Simpson 2004

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

Journal of Applied Psychology Copyright 2004 by the American Psychological Association

2004, Vol. 89, No. 4, 715–721 0021-9010/04/$12.00 DOI: 10.1037/0021-9010.89.4.715

RESEARCH REPORTS

Gender Differences: Emotional Expression and Feelings


of Personal Inauthenticity

Patricia A. Simpson and Linda K. Stroh


Loyola University Chicago

This survey study examined the emotional expression content of human resources jobs and how the
content varies by gender. On the basis of findings, it appeared that women more often conform to
feminine display rules, which require the suppression of negative emotions and the simulation of positive
emotions. In contrast, men more often adopted masculine display rules, which require the suppression of
positive emotions and the simulation of negative ones. For both men and women, emotional dissonance
generated by a feminine display-rule pattern was positively correlated with feelings of personal inau-
thenticity at work. Gender modified the relationship between emotional dissonance and gender only in
that women who adopted the masculine display-rule pattern reported feeling the least personally
inauthentic of all.

Emotional labor—the regulation of feelings and expressions for whom are women, and bill collectors, most of whom are men, offer
organizational purposes (Grandey, 1999)— has been discussed for an example. Flight attendants, in their interactions with customers,
almost 20 years. As Erickson and Ritter (2001) have observed, are routinely expected to suppress negative emotions, such as
however, most of the studies on this subject focus on employees in anger, in favor of displaying positive emotions, such as enthusiasm
nonprofessional jobs in the service sector or have been qualitative and warmth (Hochschild, 1983). In contrast, bill collectors are
in nature. This study attempts to rectify these problems by using expected to demonstrate negative emotions, such as detachment or
time-tested quantitative measures to evaluate the emotional labor even anger and irritation, when interacting with debtors, while
of individuals in professional and managerial jobs. suppressing positive emotions, such as sympathy (Sutton, 1991).
Our first research goal was to establish whether men and women Similar results have been uncovered in other qualitative studies
report conforming to different display rules at work. Display rules (for a review, see Steinberg, 1999).
have been defined as prescriptive social norms, dictating how, Not coincidentally, these display-rule patterns complement gen-
when, and where emotions should be suppressed or expressed. The dered norms of emotional expression that have been repeatedly
content of display rules may conform to racial, ethnic, and gender associated with North American and many European cultures
stereotypes (Brody, 2000) and appears also to vary as a function of (Brody, 2000). Arguably because of their adverse impact on social
objective job characteristics (Humphrey, 2000). relationships, the expression of anger and aggression are generally
Our second research goal was to determine the nature of the seen as acceptable for men but not for women (e.g., Crick, 1997;
interface among gender, emotional expression, and mental well- Hart, De Wolf, & Burts, 1993). Conversely, emotions that facili-
being. The empirical research exploring this relationship has gen- tate social relationships, such as warmth and cheerfulness, are
erated equivocal and contradictory findings (Pugliesi, 1999). typically regarded as appropriate for women (e.g., Eder & Parker,
1987; Graham, Gentry, & Green, 1981), whereas the expression of
A Gender Division of Emotional Expression and Its positive emotions is thought to be less desirable for men (e.g.,
Possible Sources Brody, 1997, 1999).
One of the most consistent findings in the literature is the On the face of it, gendered emotional expression norms seem to
correspondence between gender differences in workplace emo- be linked to the labor market allocation process. Certainly re-
tional expression and differences in the types of work that men and searchers have argued that the perceived match between a job
women perform in the labor force. Flight attendants, most of candidate’s gender-appropriate forms of emotional regulation and
the objective emotional labor requirements of a particular job exert
a strong influence on the selection process (e.g., Arvey, Renz, &
Watson, 1998; Grandey & Brauburger, 2002). Thus, what might be
Patricia A. Simpson and Linda K. Stroh, Institute of Human Resources called a feminine display-rule pattern conforms well to the emo-
and Industrial Relations, Loyola University Chicago. tional labor requirements identified with jobs in which women
All supplemental data analyses are available from Patricia A. Simpson
predominate, whereas a masculine display-rule pattern conforms
and Linda K. Stroh on request.
Correspondence concerning this article should be addressed to Patricia well to the requirements associated with many jobs in which men
A. Simpson, Institute of Human Resources and Industrial Relations, Loyola predominate.
University Chicago, 820 North Michigan Avenue, Chicago, IL 60611. Complicating our understanding of the relationship between
E-mail: psimpso@luc.edu gender and emotional labor is evidence indicating that women and

715
716 RESEARCH REPORTS

men also demonstrate different patterns of emotional expression 1995). On the other hand, quantitative studies have not been able
even within the same jobs (Leidner, 1991, 1993; Rafaeli, 1989; to confirm a direct correlation between emotional labor and job
Wichroski, 1994). In a study of store clerks, for example, Rafaeli burnout (Wharton, 1993). Further, the quantitative literature has
(1989) found that the female clerks displayed higher levels of demonstrated that job characteristics may be more determinative
positive emotions to customers than did the male clerks. than emotional labor of adverse effects on employees (Erickson &
The issue of within-job variability in emotional expression Wharton, 1997; Parkinson, 1991).
raises an interesting question. Namely, it casts doubt on the value Contradictory results may be explained by the fact that research-
of conceptualizing emotional expression at work purely in terms of ers have often focused on different outcomes and different ap-
fixed and objective characteristics of jobs. Instead, emotional proaches to measuring dependent variables. We offer another
expression may be a fluid and subjective phenomenon that varies possible explanation for these equivocal results. In operationaliz-
considerably on the basis of assumptions about gender-appropriate ing emotional labor, most of the previous studies used discrete
behavior held either by job incumbents (e.g., Putnam & Mumby, measures to capture the incidence of either suppressing felt emo-
1993; Rafaeli & Sutton, 1991) or by their coworkers and customers tion or displaying unfelt emotion. Harking back to a concept first
(e.g., Rafaeli, 1989; Smith-Lovin, 1990). To the extent that these introduced by Hochschild (1983), we suggest that the critical
influences directly impact emotional expression, the whole notion dimension linking emotional labor to worker well-being is emo-
of emotional labor may be thrown into question. After all, emo- tional dissonance. Hochschild defined emotional dissonance as the
tional labor requires emotional regulation that brings rewards to a separation of felt emotion from emotion expressed to meet external
business organization through productivity enhancement (Diefen- expectations. She argued that the effort to maintain a “difference in
dorff & Richard, 2003), therefore it is not clear that the concept feeling and feigning over the long run leads to strain” (p. 90),
can accommodate workplace behavior that operates independently ultimately posing threats to the physical and psychological well-
of (and perhaps even at odds with) this process. being of employees.
The data confirming a gender division of emotional expression Morris and Feldman (1997) offered a further refinement of the
and generating a debate about possible sources of that division emotional dissonance concept. They argued that emotional disso-
derive almost exclusively from studies of workers in low- to nance is linked to adverse outcomes for employees in which the
midlevel service work. Analysis of professional and managerial emotional regulation pattern requires not just the simultaneous
employees has been exceedingly rare (Erickson & Ritter, 2001). suppression and display of different emotions but also the simul-
On the basis of the existing literature, however, we predict that a taneous suppression and display of conflicting emotions. The
similar gender division in emotional expression exists among illustrative example of their concept of emotional dissonance in-
professional and managerial employees. volved a situation in which an employee had to show positive
emotion to angry customers when the employee’s real feelings
Hypothesis 1: Women in professional and management jobs would have dictated a reciprocal expression of anger.
will report more often suppressing negative feelings and Although Morris and Feldman’s (1997) conceptualization of
displaying positive feelings than men in professional and emotional dissonance is a bit vague, positive and negative emo-
management jobs will report. tional responses have been correlated with different neuropsycho-
logical and behavioral activation systems (Davidson, 1994, 1995;
Hypothesis 2: Men in professional and management jobs will Tellegen, Watson, & Clark, 1999; Tomarken, Davidson, Wheeler,
report more often suppressing positive feelings and display- & Doss, 1992). Because they may trigger wholly or partly different
ing negative feelings than women in professional and man- cognitive and physiological processes, opposed suppression and
agement jobs will report. display requirements across the domains of positive and negative
emotions are arguably a critical factor affecting employee
We also draw upon the previous literature in anticipating that at
well-being.
least some portion of the gender division in emotional expression
Previous quantitative studies of emotional dissonance have un-
will be a function of job effects and therefore will be linked to
fortunately incorporated problematic measures of this construct.
differential job allocation by gender. Our specific hypothesis is as
Zerbe (2000), for example, calculated emotional dissonance as the
follows:
measured difference between a respondent’s score on an index
Hypothesis 3: Differences in how male and female managers capturing expected displays of pleasurable emotions and the score
and professionals assess the emotional labor requirements of on an index capturing actually felt pleasurable emotions. This
their jobs will correspond with differences in the type of work approach is insufficient because it focuses on the contrast between
undertaken by male and female managers and professionals. actually felt and expected displays of emotions within the same
broad affective category rather than concurrent demands to sup-
Emotional Dissonance: The Link Between Emotional press and display opposed combinations of positive and negative
emotions.
Labor and Worker Well-Being
Such flaws warrant revisiting the relationship between emo-
Another line of research that has received a great deal of tional dissonance and worker well-being. In the present study we
attention is the relationship between emotional labor and workers’ elect to focus on the relationship between emotional dissonance
mental and physical health (Adelmann, 1989; Erickson, 1991; and feelings of inauthenticity at work. As noted earlier, inauthen-
Grandey, 2003; Hochschild, 1983). On the one hand, case-study ticity has been identified as one possible type of negative outcome
research suggests that emotional labor adversely affects employee in the previous literature. Further, select studies have demonstrated
well-being and leads to job stress, burnout, and feelings of per- that inauthenticity functions as an antecedent to stress, depression,
sonal inauthenticity (Hochschild, 1983; Leidner, 1993; Pierce, and job dissatisfaction (Erickson & Wharton, 1997; Parkinson,
RESEARCH REPORTS 717

1991; Sutton, 1991). In line with Hochschild’s (1983) original sample worked in the subfield of human resources; 35% worked in indus-
formulation, we also offer the following hypothesis regarding the trial relations.1
relationship between the two variables:
Variables
Hypothesis 4: Emotional dissonance is positively related to
Emotional expression measures. The core source for our emotional
feelings of inauthenticity at work.
labor measures was the Job-Related Affective Well-Being Scale (JAWS)
developed by Van Katwyk, Fox, Spector, and Kelloway (2000). JAWS
Gender as a Moderator of the Effects of Emotional represents a context-specific means of measuring affect and covers 20
Dissonance emotions experienced at work: angry, anxious, at ease, bored, calm, con-
tent, depressed, discouraged, disgusted, ecstatic, energetic, enthusiastic,
Finally, we expect that premarket socialization processes oper- excited, fatigued, frightened, furious, gloomy, inspired, relaxed, and satis-
ate to make women more vulnerable to the effects of emotional fied. JAWS is based on Russell’s (1980) circumplex model of emotion that
dissonance. In addition to the contrasting feminine and masculine categorizes affect on the basis of two bipolar dimensions, namely pleasure
display rule-patterns described previously, researchers have also and arousal.
uncovered differential norms regarding emotional expressiveness. Respondents were asked to self-report the extent to which they were
required to cover up each emotion while they were at work (emotional
Emotional expressiveness refers to the degree to which actually
suppression) and the extent to which they were required to pretend they
felt emotions are communicated in interpersonal interactions have the emotion (emotional display). Respondents indicated their answers
(Ganong & Coleman, 1985). Research on emotional expressive- on a 5-point scale, ranging from 1 ⫽ never to 5 ⫽ every day.
ness has demonstrated that being female is associated with the Factor analysis was then used to group the emotions we measured.
unconstrained communication of actually felt emotions to others. Conducted separately for the emotional suppression and the emotional
In contrast, being male is associated with the ability to hide display items, each varimax rotation analysis extracted three factors. The
experienced emotions from others and feign alternative emotional first factor, which we labeled positive emotions, contained at ease, calm,
states (e.g., Ashmore, Del Boca, & Wohlers, 1986; Fabes & content, ecstatic, energetic, enthusiastic, excited, inspired, relaxed, and
Martin, 1991). satisfied (Cronbach’s ␣ ⫽ .96; eigenvalue ⫽ 7.69).2 The second factor,
The coexistence of gendered norms about emotional expressive- negative emotions, contained angry, anxious, depressed, discouraged, dis-
gusted, fatigued, frightened, furious, and gloomy (Cronbach’s ␣ ⫽ .90;
ness and content-specific display rules puts women into a double
eigenvalue ⫽ 5.23). The third factor referred only to the emotional state of
bind that potentially exaggerates the negative impact of emotional boredom. Because boredom appears to be independent of the emotions that
labor. As we argued earlier, both men and women may experience fit easily into positive and negative categories, we decided not to subject
adverse effects when they engage in emotionally dissonant behav- this factor to further analysis. Further supplemental series of equimax,
iors. Yet the adverse effects are potentially exaggerated for women quartermax, and varimax rotation analyses (with eigenvalue thresholds
because when they suppress actually felt emotion or simulate an greater than 1 and less than 1) yielded similar results.
unfelt emotion, they are contradicting a social identity linked to Once we had identified factors, we measured total scores across the
forthright and open communication of emotional states. Previous positive and negative emotion indices for both the suppression of emotion
research has demonstrated that adverse psychological effects can and the display of emotion. That is, we developed four total scores.
occur when individuals’ identity confirmation is threatened as a Suppress positive referred to the total score associated with suppressing
positive emotions. Suppress negative referred to the total score associated
result of behaving in a manner that contradicts a salient social
with suppressing negative emotions. Display positive referred to the total
identity (Cropanzano, Weiss, & Elias, in press; Weick, 1995). score associated with displaying positive emotions. Display negative re-
Relatedly, there is some empirical evidence to indicate that women ferred to the total score associated with displaying negative emotions.
experience greater levels of stress when they cannot express gen- Emotional dissonance measures. We created two distinct emotional
uinely held feelings (Pugliesi & Shook, 1997). Consequently, we dissonance measures because we wanted to capture the discrete effects of
hypothesize the following: the feminine and masculine display rule patterns separately. To assess the
emotional dissonance generated by the feminine display rule pattern, we
Hypothesis 5: Women will report comparatively higher levels summed the standardized score values for the suppress negative variable
of inauthenticity than will men in response to performing and for the display positive variable. This new variable was labeled
emotionally dissonant labor. Dissonance Type 1. To assess the emotional dissonance generated by the
masculine display rule pattern, we summed the standardized scores for the
suppress positive variable and for the display negative variable. This new
Method variable was labeled Dissonance Type 2. This measurement approach
Sample
1
Data for this study were derived from a survey of all graduates of a The demographic data for our respondent sample were comparable to
master’s program in human resources at a major university in the Midwest. demographic data obtained from alumni salary surveys conducted in 1989 –
This survey included sections covering demographic characteristics, type 1990 (48.4% men, 50.1% women; mean age 41 years; mean experience 13
of work, rank in the organizational hierarchy, job characteristics, number of years; 62% in human resources subfield, 38% in industrial relations). The
subordinates, emotional-expression requirements, and feelings of personal similarity in demographic attributes across the surveys suggests that the
inauthenticity. Respondents were told that the study was designed to potentially reactive content of our study did not lead to sample selection
examine the current status of human resources professionals within busi- bias.
2
ness organizations. Surveys were mailed to 1,000 currently employed Although our inclusion of calm within our roster of positive emotions
alumni. A total of 303 surveys were returned, representing a response rate runs counter to the recent work of Tellegen et al. (1999), these researchers
of approximately 30%. The sample was 49.7% women and 51.3% men. also observed that the confirmatory factor analysis values derived for the
The average age was 45.4 years (SD ⫽ 12.43). Respondents had an average emotions calm, relaxed, and placid are borderline enough to put their
tenure of 15.3 years’ experience (SD ⫽ 10.28). Sixty-five percent of the approach of classifying them as negative emotions into question.
718 RESEARCH REPORTS

represents an improvement over past approaches to capturing emotional

14


dissonance.
Dissonance Type 1 had a mean of – 0.05 and ranged in value from –2.87

.140*
13


to 4.86. Dissonance Type 2 had a mean of – 0.09 and ranged in value from
–1.61 to –7.27. Thus, the mean values for both variables suggest that for
this sample, the level of emotional dissonance resulting from both display

.484*
.270*
rule patterns is actually somewhat closer to the minimum end of the range

12


than to the maximum end of the range.
Inauthenticity. Inauthenticity at work was measured by incorporating

.097
.095
.118
two items borrowed from a 6-item scale developed by Erickson and Ritter

11


(2001). The associated 5-point scale had extreme anchors of 1 (never)
and 5 (every day). The alpha reliability for our inauthenticity scale was .73.

.645*
.124*

.313*
Gender. Gender was measured with a dummy variable coded 1 for

.109
10


women or 0 for men.
Type of work. To control for the effects of job allocation, respondents

.203*

.655*
.649*
were asked to classify their jobs into the subfield of either human resources

.090

.097

9
(1) or industrial relations (0). On the basis of a similar self-report measure,
we also created a series of functional category variables: (a) administration,
(b) human relations, and (c) strategic. These were dummy variables; the

⫺.186*
⫺.135*

.138*
⫺.154*
.070

.104

reference category was other. The administration category included task-

8
oriented activities such as administering compensation and benefits pro-
grams. The human relations category included activities such as equal
opportunity employment, performance assessment, and conflict resolution.

⫺.185*
.127*
.657*
.781*
.253*
.161*
.713*

The strategic category included activities involving the assessment of how

7
the use and development of human resources contribute to achieving
overall organizational goals and business strategies.

.240*
.138*
.757*
.136*

.862*
.861*
.238*
Control variables. The control variables included autonomy, task com-

.116

plexity, task identity, task significance, and feedback. An example of an 6
autonomy item is “I decide on my own how to go about doing the work.”
Response choices ranged from 1 (very inaccurate) to 7 (very accurate) on

.283*

.199*
.346*
.012

⫺.427
.005

.018
⫺.002
.074

a 7-point scale. Alpha reliabilities for these scales were .92 (autonomy), .91
5

(task complexity), .82 (task identity), .90 (task significance), and .86
(feedback). Because none of the job characteristic measures was significant
⫺.056

.029

.052
.092
.050
⫺.013
.026
.016
.037
.037
when entered as a discrete variable in preliminary data analyses, we created

4

a single multidimensional scale variable representing the average score by


combining the discrete variables and dividing by five. On the basis of the
Job Diagnostic Scale developed by Hackman and Oldham (1975), this
.125*

.146*
.100
.027

.166
⫺.052
.066
.098
.032
.104
.112

average score was labeled the Job Diagnostic Scale variable.


3

Because they generally occupy lower status positions in organizational


hierarchies, Hochschild (1983) maintained that women lack what she has
.152*

.122*
defined as a status shield. Their jobs may routinely require them to cover .148*
.038
.078

.034

.075
.036
.018

⫺.003
.061
⫺.059

2

up emotions such as anger and irritation, while at the same time they often
bear the brunt of higher-status, often male, coworkers’ emotional outbursts.
To control for the status shield dynamic, we used a continuous variable
⫺.137*

.478*

.308*

.387*
.190*
.036
⫺.072
⫺.036

.075
⫺.016

⫺.014
⫺.096
.440

indicating the number of subordinates reporting to a survey respondent.


1

Following Erickson and Ritter (2001), we also controlled for the fre-
quency of interaction with others at work. Social interaction at work has
1.025
1.051
19.370
0.157
10.030

1.693

1.423
0.500
1.260
0.904
0.785
1.130
1.276
0.401

been shown to have positive effects on employee well-being and may also
SD

moderate the effects of emotional labor (Wharton, 1993). Our social


interaction measure was a continuous variable capturing the percentage of
time respondents spent interacting with others while at work.
Note. n ⫽ 303. JDS ⫽ Job Diagnostic Scale.
2.260
5.898
5.760
0.905
14.790

⫺0.046

⫺0.085
0.520
0.090
⫺0.170
2.322
2.357
2.400
1.190

Finally, we controlled for experience by asking respondents how many


M

years they had spent in the workforce. Research has suggested that the
effects of emotional labor may covary with this variable (Erickson &
Ritter, 2001). Table 1 presents basic descriptive data on the variables used
Gender ⫻ Dissonance Type 1
Gender ⫻ Dissonance Type 2
Dissonance Type 1 (suppress

Dissonance Type 2 (suppress

in this study.
negative–display positive)

positive–display negative)
Number of subordinates
Means and Correlations

Data Analysis
Suppress negative
Variable

Suppress positive
Display negative
Display positive

In examining the gender division of labor, we used two series of


Inauthenticity

univariate analyses of variance (ANOVA; SPSS, Version 6.1.27.0). In the


Experience
Interaction

first series, the dependent variables were the four separate emotional
Gender

expression dimensions, namely suppress positive, suppress negative, dis-


* p ⱕ .05.
Table 1

JDS

play positive, display negative, and gender was the independent variable. In
the second series, the subfield and functional category variables were
1.
2.
3.
4.
5.
6.

7.

8.
9.
10.
11.
12.
13.
14.

added as covariates.
RESEARCH REPORTS 719

We used ordinary least squares regression techniques to analyze the Table 3


interface between emotional dissonance and feelings of inauthenticity at Results of Regression With Inauthenticity as Dependent Variable
work. In Model 1 of the regression analysis, the dependent variable,
inauthenticity, was regressed on Dissonance Type 1, Dissonance Type 2, Variable Model 1 Model 2
gender, and a set of controls. In Model 2, two interaction terms, combining
gender and the two emotional dissonance measures, were added to the Dissonance Type 1 (suppress
other independent variables. negative–display positive) 0.33* 0.37*
Dissonance Type 2 (suppress
positive–display negative) ⫺0.07 ⫺0.01
Results JDS ⫺0.14* ⫺0.12*
No. of subordinates ⫺0.01 ⫺0.01
Gender Differences in the Content of Emotional Interaction ⫺0.45 ⫺0.46
Expression Experience 0.01 ⫺0.01
Gender ⫺0.25* ⫺0.30*
Table 2 presents univariate ANOVA results associated with the Gender ⫻ Dissonance Type 1 ⫺0.01
four separate emotional expression dimensions. These results in- Gender ⫻ Dissonance Type 2 ⫺0.18*
Constant 3.66* 3.60*
dicate that there was a significant mean difference across all four R2 0.30 0.33
emotional labor dimensions. For suppressing negative emotions, ⌬R2 0.03*
the means scores for the women were significantly higher than the
mean scores for the men (Mwomen ⫽ 2.45, Mmen ⫽ 2.23, F(1, Note. n ⫽ 303. JDS ⫽ Job Diagnostic Scale.
* p ⱕ .05.
301) ⫽ 5.90, p ⱕ .05). This was also true for displaying positive
emotions (Mwomen ⫽ 2.56, Mmen ⫽ 2.21, F(1, 301) ⫽ 6.92, p ⱕ
.05). Thus, Hypothesis 1 was supported.
The opposite was true for the category covering up positive Only the display rule pattern requiring the suppression of negative
emotions (Mwomen ⫽ 2.24, Mmen ⫽ 2.63, F(1, 301) ⫽ 7.22, emotions and the display of positive emotions was positively
p ⱕ .05) and displaying negative emotions (Mwomen ⫽ 1.15, associated with feelings of inauthenticity at work (b ⫽ .33),
Mmen ⫽ 1.27, F(1, 301) ⫽ 4.78, p ⱕ .05). Thus, men are more however. These results provide only partial support for Hypothesis
likely to cover up positive emotions and display negative emotions 4; our expectation was that both forms of emotional dissonance
than are women. This evidence supports Hypothesis 2. would be positively associated with inauthenticity.
In the second ANOVA series, the subfield and functional work
category dummy variables were entered as covariates. In this The Role of Gender
analysis, the F value associated with the gender variable was no
Adding interaction terms combining gender and the emotional
longer significant for three of the emotional-expression dimen-
dissonance variables in Model 2 resulted in a significant change in
sions. The only emotional expression dimension for which a gen-
explanatory power (‚R ⫽ .03; p ⫽ .03). Further, in this model the
der difference persisted was the display of positive emotions
display-rule pattern requiring the simultaneous suppression of neg-
(Mwomen ⫽ 2.54, Mmen ⫽ 2.19, F(1, 291) ⫽ 5.11, p ⱕ .05). These
ative feelings and the expression of positive ones remained posi-
results offer partial support for Hypothesis 3, in which we postu-
tively correlated with the dependent variable (b ⫽ .37). As was the
lated that gender differences in representation patterns within jobs
case in Model 1, the correlation between the main effects gender
would correspond with gender differences in the perceived emo-
variable and inauthenticity (b ⫽ –.30) was negative. Finally, only
tional expression content of work.
one of the interaction terms, the one combining gender and the
masculine display-rule pattern, was associated with inauthenticity,
Emotional Dissonance and Feeling Inauthentic and that association was negative (b ⫽ –.18).
Table 3 reports results for two separate regression analyses. These results suggest that when the effects of the feminine
Model 1 was significant (R2 ⫽ .30, F(6, 296) ⫽ 14.49, p ⫽ .01). display-rule pattern are controlled for, women actually feel more
personally authentic at work than do men. In addition, women who
reported that their jobs require them to adopt a masculine display-
rule pattern seem to be the most likely of all to feel personally
Table 2 authentic. Gender seems to have a moderating effect on the rela-
Results of ANOVA of Emotional Expression Dimensions tionship between emotional dissonance and feelings of inauthen-
Gender with subfield ticity but not in the expected direction and only in association with
and functional category the masculine display-rule pattern. The results fail to confirm
Gender alone of work covariates Hypothesis 5.
Dependent Male Female Male Female
variable M M Fa M M Fb Conclusions

Suppress negative 2.23 2.45 5.90* 2.28 2.42 1.55 Our results suggest that male and female managers conform to
Display positive 2.21 2.56 6.92* 2.19 2.54 5.11* different emotional display patterns. The results also suggest that
Suppress positive 2.63 2.24 7.22* 2.57 2.23 3.18 apart from displays of positive emotions, the gender division of
Display negative 1.27 1.15 4.78* 1.25 1.13 3.16 emotional expression appears to be closely associated with vari-
Note. n ⫽ 303. ANOVA ⫽ analysis of variance.
able patterns of gender representation across jobs. This finding
a
df ⫽ 1, 301. b df ⫽ 1, 291. lends support to the view that job effects have more of a determi-
* p ⱕ .05. native impact on emotional expression at work than do gender
720 RESEARCH REPORTS

effects. The anomalous finding for the displaying positive emo- differences may exist in regard to what constitutes emotional
tions category does give pause, however. It raises doubts about displays, self-report measures are also problematic. Some of the
whether emotional expression in the workplace should be concep- gender results on the inauthenticity measures in this study may be
tualized purely as a function of objective job characteristics or, confounded by such differences in judgment standards. Neverthe-
more important, whether all emotional displays at work should be less, self-report measures have been used repeatedly in quantitative
conceptualized strictly as emotional labor. The results of this study studies of emotional labor (Erickson & Ritter, 2001).
also indicate that women are no more vulnerable than men to the Given that our study sample focuses exclusively on human
negative outcomes associated with emotional labor. Both men and resources professionals, future researchers should attempt to test
women appear to feel less personally authentic in the workplace to these hypotheses, using populations and samples outside the hu-
the extent that they simultaneously suppress negative emotions and man resources field. Individuals attracted to this field and its two
display positive emotions on the job. However, this finding must broad functional arenas may have certain personality attributes that
still be weighed against the data indicating that women are re- moderate the effects of emotional dissonance in some fashion. We
quired to conform to this display-rule pattern more often than men. have already discussed how personality attributes may shed light
In this indirect fashion, women as a group would be more suscep- on the finding regarding women and the masculine display-rule
tible than men to experiencing feelings of inauthenticity at work. pattern. Further studies based on samples drawn from other types
Indeed, in supplemental t test and ANOVA analyses, we found that of employees can help clarify the extent to which this study’s
women reported experiencing more of this type of emotional findings are generalizable.
dissonance than men. Direct measures of personality attributes should be included in
Two surprising findings emerged from this study. The first, that future emotional labor studies. One important starting point would
women who reported adopting a masculine display-rule pattern be disposition measures. Previous research demonstrates that neg-
were least likely to report feelings of inauthenticity, may somehow ative affectivity is correlated with greater stress, depression, and
be linked to the same attributes (e.g., assertiveness) that account somatic complaints at work (Burke, Brief, & George, 1993;
for these women’s success in male-typed jobs. The second sur- George, 1990). By extension, it is plausible that negative affectiv-
prising finding, that only the feminine display-rule pattern is ity might moderate the relationship between emotional labor and
correlated with feelings of inauthenticity, suggests that the rela- feelings of inauthenticity.
tionship between emotional dissonance and well-being is more Finally, this study also offers applied contributions. Our find-
content specific than we initially proposed. This finding comple- ings suggest that organizations need to educate managers about the
ments the work of Erickson and Ritter (2001), who found that potential ill effects associated with a feminine display-rule pattern
suppressing negative emotions such as anger is uniquely linked to and that this problem is of particular importance to female man-
feeling inauthentic at work. A related finding came from a sup- agers who are often concentrated in jobs associated with this
plementary regression analysis that examined the effects of sup- display-rule pattern. Women may also experience more emotional
pression of negative emotions and expression of positive emotions dissonance to the extent that they report higher requirements to
separately. The former type of emotional labor had the stronger display positive emotions even when they perform the same work
relationship with inauthenticity. Further research is clearly war- as men. In this regard, women may suffer from false assumptions
ranted to replicate our findings and to develop an explanation for about the relationship between emotional management behaviors
such a relationship. and optimal job performance. These false assumptions might be a
good point to start from in developing strategies designed to help
Limitations and Future Research female managers avoid the negative consequences of emotional
dissonance.
One of the strengths of this study is that it used an improved
measure of emotional dissonance. At the heart of our measure is a References
recognition that emotional dissonance arises from the simulta-
neous suppression and display of opposed combinations of posi- Adelmann, P. K. (1989). Emotional labor and employee well-being. Un-
tive and negative emotions. However, our measure remains a published doctoral dissertation, University of Michigan, Ann Arbor.
Arvey, R. D., Renz, G. L., & Watson, T. W. (1998). Emotionality and job
global one. It is based on examining the absolute difference be-
performance: Implication for personnel selection. Research in Human
tween overall intensity measures for suppressing positive emotions Resource Management, 16, 103–147.
and displaying negative emotions on the one hand and for sup- Ashmore, R. D., Del Boca, F. K., & Wohlers, A. J. (1986). The social
pressing negative emotions and displaying positive emotions on psychology of female–male relations: A critical analysis of central
the other hand. Future efforts should be directed toward the de- concepts. Orlando, FL: Academic Press.
velopment of a more direct measure of discrete instances of Brody, L. R. (1997). Beyond stereotypes: Gender and emotion. Journal of
emotional dissonance experienced at work. Social Issues, 53, 369 –394.
Future researchers may also want to include larger numbers of Brody, L. R. (1999). Gender, emotion, and the family. Cambridge, MA:
items on the inauthenticity scale. This might lead to an even higher Harvard University Press.
reliability coefficient. Brody, L. R. (2000). The socialization of gender differences in emotional
expression: Display rules, infant temperaments, and differentiation. In
The use of self-report measures in this study also raises some
A. Fisher (Ed.), Gender and emotion (pp. 24 – 47). Cambridge, MA:
important concerns. The extent to which conforming to display Cambridge University Press.
rules occurs at a conscious or subconscious, even unconscious, Burke, M. J., Brief, A. P., & George, J. M. (1993). The role of negative
level is open to debate. The possibility that this type of emotional affectivity in understanding relations between self-reports of stressors
management may be virtually automatic casts doubt on the effi- and strains: A comment on the applied psychology literature. Journal of
cacy of self-report measures. Further, to the extent that gender Applied Psychology, 78, 401– 412.
RESEARCH REPORTS 721

Crick, N. (1997). Engagement in gender normative versus nonnormative Leidner, R. (1991). Serving hamburgers and selling insurance: Gender,
forms of aggression: Links to social-psychological adjustment. Devel- work, and identity in interactive service jobs. Gender & Society, 5,
opmental Psychology, 33, 610 – 617. 154 –177.
Cropanzano, R., Weiss, H. W., & Elias, S. M. (in press). The impact of Leidner, R. (1993). Fast food, fast talk: Service work and the routinization
display rules and emotional labor on psychological well-being at work. of everyday life. Berkeley: University of California Press.
In P. L. Perrewé & D. C. Ganster (Eds.), Research in occupational stress Morris, J. A., & Feldman, D. C. (1997). Managing emotions in the
and well being. Greenwich, CT: JAI Press. workplace. Journal of Managerial Issues, 9(3), 257–274.
Davidson, R. J. (1994). Honoring biology in the study of affective style. In Parkinson, B. (1991). Emotional stylists: Strategies of expressive manage-
P. Ekman & R. J. Davidson (Eds.), The nature of emotion: Fundamental ment among trainee hairdressers. Cognition and Emotion, 5, 419 – 434.
questions (pp. 321–328). New York: Oxford University Press. Pierce, J. L. (1995). Gender trials: Emotional lives in contemporary law
Davidson, R. J. (1995). Cerebral asymmetry, emotion and affective style. firms. Berkeley: University of California Press.
In R. J. Davidson & K. Hugdahl (Eds.), Brain asymmetry (pp. 361–387). Pugliesi, K. (1999). The consequences of emotional labor: Effects of work
Cambridge, MA: MIT Press. stress, job satisfaction, and well-being. Motivation and Emotion, 23(2),
Diefendorff, J. M., & Richard, E. M. (2003). Antecedents and conse- 125–154.
quences of emotional display rule perceptions. Journal of Applied Psy- Pugliesi, K., & Shook, S. L. (1997). Gender, jobs, and emotional labor in
chology, 88, 284 –294. a complex organization. Social Perspectives on Emotion, 4, 283–316.
Eder, D., & Parker, S. (1987). The cultural production and reproduction of Putnam, L., & Mumby, D. (1993). Organizations, emotion and the myth of
gender: The effect of extracurricular activities on peer group culture. rationality. In S. Fineman (Ed.), Emotions and organizations (pp. 36 –
Sociology of Education, 60, 200 –213. 57). London: Sage.
Erickson, R. J. (1991). When emotion is the product: Self, society, and Rafaeli, A. (1989). When cashiers meet customers: An analysis of the role
(in)authenticity in a postmodern world. Unpublished doctoral disserta- of supermarket cashiers. Academy of Management Review, 32(2), 245–
tion, Washington State University, Pullman. 273.
Erickson, R. J., & Ritter, C. (2001). Emotional labor, burnout, and inau- Rafaeli, A., & Sutton, R. I. (1991). Emotional contrast strategies as means
thenticity: Does gender matter? Social Psychology Quarterly, 64, 146 – of social influence: Lessons from criminal interrogators and bill collec-
164. tors. Academy of Management Review, 34, 749 –775.
Erickson, R. J., & Wharton, A. S. (1997). Inauthenticity and depression: Russell, J. A. (1980). A circumplex model of affect. Journal of Personality
Assessing the consequences of interactive service work. Work and and Social Psychology, 39, 1161–1178.
Occupations, 24, 188 –213. Smith-Lovin, L. (1990). Emotion as the confirmation and disconfirmation
Fabes, R. A., & Martin, C. L. (1991). Gender and age stereotypes of of identity: An affect control model. In T. D. Kemper (Ed.), Research
emotionality. Personality and Social Psychology Bulletin, 17, 532–540. agendas in the sociology of emotions (pp. 238 –270). Albany, NY:
Ganong, K. G., & Coleman, M. (1985). Sex, sex roles, and emotional SUNY Press.
expressiveness. Journal of Genetic Psychology, 146(3), 405– 411. Steinberg, R. J. (1999). Emotional labor since The Managed Heart. Annals
George, J. M. (1990). Personality, affect, and behavior in groups. Journal of the American Academy of Political and Social Science, 561, 8 –26.
of Applied Psychology, 75, 107–116. Sutton, R. I. (1991). Maintaining norms about expressed emotions: The
Graham, J. W., Gentry, K. W., & Green, J. (1981). Some evidence for the case of bill collectors. Administrative Science Quarterly, 36, 245–268.
self-presentational nature of emotional expression. Personality and So- Tellegen, A., Watson, D., & Clark, L. A. (1999). On the dimensional and
cial Psychology Bulletin, 7, 467– 474. hierarchical nature of affect. Psychological Science, 10, 297–303.
Grandey, A. (1999). Emotional regulation in the workplace: A new way to Tomarken, A. J., Davidson, R. J., Wheeler, R. E., & Doss, R. C. (1992).
conceptualize emotional labor. Journal of Occupational Health Psychol- Individual differences in anterior brain asymmetry and fundamental
ogy, 55(1), 95–110. dimensions of emotion. Journal of Personality and Social Psychol-
Grandey, A. (2003). When the “show must go on”? Surface acting and ogy, 62, 676 – 687.
deep acting as determinants of emotional exhaustion and peer-rated Van Katwyk, P. T., Fox, S., Spector, P. E., & Kelloway, E. K. (2000).
service delivery. Academy of Management Journal, 4, 86-96. Using the Job-Related Affective Well-Being Scale (JAWS) to investi-
Grandey, A., & Brauburger, A. (2002). The emotion regulation behind the gate affective responses to work stressors. Journal of Occupational
customer service smile. In R. Lord, R. Klimoski, & R. Kanfer (Eds.), Health Psychology, 5, 219 –230.
Emotions in the workplace: Understanding the structure and role of Weick, K. E. (1995). Sensemaking in organizations. Thousand Oaks, CA:
emotions in organizational behavior (pp. 260 –294). San Francisco: Sage.
Jossey-Bass. Wharton, A. S. (1993). The affective consequences of service work:
Hackman, J. R., & Oldham, G. R. (1975). Development of the Job Diag- Managing emotions on the job. Work and Occupation, 20, 205–232.
nostic Survey. Journal of Applied Psychology, 55, 259 –286. Wichroski, M. A. (1994). The secretary: Invisible labor in the work world
Hart, C. H., De Wolf, M., & Burts, D. (1993). Parental disciplinary of women. Human Organization, 53, 33– 41.
strategies and preschoolers’ play behavior in playground settings. In Zerbe, W. J. (2000). Emotional dissonance and employee well-being. In
C. H. Hart (Ed.), Children in playgrounds (pp. 271–313). Albany, NY: N. M. Ashkanasy, C. E. J. Hartel, & W. J. Zerbe (Eds.), Emotions in the
SUNY Press. workplace: Research, theory, and practice (pp. 189 –214). Westport,
Hochschild, A. R. (1983). The managed heart. Berkeley: University of CT: Quorum Books.
California Press.
Humphrey, R. H. (2000). The importance of job characteristics to emo-
tionality displays. In N. M. Ashkanasy, C. E. J. Hartel, & W. J. Zerbe Received April 16, 2002
(Eds.), Emotions in the workplace: Research, theory, and practice (pp. Revision received June 26, 2003
236 –249). Westport, CT: Quorum Books. Accepted July 2, 2003 䡲

You might also like