears of psychological treatment and its relation to
mental health service avoidance. Professional Psychology: Research and Practice, 20,
251–257.
Lang, P. J., Greenwald, M. K., Bradley, M. M., & Hamm, A. O. (1993). Looking at pictures:
Affective, facial, visceral, and behavioral reactions. Psychophysiology, 30, 261–273.
Lazur, R.F., & Majors, R. (1995). Men of color: Ethnocultural variations of male gender role
strain. In R.F. Levant & W.S. Pollack (Eds.), A new psychology of men (pp. 337–358).
New York: Basic Books.
Madon, S., Jussim, L., Kieper, S., Eccles, J., Smith, A., & Palumbo, P. (1998). The accuracy
and power of sex, social class, and ethnic stereotypes: A naturalistic study in person
perception. Personality and Social Psychology Bulletin, 24, 1304–1318.
McCauley, C., & Stitt, C. L. (1978). An individual and quantitative measure of stereotypes.
Journal of Personality and Social Psychology, 36, 929–939.
OVERESTIMATING GENDER DIFFERENCES 331
Nisbett, R. E., & Kunda, Z. (1985). Perception of social distributions. Journal of Personality &
Social Psychology, 48, 297–311.
Plant,E. A., Hyde, J. S.,Keltner, D.,& Devine, P. G. (2000).The gender stereotyping of emo
-
tions. Psychology of Women Quarterly, 24, 81–92.
Robertson, J., & Fitzgerald, L. F. (1990). The (mis)treatment of men: Effects of client gender
role and life–style on diagnosis and attribution of pathology. Journal of Counseling
Psychology, 37, 3–9.
Robinson M. D., & Clore, G. L. (2002). Belief and feeling: Evidence for an accessibility
model of emotional self–report. Psychological Bulletin, 128, 934–960.
Robinson, M. D., Johnson, J. T., & Shields, S. A. (1998). The gender heuristic and the data
-
base: Factors affecting the perception of gender–related differences in the experi
-
ence and display of emotions. Basic & Applied Social Psychology, 20, 206–219.
Ryan, C. S. (1996). Accuracy of black and white college students’ in–group and out–group
stereotypes. Personality and Social Psychology Bulletin, 22, 1114–1127.
Ryan, C. S. (2002). Stereotype accuracy. European Review of Social Psychology, 13, 75–109.
Ryan, C. S., & Bogart, L. M. (2001). Longitudinal changes in the accuracy of new group
members’ in–group and out–group stereotypes. Journal of Experimental Social Psy
-
chology, 37, 118–133.
Shields, S. A. (1995). The role of emotion beliefs and values on gender development. In N.
Eisenberg (Ed.), Review of personality and social psychology (Vol. 15, pp. 212–232).
Thousand Oaks, CA: Sage.
Snell, W. E., Jr., Miller, R. S., & Belk, S. S. (1988). Development of the emotional self–disclo-
sure scale. Sex Roles, 18, 59–74.
Snell, W. E., Jr., Miller, R. S., Belk, S. S., Garcia–Falconi, R., & Hernandez–Sanchez, J. E.
(1989). Men’s and women’s emotional disclosures: The impact of disclosure recipi-
ent, culture, and the masculine role. Sex Roles, 21, 467–486.
Stabb, S. D., Cox, D. L., & Harber, J. L. (1997). Gender–related therapist attributions in cou-
ples therapy: A preliminary multiple case study. Journal of Marital and Family Ther-
apy, 23, 335–346.
Tredinnick, M. G., & Fowers, B. J. (1997).Gender bias andindividualism: Responses to case
vignettes among psychologists, professors, and students. Current Psychology: Devel
-
opmental, Learning, Personality, Social, 15, 291–299.
Vogel, D. L. (2004). Self–disclosure and emotional expression. Unpublished raw data.
Vogel, D. L., Epting, F., & Wester, S.R. (2003). Counselors’ perceptions of female and male
clients: Do we reinforce traditional gender roles? Journal of Counseling and Develop
-
ment, 81, 131–141.
Vogel, D. L., & Wester (2003). To seek help or not to seek help: The risks of self–disclosure.
Journal of Counseling Psychology, 50, 351–361.
Vogel, D. L., Wester, S. R., & Heesacker, M. (1999). Dating relationships and the de
-
mand/withdraw pattern of communication. Sex Roles, 41, 297–306.
Vogel, D. L., Wester, S. R., Heesacker, M., & Madon, S. (2003). Confirming sex stereotypes:
A social role perspective. Sex Roles, 48, 519–528.
Watson, D. (2000). Mood and temperament. New York: Guilford Press.
Wester, S. R., Vogel, D. L., Pressly, P. K., & Heesacker, M. (2002). Sex differences in emo
-
tion: A critical review of the literature and implications for counseling psychology.
The Counseling Psychologist, 30, 630–652.
332 VOGEL ET AL.
