Journal of Business Research 67 (2014) 100–107

Contents lists available at ScienceDirect

Journal of Business Research

The role of subordinate emotional masking in leader–member

exchange and outcomes: A two-sample investigation☆
Jun Xu a,⁎, 1, Yongmei Liu b, 2, Yi Guo c, 3
a
Marketing Department, DePaul University, 1 East Jackson Boulevard, Suite 7500, Chicago, IL 60604-2287, United States
b
Department of Management and Quantitative Methods, College of Business, Illinois State University, Normal, IL 61790, United States
c
Department of Business Administration, School of Business, East China University of Science and Technology, Shanghai 200237, China

a r t i c l e i n f o a b s t r a c t

Article history: Recent developments in the leadership literature call for research that emphasizes the role of the follower as well as
Accepted 1 October 2012 that of emotion in the leader–member exchange (LMX). In response to these calls, this research examines the rela-
Available online 6 December 2012 tionship between subordinate emotional masking and LMX quality, as well as the mediating role of LMX quality in
emotional masking/outcome relationships. We collect data from two samples of MBA students from the United
Keywords:
States and China, respectively. In both samples, we find that emotional masking by subordinates is negatively related
Emotion
Leadership
to LMX quality. The mediating role of LMX quality in the relationship between emotional masking and work-related
Emotion regulation outcomes is also supported. Suggestions for future research and practice are discussed.
Emotional masking © 2012 Elsevier Inc. All rights reserved.
Leader–member Exchange (LMX)

1. Introduction
Leader–member exchange (LMX) theory has been a prominent
approach to understanding leadership ever since its inception in
1970s. Based on role theory and social exchange theory, LMX theory
emphasizes the role-making process during the leader–member inter-
action, the supervisor's differentiation among subordinates, and the
formation of in-groups and out-groups in the vertical dyadic relation-
ship. At the heart of LMX theory is a relational view of leadership, and
a closer attention to the role of subordinates in leadership (House &
Aditya, 1997). This relational view of leadership is echoed by a recent
conceptual development in the leadership literature that emphasizes
the role of followership. It is argued that subordinates may demonstrate
an array of proactive behaviors that influence leader–member interac-
tions, and more research is needed to examine how subordinates
approach their role as followers in the vertical dyad (e.g., Carsten,
Uhl-Bien, West, Patera, & McGregor, 2010; Howell & Shamir, 2005).
Another recent development in the leadership literature is the
exploration of the role of emotion in leadership. Scholars have increasing-
ly recognized that emotion is an integral part of leading (Brotheridge &
Lee, 2008; McColl-Kennedy & Anderson, 2005). In a recent literature re-
view, Gooty, Connelly, Griffith, and Gupta (2010) identified sixty-three
☆ The authors would like to thank Jim Jawahar and Pam Perrewé for their helpful
comments on an earlier version of this manuscript.
⁎ Corresponding author. Tel.: +1 312 362 8421.
E-mail addresses: jxu11@depaul.edu (J. Xu), yliu2@ilstu.edu (Y. Liu),
yiguo@ecust.edu.cn (Y. Guo).
1
All authors have contributed equally to this paper.
2
Tel.: +1 309 438 2473.
3
Tel.: +86 21 6425 3210.
scholarly articles on the linkages between leadership and emotion. This
review indicates that researchers have focused primarily on leaders.
Gooty et al. (2010) concluded that current research on leadership and
emotion assumes a very passive role for followers, that “little to nothing
is known regarding follower behaviors that impact leader emotions and
affect leader's style, behaviors and perhaps quality of exchanges,” and
that “an explicit focus on followership is required” in future research
(p.1000).
In response to these recent calls for more attention on the role of
followership and that of emotion in leadership research, we focus on
subordinate emotional regulatory behavior in leader–member interac-
tions. Specifically, we examine the role of subordinate emotional masking
in LMX and work-related outcomes in a number of ways. First, we focus
on the relationship between emotional masking and LMX. Emotional
masking refers to emotional expressive behaviors that involve “perceived
discrepancies between the inner experience and the outer expression of
emotion or attempts at masking the expression of one's inner feelings
for self-presentational purposes” (Gross & John, 1998, p.175). Emotional
masking in leader–member interactions represents an emotional regula-
tion behavior in the within-organization context. Whereas research to
date has expanded the popular concept of emotional labor (Hochschild,
1983) to the within-organization context (e.g., Diefendorff & Greguras,
2009), the dynamics involved in emotional regulatory behaviors that
occur in interpersonal interactions within organizations remains less
understood.
Second, we examine a number of work-related outcomes, including
affective wellbeing, job satisfaction, and turnover intention, as they
relate to emotional masking. Due to the predominantly rational view of
organizations, emotions are typically considered the opposite of rational-
ity and professionalism (Ashforth & Humphrey, 1995). Thus, emotions

0148-2963/$ – see front matter © 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jbusres.2012.11.011
 J. Xu et al. / Journal of Business Research 67 (2014) 100–107
are closely managed at the workplace, and are frequently experienced
and expressed in inauthentic, or even distorted, ways. Trying to under-
stand the implications of such emotion management on workers'
wellbeing, work attitudes, and behaviors is an important endeavor
because, when armed with such knowledge, organizations can then
engage in practices that encourage the types of emotional expressive be-
haviors that are beneficial to organizational effectiveness and employee
wellbeing, and remove factors that facilitate dysfunctional emotional
management effort.
Lastly, we examine the mediating role of LMX in the relationships
between emotional masking and work-related outcomes (see Fig. 1
for the research model). Contemporary jobs require a great deal of
interdependence among organizational members. Thus, work relation-
ships play an important role in members' access to resources and their
ability to get things done in organizations (Brass & Burkhardt, 1993).
Past research suggests significant relationships between emotional regu-
latory behaviors and interpersonal relationship quality (e.g., Butler et al.,
2003). We expect that LMX, as a form of work relationship influential to
important contingencies at work, will mediate the relationship between
emotional masking and subordinate wellbeing, work attitudes, and inten-
tion to leave the organization. An understanding of such mediation mech-
anisms is important, because it offers insight into the important role
leader–member relationships play in subordinate wellbeing, and inform
leadership practices aimed at improving workplace effectiveness.
2. Theory and hypotheses
2.1. Emotional expression and emotional masking
Emotional expression plays an important role in interpersonal rela-
tionships. Emotional expressive cues help interaction partners under-
stand each other and adjust interpersonal behaviors according to the
relational dynamics, which forms the foundation of effective interperson-
al coordination and social exchange (Frijda & Mesquita, 1994). Important-
ly, people often express or suppress emotions for self-representation or
social-influence purposes (Goffman, 1959). In the context of leader–
member exchange, subordinates may hide their anger toward a leader
for fear of negative consequences, or they may exaggerate their liking of
a leader simply to ingratiate. In such situations, emotional masking has
occurred. We follow Gross and John (1998) to refer to emotional masking
as an emotional behavior in which individuals try to disguise their true
inner feelings by exaggerating, suppressing, or neutralizing their emo-
tional expressions. Emotional masking is common in people's daily
work lives. Mann (1999) reported with a British sample that over half
of all communications at work involve expressions of unfelt emotions.
Like emotional labor that is exercised in the service context, emo-
tional masking within organizations is conducted based upon one's per-
ceptions of the display rules, which prescribe what one should do with
one's emotions (Ekman & Friesen, 1975). Diefendorff and Richard
(2008) differentiated between prescriptive and contextual display
rules. Prescriptive display rules concern organizational expectations
for emotional expression, which often translate into the expression of
socially pleasant emotions and suppression of socially unpleasant
emotions. Contextual display rules, in contrast, concern the rules
that guide individuals' emotional expressions at any given moment.
The adoption of contextual display rules, rather than being induced
by externally prescribed expectations of organizations, is influenced
by specific contextual features such as the power dynamics between
interactive partners (Matsumoto, Yoo, Hirayama, & Petrova, 2005),
and one's interpersonal intentions and goals (Thoits, 1996). Thus, individ-
uals have much discretion in their adoption of various emotional regula-
tory strategies when trying to comply with contextual display rules.
While emotional labor in the customer service context (e.g., Grandey,
2003; Van Dolen, De Ruyter, & Lemmink, 2004) is governed closely by
prescriptive display rules, we argue that emotional regulatory behavior
101
in the within-organization context (i.e., in leader–member interactions),
is governed primarily by contextual display rules.
2.2. Emotional masking and LMX
From the interpersonal perspective, emotional masking hinders
interpersonal understanding because it distorts true feelings with
false emotional expressions. In emotional masking, inner feelings of
individuals are either suppressed, amplified, or neutralized, all of which
decrease the ability of others to know an individual's true feelings. This
decreased interpersonal understanding in turn reduces the levels of close-
ness in a relationship (cf., Gross & John, 2003). Subordinates' emotional
expression is an integral part of leader–member communication. Leaders
may, for example, use subordinates' emotional expressions to infer their
reactions to certain issues, based upon which they may be able to make
changes and adjustments. Without effective communication of emotions,
mutual understanding becomes problematic, which in turn hinders inter-
personal liking, trust, and commitment.
Moreover, emotional masking may not be effective in hiding true
feelings (Gross & John, 1998). Emotions are expressed through multiple
channels. “Leaky” channels such as body gesture or vocal tone that are
less subjective to conscious monitoring may reveal hidden emotional
information (Ekman & Friesen, 1975). Prior research suggests that
observers can detect authenticity of emotional expression, and react
negatively to inauthentic displays (Frank, Ekman, & Friesen, 1993).
When untruthful expressions are detected, the quality of an interperson-
al relationship is harmed (Gross & John, 2003). The development of
high-quality LMX relations, like all other interpersonal relationships, de-
pends in part on how emotions are experienced and expressed. In an ex-
perimental study, Newcombe and Ashkanasy (2002) found that
presenting positive information with negative emotional expressions
led to negative follower evaluations of, and a more negative relationship
quality with, the leader. We expect similar patterns to be observed when
subordinates demonstrate such emotional inauthenticity in leader–
member interactions.
Note that leaders' reactions to follower emotional expressivity may
be influenced by the perceived power differential between them. Past
research suggests that low-status individuals tend to inhibit overt
negative emotions when they interact with high-status individuals
(Diefendorff & Richard, 2008). Thus, it is likely that followers' emotional
masking may at times be well-received by leaders. However, this benefi-
cial outcome of emotional masking is unlikely to overcome the impaired
communication effectiveness due to lack of expressive emotional cues
available within the relationship. Thus, taking everything into consider-
ation, we argue that the overall relationship between emotional masking
and LMX will be negative. Thus,
Hypothesis 1. Subordinate emotional masking is negatively related to
LMX quality.
2.3. Emotional masking and work outcomes
Affective wellbeing refers to the degree to which one experiences
more positive and less negative emotions. Past research suggests that
felt authenticity is conducive to positive feelings of satisfaction, enjoy-
ment and happiness, because it satisfies the fundamental human needs
for autonomy and self-expression (Deci & Ryan, 2000). It is reasonable
to expect that individuals who mask their emotions will feel a sense of
inauthenticity, resulting in decreased level of affective wellbeing. In
line with this argument, Gross and John (1998) reported that emotional
masking was associated with increased negative emotional experiences
and decreased positive emotional experiences. Thus, we propose that
emotional masking is negatively related to affective wellbeing.
According to the Affective Event Theory (Weiss & Cropanzano, 1996),
employees' emotional experiences at work have direct influence on
work attitudes. Prior research has shown a significant link between
102 J. Xu et al. / Journal of Business Research 67 (2014) 100–107
Leader-member exchange
Subordinate emotional masking
quality
Fig. 1. Research model.
employees' moment-to-moment emotional experiences at work and job
satisfaction (e.g., Fisher, 2000). Because of the tendency of emotional
masking to give rise to negative emotions (Gross & John, 1998), emotion-
al masking should be negatively related to job satisfaction.
Furthermore, because emotional masking is related to decreased
interpersonal relationship quality (Butler et al., 2003), employees
who mask their emotions may experience difficulties in building
effective work relationships with others in the workplace. The
resulting low-quality work relationships may make it a challenge
for them to navigate the social environment at work effectively and
reduce task performance (cf., Seibert, Kraimer, & Liden, 2001). Such
negative experiences will in turn prompt those who habitually mask
their emotions to think of quitting. Thus,
Hypothesis 2. Subordinate emotional masking is negatively related to
(a) affective wellbeing, (b) job satisfaction, and positively related to
(c) turnover intention.
2.4. The mediating role of LMX
The positive psychology literature suggests that interactions with
others are important to individual wellbeing (e.g., Reis & Gable, 2003).
Quinn and Dutton (2005) suggested that high-quality interpersonal
connections increase positive affect, because they enhance a person's
sense of belonging, competency, and autonomy. Thus, given the
frequent interactions between leader and subordinates at work, high-
quality LMX should enhance the subordinate's affective wellbeing.
It is well-established in the literature that LMX is related to outcomes
of job satisfaction and turnover (Aryee & Chen, 2006; DeConinck, 2009;
Dulebohn, Bommer, Liden, Brouer, & Ferris, 2012). Given the discretion
leaders have over important contingencies such as resources and reward
allocations, the relationship with the leader is an important work relation-
ship for subordinates. Subordinates who have quality relationships with
their leaders have higher levels of communication, are assigned more
challenging tasks, and receive greater rewards (e.g., Liden, Wayne, &
Sparrowe, 2000). Thus, they should be more satisfied with their work,
and less likely to develop turnover intentions, compared with people
who are members of out-groups.
We further propose that LMX quality mediates the relationships
between subordinate emotional masking and the above-discussed
work-related outcomes. The social interaction model of emotion regula-
tion (Côté, 2005) suggests that emotional regulatory behavior has its
effects on intrapersonal outcomes through interpersonal mechanisms.
Specifically, it is argued that when individuals regulate their emotions,
their behaviors will induce reactions from an observer in either favorable
or unfavorable ways. Authentic emotional displays generate positive
reactions, because they indicate personalized attention, interest in
building interpersonal connectedness, and a lack of manipulative motives
(Grandey, Fisk, Mattila, Jansen, & Sideman, 2005; Rafaeli & Sutton, 1989).
On the other hand, when individuals express inauthentic emotions,
observers detect the inauthenticity, and interpret it as emotional dishon-
esty and a lack of interest in interpersonal interaction, leading to negative
reactions to the focal person (Frank et al., 1993). These observer reactions
will subsequently influence how focal individuals interpret their own
attitudes and behaviors.
The social interaction model of emotion regulation provides a
meaningful framework to explain intrapersonal outcomes of emotion
regulation by positing observer reactions as an interpersonal feedback
mechanism through which individuals make sense of their own
Subordinate Outcomes
Affective wellbeing at work
Job satisfaction
Turnover intention
emotional regulatory behaviors. Applying the framework to the
LMX context, a high-quality LMX relationship represents a favorable
response from the leader to their interpersonal effort. In contrast, a
low-quality LXM relationship indicates an unfavorable response.
Such interpersonal feedback will in turn influence subordinates' af-
fective wellbeing, job satisfaction, and turnover intentions. Thus,
Hypothesis 3. LMX quality mediates the relationships between subordi-
nate emotional masking and (a) affective wellbeing, (b) job satisfaction,
and (c) turnover intention.
3. Method
3.1. Samples and procedures
Data in Sample 1 were collected from two MBA programs affiliated
with two large state universities in Southeast US. Among 171 participants,
140 provided complete responses (82% response rate). Among the sam-
ple, 46% were female. Most of the participants (78%) were in their 20s,
14% were in their 30s, and 8% were over 40. Approximately 22% of the
participants held managerial positions. The average position tenure of
the participants was 27 months, ranging from 1 month to 15 years.
Data in Sample 2 were collected from students in two executive MBA
programs affiliated with two large universities in eastern China. The
questionnaire was translated into Chinese by the authors, and then
back-translated by an expert who is fluent in both Chinese and English.
Among 225 participants, 207 provided complete responses (92% re-
sponse rate). Turnover intention data were only collected with one of
the samples (N=139) due to concerns of the program coordinator of
the other program. Among the sample, 37% were female. About 27% of
the participants were in their 20s, 56% were in their 30s, and 17% were
over 40. The majority (84%) of the participants held managerial posi-
tions. The participants had been in their current positions for an average
of 41 months, ranging from 2 months to 24 years.
3.2. Measures
All measures were assessed using a 5-point Likert-type scale (1=
strongly disagree, and 5=strongly agree) except for the affective
wellbeing scale, which is explained below.
3.2.1. Emotional masking
Seven items assessing emotional masking were adapted from Gross
and John's 13-item scale (1998). The items used in the current study
were selected based on their appropriateness in leader–member ex-
change context. Sample items include “In my interactions with my super-
visor, even if I am not enjoying myself, I often pretend to be having a good
time.”
Because some items were removed from the original scale, we
conducted Confirmative Factor Analyses (CFAs) using LISREL 8.80
(Jöreskog & Sörbom, 2006) to further assess the construct validity. The
maximum likelihood method was used. The one-factor model had
acceptable fit in both samples. For the American sample, the fit indices
are as follows: χ2(9) = 21.99, p b .01, CFI = .96, NNFI = .94, and
SRMR= .045. For the Chinese sample, the fit indices are as follows:
χ2(9) = 27.8, p b .01, CFI = .97, NNFI = .95, and SRMR=.041. All the
factor loadings were significant (p b .05) and greater than twice of
their standard errors, indicating good convergent validity (Anderson &
 J. Xu et al. / Journal of Business Research 67 (2014) 100–107
Gerbing, 1988). Thus, it appears that there is adequate evidence for con-
struct validity of the adapted version of the scale.
3.2.2. LMX
The seven-item LMX scale (Scandura, Graen, & Novak, 1986) was
utilized to measure subordinates' perceptions of the quality of their
relationship with their supervisors. A sample item is “I think that
my supervisor understands my problems and needs.”
3.2.3. Affective wellbeing
Affective wellbeing was measured by the Job-related Affective
Wellbeing Scale (JAWS) developed by Van Katwyk, Fox, Spector,
and Kelloway (2000). The scale has twenty items, ten for positive
emotions and ten for negative emotions. Respondents were asked to
indicate the extent to which they experienced these emotions in gen-
eral at work. The words “at work” were highlighted in the question-
naires to ensure that the respondents will focus on their emotional
experiences at work, rather than their general emotional experiences
in life. A five-point response format was used (1= very little or not
at all; 5 = extremely often). The negative emotion items were
reverse-coded so that higher scores indicate greater affective wellbeing.
3.2.4. Job satisfaction
Job satisfaction was measured using the 3-item measure from the
Michigan Organizational Assessment. Questions in this scale include:
“All in all, I am satisfied with my job,” “In general, I don't like my job
(reverse coded),” and “In general, I like working here.”
3.2.5. Intention to turnover
Intention to turnover was measured using two items from House
and Rizzo (1972). The items are, “I often think about quitting this
job,” and “I will probably look for a new job in the next year.”
3.2.6. Control variables
Educational background may be associated with the degree to
which individuals are aware of socio-cultural and organizational
emotional display rules and the tendency to internalize these rules.
Employees holding higher positions may have more frequent interac-
tions with their supervisors, and may therefore develop closer work
relationships, which are likely governed by different emotional
norms. As tenure increases, employees may develop closer relation-
ships with the leader, which influence the emotional dynamic in the
dyad. Thus, these variables are controlled for data analyses. Educa-
tional background was coded as “0” = without university degree,
and “1” = with university degree. Position was coded as ‘0” =
non-managerial position, and “1” = managerial position. And tenure
was measured in months.
Table 1
Means, standard deviations, reliability estimates, and intercorrelations among variables of interest.
Sample 1 — American sample (N = 140)
1 2 3 4
1 Emotional masking .84
2 LMX quality −.54⁎⁎ .87
3 Affective wellbeing −.17⁎ .35⁎⁎
4 Job satisfaction −.19⁎ .48⁎⁎
5 Intention to turnover^ .16 −.35⁎⁎ −.39⁎⁎
Mean 2.52 3.55 3.85
SD .81 .76
Note: two-tail test; ⁎p b .05, ⁎⁎p b .01. Coefficient alphas are on the diagonal.
^: N = 139 for turnover in sample 2.
103
3.3. Data analyses and results
Means, standard deviations, scale reliability and intercorrelations
for the study variables are reported in Table 1. An examination of
the correlations indicates that all variables were correlated at a mod-
erate level and in the predicted directions.
3.3.1. Common method variance
Because all measures in the study were collected via subordinate
self-reports, we conducted a series of analyses to assess the potential
concern for common method biases (Podsakoff, MacKenzie,
Podsakoff, & Lee, 2003). First, the Harman (1967) one-factor test
was used. The results revealed no dominant common method factor.
The first factor accounted for 24% and 19% of variance explained in
Sample 1 and Sample 2, respectively. Second, the partial-correlation
technique is used to assess common method variance (Lindell &
Whitney, 2001). We use an item, “I can easily let others know how
I'm feeling without using words,” as the marker variable. This item
taps into one's general nonverbal encoding skill. Conceptually, there
is no plausible link between one's nonverbal encoding skill and emo-
tional regulatory behaviors at work, nor with other variables of inter-
est in the study. All significant zero-order correlations in both
samples remained significant after the partial correlation adjustments
were made. Collectively, the above evidence suggests that common
method variance is not a serious problem in the studies.
3.3.2. Hypothesis testing
Hypotheses 1 and 2 were tested using hierarchical regression
analysis. In the first step, the control variables were entered. The pre-
dictive variables were entered in Step 2. Hypothesis 1 predicted that
emotional masking is negatively related to LMX. As seen in Table 2,
subordinate emotional masking was significantly associated with
LMX in both Sample 1(β = − .53, p b .01) and Sample 2 (β = − .30,
p b .05). Thus, Hypothesis 1 was supported.
Hypothesis 2a predicted that emotional masking is negatively
related to affective wellbeing. As seen in Table 3, Step 2, in the American
sample, emotional masking was not significantly related to affective
wellbeing (β = −.14, n.s.); in the Chinese sample, however, the rela-
tionship was significant and negative (β = −.22, p b .05). Thus,
Hypothesis 2a was supported only in the Chinese sample. Hypothesis
2b predicted that emotional masking is negatively related to job satis-
faction. As can been seen in Table 4, Step 2, emotional masking was neg-
atively related to job satisfaction in both Sample 1 (β = −.17, p b .05)
and Sample 2 (β = −.22, p b .05). Thus, Hypothesis 2b was supported.
Hypothesis 2c predicted that emotional masking is positively related
to turnover intention. As seen in Table 5, Step 2, emotional masking
was significantly related to turnover intention in Sample 1 (β = .17,
p b .05), but not in Sample 2 (β = .15, n.s.). Thus, Hypothesis 2c was
supported only in the American sample.
Hypothesis 3 predicted that LMX quality mediates the relation-
ships between emotional masking and the above-mentioned three
Sample 2 — Chinese sample (N = 207)
5 1 2 3 4 5
.74
−.30⁎⁎ .77
.84 −.25⁎⁎ .34⁎⁎ .72
.58⁎⁎ .85 −.25⁎⁎ .50⁎⁎ .51⁎⁎ .84
−.67⁎⁎ .75 .12 −.36⁎⁎ −.26⁎⁎ −.64⁎⁎ .71
3.42 3.05 2.97 3.38 3.60 3.52 2.63
.46 1.02 1.19 .65 .58 .36 .82 .92
104 J. Xu et al. / Journal of Business Research 67 (2014) 100–107

Table 2 3.3.3. Additional analysis

Regression analyses on the relationship between emotional masking and LMX. Note that results relating to Hypothesis 2 indicate that some inconsis-
Variable LMX Quality tencies exist across two samples, including the relationship between
emotional masking and affective wellbeing (Hypothesis 2a), and the
Sample 1 — American Sample 2 — Chinese
sample (N = 140) sample (N = 207) relationship between emotional masking and turnover intentions
(Hypothesis 2c). To further understand these inconsistencies, we
B SE β R2 ΔR2 B SE β R2 ΔR2
B B conducted additional analysis to assess whether cultural differences
played a role. Specifically, we created a dummy variable of country
Step 1: .04 .02 .03 .02
Educational .30 .18 .14 .08 .11 .11
and combined data from the two samples. With the combined data,
background we tested a moderated mediation model with country as the potential
Position .24 .14 .14 .16 .09 .11 moderator, emotional masking as the independent variable, LMX quality
Tenure .00 .00 .05 −.00 .00 −.09 as the mediator variable, and work-related outcomes (i.e., affective
wellbeing, job satisfaction, and turnover intention) as the dependent
Step 2: .31 .29⁎⁎ .09 .06⁎⁎
Educational .03 .16 .01 .09 .11 .10 variables. We followed the Edwards and Lambert's (2007) moderated
background mediation testing approach to test both the first-stage and second
Position .17 .12 .10 .12 .09 .07 stage effects. The moderated mediation analysis results (see Table 6)
Tenure .00 .00 .01 −.00 .00 −.12 indicate that the country did not moderate the path from emotional
Emotional −.50 .07 −.53⁎⁎ −.22 .06 −.30⁎
masking
masking to LMX quality (first-stage moderation model) (axz5 =−.093,
p>.05). Similarly, country did not moderate the path from LMX quality
⁎⁎ p b .01.
⁎ p b .05.
to affective wellbeing (bmz20 =−.118, p>.05), to job satisfaction
(bmz20 =−.852, p>.05), or to turnover intention (bmz20 =−.067,
p>.05) (second-stage moderation model). These insignificant moderat-
ing effects of country indicate that country did not serve as a significant
work-related outcome variables. To best test these proposed media- factor moderating the emotional masking/LMX quality/work-related
tion hypotheses, we adopted the bootstrapping approach developed outcome relationships. Thus, cultural differences did not appear to
by Preacher and Hayes (2004). The results (see the lower part of account for the inconsistency between American and Chinese samples
Tables 3–5) showed that all of the six investigated indirect effects in in Hypotheses 2a and 2c.
the mediation analysis (three for each sample) were significant,
supporting the Hypothesis 3. Specifically, in the American sample, 4. Discussion
the bootstrapped 95% confidence interval around the three indirect
effects between emotional masking and affective wellbeing (− .17, In this research we examined the relationships among subordinate
− .05), between emotional masking and job satisfaction (− .48, emotional masking, LMX, and work-related outcomes, as well as the me-
− .22), and between emotional masking and turnover intention diating role of LMX in emotional masking/outcome relationships. Two
(.15, .44) did not include zero, indicating significant indirect effects samples, one from America and the other from China, provided largely
between emotional masking and the three work-related outcomes consistent support to the relationships proposed. Across these two sam-
via the mediating variable (LMX quality). In the Chinese sample, the ples, we found that subordinate emotional masking was negatively relat-
bootstrapped 95% confidence interval around the three indirect ed to LMX quality and job satisfaction, and that LMX mediates the
effects between emotional masking and affective wellbeing (− .08, relationships between emotional masking and the work-related out-
−.02), between emotional masking and job satisfaction (−.30, −.10), comes of affective well being, job satisfaction, and turnover intention.
and between emotional masking and job satisfaction (.07, .29) did not Our research findings indicate some levels of inconsistency between
cross zero, indicating significant indirect effects between emotional the American and Chinese samples, including the relationship between
masking and three work-related outcomes via the mediating variable emotional masking and affective wellbeing in Hypothesis 2a, and be-
(LMX quality). tween emotional masking and turnover intentions in Hypothesis 2c.

Table 3
Regression and mediation analyses on the relationships among emotional masking, LMX and affective wellbeing.

Variable Affective wellbeing

Sample 1 — American sample (N = 140) Sample 2 — Chinese sample (N = 207)

B SE B β R2 ΔR2 B SE B β R2 ΔR2

Step 1: .06 .04 .06 .04

Educational background .15 .11 .11 .10 .07 .10
Position .17 .09 .16 .18 .06 .21
Tenure .00 .00 .14 .00 .00 .04

Step 2: .08 .02⁎ .11 .05⁎

Educational background .10 .11 .08 .09 .07 .09
Position .16 .09 .15 .16 .06 .18⁎
Tenure .00 .00 .13 .00 .00 .01
Emotional masking −.08 .05 −.14 −.13 .04 −.22⁎
Bootstrap results for the indirect Point estimate SE BCA LL BCA UL Point estimate SE BCA LL BCA UL
effect of emotional masking on −.11 .03 −.17 −.05 −.05 .02 −.08 −.02
affective wellbeing via LMX quality^

^: the coefficient are based on bootstrapping 5000 re-sampling. BCA LL = bias corrected and accelerated lower limit; BCA UL = bias corrected and accelerated upper limit.
⁎ p b .05.
 J. Xu et al. / Journal of Business Research 67 (2014) 100–107 105

Table 4
Regression and mediation analyses on the relationships among emotional masking, LMX and job satisfaction.

Variable Job satisfaction

Sample 1 — American sample (N = 140) Sample 2 — Chinese sample (N = 207)

2 2
B SE B β R ΔR B SE B β R2 ΔR2

Step 1: .04 .02 .04 .03

Educational background .18 .24 .06 −.02 .16 −.00
Position .39 .19 .17⁎ .34 .13 .22
Tenure .00 .00 .08 .00 .00 .07

Step 2: .07 .03⁎ .08 .04⁎

Educational background .07 .25 .02 −.00 .15 −.01
Position .36 .19 .16 .30 .13 .19
Tenure .00 .00 .07 .00 .00 .05
Emotional masking −.21 .11 −.17⁎ −.25 .09 −.22⁎
Bootstrap results for the indirect effect of emotional Point estimate SE BCA LL BCA UL Point estimate SE BCA LL BCA UL
masking on job satisfaction via LMX quality^ −.35 .07 −.48 −.22 −.19 .05 −.30 −.10

^: the coefficient are based on bootstrapping 5000 re-samping. BCA LL = bias corrected and accelerated lower limit; BCA UL = bias corrected and accelerated upper limit.
⁎ p b .05.

Our additional analyses indicate that the inconsistent findings may be
due to reasons other than cultural differences. These may include the
relatively small sample size and differences in some demographic char-
acteristics (e.g., age, position) in the two samples. Further research with
larger and more representative samples is needed to further explore
these differences.
inherent element in the leadership process. For the most part, however,
just as with leadership research in general, that on the linkage between
leadership and emotions has also focused on the leaders. The current re-
search serves as a first step into an inquiry of the role of subordinate
emotional regulatory behaviors in the leader-follower vertical dyad.

4.1. Theoretical implication
Recent research on leadership has recognized the proactive role of
subordinates in the leadership process. Some have suggested that lead-
ership is a relationship that is co-produced by those who lead and those
who are led (Carsten et al., 2010; Howell & Shamir, 2005). However,
most leadership research has mainly focused on the influence of leaders
on subordinates (Gooty et al., 2010). Because subordinates are an inte-
gral part of the relationship, their personalities, values, motivations, and
interpersonal styles may also influence leadership effectiveness. It is
thus important to understand how subordinates may shape leader–
member interactions and leadership outcomes (Carsten et al., 2010).
Our research contributes to this endeavor with an examination of how
one type of subordinate emotional behavior, emotional masking, is as-
sociated not only with LMX, but also with important work-related
outcomes.
The role of emotion in leadership has drawn some scholarly atten-
tion in recent years. Researchers have started to view emotion as an
4.2. Limitations
Our research design is limited by the use of cross-sectional data,
which precludes determination of direction of causality. Theoretically,
there should be a feedback loop from LMX to emotional masking,
because a higher-quality relationship can accommodate a wider range
of emotional expressions without risking damaging the relationship
(cf., Parks & Floyd, 1996). Future research should utilize longitudinal
designs to examine whether LMX quality leads to greater freedom of
expression in leader–member interactions over time and, if so, how
the strengths of the two causal directions compare.
Another limitation of our research design is that no data were col-
lected from leaders, although our line of reasoning depends heavily on
how leaders perceive and react to subordinates' emotional masking.
The data being collected from a single source, the follower, also causes
concern for common method biases. Although our analyses indicate
that common method biases did not appear to be a serious concern in
this research, we encourage future research to collect data from both
leaders and followers.

Table 5
Regression and mediation analyses on the relationships among emotional masking, LMX and turnover intention.

Variable Turnover intention

Sample 1 — American sample (N = 140) Sample 2 — Chinese sample (N = 139)

B SE B β R2 ΔR2 B SE B β R2 ΔR2

Step 1: .07 .05 .03 .01

Educational background .31 .28 .09 .17 .19 .08
Position −.39 .22 −.15 −.28 .16 −.15
Tenure −.01 .00 −.18⁎ −.00 .00 −.04

Step 2: .10 .03⁎ .05 .02

Educational background .44 .29 .13 .16 .19 .07
Position −.36 .22 −.14 −.26 .16 −.14
Tenure −.01 .00 −.16 −.00 .00 −.03
Emotional masking .24 .12 .17⁎ .21 .12 .15
Bootstrap results for the indirect effect of emotional Point estimate SE BCA LL BCA UL Point estimate SE BCA LL BCA UL
masking on turnover attention via LMX quality^ .29 .07 .15 .44 .16 .06 .07 .29

^: the coefficient are based on bootstrapping 5000 re-samping. BCA LL = bias corrected and accelerated lower limit; BCA UL = bias corrected and accelerated upper limit.
⁎ p b .05.
106 J. Xu et al. / Journal of Business Research 67 (2014) 100–107
Table 6
Results of the First and Second Stage Moderated Mediation Model following Edwards
and Lambert (2007) approach.
Difference
1. First stage moderation (emotional masking's effect on LMX −.093 (p > .05)
quality via country)
2. Second stage moderation (LMX quality's effect on affective −.118 (p > .05)
wellbeing via country)
3. Second stage moderation (LMX quality's effect on job satisfaction −.852 (p > .05)
via country)
4. Second stage moderation (LMX quality's effect on turnover −.067 (p > .05)
intention via country)
Note: 1000 bootstrap samples were used to generate estimates; unstandardized path
coefficients are reported in the table.
4.3. Directions for future research
Several research questions are worth of future exploration. First, the
interpersonal effects of different types of emotional masking can be
examined. Côté (2005) suggested that the directions of emotion regula-
tion (i.e., amplification vs. suppression), as well as the valence of
the emotions being regulated, likely influence both target reactions. In
leader–member interactions, followers may feel pressure both to
suppress negative emotions and to fake positive emotions. It would be
interesting to see if these two types of emotion regulation behaviors
are indeed associated with different outcomes.
Second, leaders' attribution of subordinate behaviors may play a role
in the way subordinates' emotional expressions influence LMX quality.
Past research has examined the relationship between subordinates' at-
tribution of leader behaviors and LMX quality (Dasborough, Ashkanasy,
Tee, & Tse, 2009). It can be expected that leader attribution of subordi-
nate emotional behaviors may also influence LMX quality.
4.4. Practical implications
This study has practical implications for managers who wish to build
high-quality relationships with their subordinates. Because leaders
must frequently lean on subordinates with whom they have strong
relationships for support and resource exchanges, it is important for
leaders to understand factors that hinder the effective formation of
quality leader–member relationships. As with all other interpersonal
relationships, being able to honestly express inner feelings is not only
indicative of a close relationship, but it is conducive to such a
high-quality relationship (Parks & Floyd, 1996). Thus, effective leaders
should create a trusting emotional atmosphere in their relationships
with subordinates, and refrain from bestowing social punishment to
subordinates who express their true feelings.
The study also has implications for employees who desire to develop
and maintain a high-quality relationship with their leaders. Although
cultural and organizational emotional display norms may lead them to
hide their true emotions when they interact with a leader, it may be
more beneficial to practice emotional authenticity, given the negative
association of emotional masking with LMX and other work-related
outcomes. The achievement of such authenticity is more likely when
an individual is equipped with emotional knowledge that enables
them to be aware of, understand, and be able to influence one's own
and others' emotions. Thus, it appears beneficial to cultivate one's emo-
tional intelligence so that one can make one's feelings known in socially
acceptable, constructive ways.
Finally, given the significant associations between emotional masking
and significant outcomes at work, it is important for organizations to
foster an overall atmosphere of freedom and openness of emotional
expression. Strigent organizational display rules that sanction the
expression of integrative emotions and restrain the expression of un-
pleasant emotions may not always be beneficial to organizational
wellbeing. It is advisable for organizations to enhance employee
emotional competencies, so that they can identify ways to experience
and express emotions to enhance the quality of work life.
References
Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A
review and recommended two-step approach. Psychological Bulletin, 103(3), 411–423.
Aryee, S., & Chen, Z. X. (2006). Leader–member exchange in a Chinese context: Ante-
cedents, the mediating role of psychological empowerment and outcomes. Journal
of Business Research, 59(7), 793–801.
Ashforth, B. E., & Humphrey, R. H. (1995). Emotion in the work place: A reappraisal.
Human Relations, 48(2), 97–125.
Brass, D. J., & Burkhardt, M. E. (1993). Potential power and power use: An investigation
of structure and behavior. Academy of Management Journal, 36, 441–470.
Brotheridge, C. M., & Lee, R. T. (2008). The emotions of managing: An introduction to
the special issue. Journal of Managerial Psychology, 23(2), 108–116.
Butler, E. A., Egloff, B., Wilhelm, F. H., Smith, N. C., Erickson, E. A., & Gross, J. J. (2003).
The social consequences of expressive suppression. Emotion, 3(1), 48–67.
Carsten, M. K., Uhl-Bien, M., West, B. J., Patera, J. L., & McGregor, R. (2010). Exploring
social constructions of followership: A qualitative study. The Leadership Quarterly,
21(3), 543–562.
Côté, S. (2005). A social interaction model of the effects of emotion regulation on work
strain. Academy of Management Review, 30(3), 509–530.
Dasborough, M. T., Ashkanasy, N. M., Tee, E. Y. J., & Tse, H. H. M. (2009). What
goes around comes around: How meso-level negative emotional contagion
can ultimately determine organizational attitudes toward leaders. The Leadership
Quarterly, 20(4), 571–585.
Deci, E. L., & Ryan, R. M. (2000). The “what” and “why” of goal pursuits: Human needs
and the self-determination of behavior. Psychological Inquiry, 11(4), 227–268.
DeConinck, J. B. (2009). The effect of leader–member exchange on turnover among retail
buyers. Journal of Business Research, 62(11), 1081–1086.
Diefendorff, J. M., & Greguras, G. J. (2009). Examining the roles of targets and discrete
emotions in shaping display rule perceptions. Journal of Management, 35, 880–898.
Diefendorff, J. M., & Richard, E. M. (2008). Not all emotional display rules are created
equal: Distinguishing between prescriptive and contextual display rules. In N. M.
Ashkanasy, & C. L. Cooper (Eds.), Research companion to emotion in organizations
(pp. 316–334). Northhampton, Maine: Edward Elgar Publishing, Inc.
Dulebohn, J. H., Bommer, W. H., Liden, R. C., Brouer, R. L., & Ferris, G. R. (2012). A
meta-analysis of antecedents and consequences of leader-member exchange:
Integrating the past with an eye toward the future. Journal of Management, 38,
1715–1759.
Edwards, J. R., & Lambert, L. S. (2007). Methods for integrating moderation and mediation:
An analytical framework using moderated path analysis. Psychological Methods, 12,
1–22.
Ekman, P., & Friesen, W. (1975). Unmasking the face: A guide to recognizing emotions
from facial clues. Englewood Cliffs, NJ: Prentice-Hall.
Fisher, C. D. (2000). Mood and emotions while working: Missing pieces of job satisfac-
tion? Journal of Organizational Behavior, 21(2), 185–202.
Frank, M. G., Ekman, P., & Friesen, W. V. (1993). Behavioral markers and recognizability
of the smile of enjoyment. Journal of Personality and Social Psychology, 64, 83–93.
Frijda, N. H., & Mesquita, B. (1994). The social roles and functions of emotions. In S.
Kitayama, & H. R. Markus (Eds.), Emotion and culture: Empirical studies of mutual
influence (pp. 51–87). Washington, DC: American Psychological Association.
Goffman, E. (1959). The presentation of self in everyday life. New York: Bantam Books.
Gooty, J., Connelly, M. S., Griffith, J., & Gupta, G. (2010). Leadership, affect and emotions: A
state of the science review. The Leadership Quarterly, 21(6), 979–1004.
Grandey, A. (2003). When “the show must go on”: Surface acting and deep acting as
determinants of emotional exhaustion and peer-rated service delivery. Academy
of Management Journal, 46, 86–96.
Grandey, A., Fisk, G., Mattila, A., Jansen, K. J., & Sideman, L. (2005). Is service with a smile
enough? Authenticity of positive displays during service encounters. Organizational
Behavior and Human Decision Processes, 96(1), 38–55.
Gross, J. J., & John, O. P. (1998). Mapping the domain of expressivity: Multimethod evidence
for a hierarchical model. Journal of Personality and Social Psychology, 74(1), 170–191.
Gross, J. J., & John, O. P. (2003). Individual differences in two emotion regulation processes:
Implications for affect, relationships, and wellbeing. Journal of Personality and Social
Psychology, 85(2), 348–362.
Harman, H. H. (1967). Modern factor analysis. Chicago: University of Chicago Press.
Hochschild, A. R. (1983). The managed heart. Berkeley: University of California Press.
House, R. J., & Aditya, R. N. (1997). The social scientific study of leadership: Quo Vadis?
Journal of Management, 23(3), 409–473.
House, R. J., & Rizzo, J. R. (1972). Toward the measurement of organizational practices:
Scale development and validation. The Journal of Applied Psychology, 56(5), 388–396.
Howell, J. M., & Shamir, B. (2005). The role of followers in the charismatic leadership
process: Relationships and their consequences. Academy of Management Review,
30(1), 96–112.
Jöreskog, K. G., & Sörbom, D. (2006). LISREL 8.80 for Windows [Computer Software].
Lincolnwood, IL: Scientific Software International, Inc.
Liden, R. C., Wayne, S. J., & Sparrowe, R. T. (2000). An examination of the mediating role of
psychological empowerment on the relations between the job, interpersonal
relationships, and work outcomes. The Journal of Applied Psychology, 85(3), 407–416.
Lindell, M. K., & Whitney, D. J. (2001). Accounting for common method variance in
cross-sectional designs. The Journal of Applied Psychology, 86(1), 114–121.
Mann, S. (1999). Emotion at work: To what extent are we expressing, suppressing, or
faking it? European Journal of Work and Organizational Psychology, 8(3), 347–369.
 J. Xu et al. / Journal of Business Research 67 (2014) 100–107
Matsumoto, D., Yoo, S. H., Hirayama, S., & Petrova, G. (2005). Development and valida-
tion of a measure of display rule knowledge: The display rule assessment invento-
ry. Emotions, 5(1), 23–40.
McColl-Kennedy, J., & Anderson, R. (2005). Subordinate-manager gender combination
and perceived leadership style influence on emotions, self-esteem and organiza-
tional commitment. Journal of Business Research, 58(2), 115–125.
Newcombe, M. J., & Ashkanasy, N. M. (2002). The role of affect and affective congruence in
perceptions of leaders: An experimental study. The Leadership Quarterly, 13, 601–614.
Parks, M. R., & Floyd, K. (1996). Meanings for closeness and intimacy in friendship.
Journal of Social and Personal Relationships, 13(1), 85–107.
Podsakoff, P. M., MacKenzie, S. B., Podsakoff, N. P., & Lee, J. -Y. (2003). Common method
biases in behavioral research: A critical review of the literature and recommended
remedies. The Journal of Applied Psychology, 88(5), 879–903.
Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect
effects in simple mediation models. Behavior Research Methods, Instruments, &
Computers, 36(4), 717–731.
Quinn, R. W., & Dutton, J. E. (2005). Coordination as energy-in-conversation. Academy
of Management Review, 30(1), 36–57.
Rafaeli, A., & Sutton, R. I. (1989). The expression of emotion in organizational life. Research in
Organizational Behavior, 11, 1–42.
107
Reis, H. T., & Gable, S. L. (2003). Toward a positive psychology of relationships. In
C. L. Keyes, & J. Haidt (Eds.), Florishing: The positive person and the good life
(pp. 129–159). Washington, DC: American Psychological Association.
Scandura, T. A., Graen, G. B., & Novak, M. A. (1986). When managers decide not to decide
autocratically: An investigation of leader–member exchange and decision influence.
The Journal of Applied Psychology, 71, 579–584.
Seibert, S. E., Kraimer, M. L., & Liden, R. C. (2001). A social capital theory of career success.
Academy of Management Journal, 44, 219–237.
Thoits, P. A. (1996). Managing the emotions of others. Symbolic Interaction, 19(2), 85–109.
Van Dolen, W., De Ruyter, K., & Lemmink, J. (2004). An empirical assessment of the in-
fluence of customer emotions and contact employee performance on encounter
and relationship satisfaction. Journal of Business Research, 57(4), 437–444.
Van Katwyk, P. T., Fox, S., Spector, P. E., & Kelloway, E. K. (2000). Using the Job-related
Affective Wellbeing Scale (JAWS) to investigate affective responses to work
stressors. Journal of Occupational Health Psychology, 5(2), 219–230.
Weiss, H. M., & Cropanzano, R. (1996). Affective events theory: A theoretical discussion of
the structure, causes and consequences of affective experiences at work. In B. M.
Staw, & L. L. Cummings (Eds.), Research in Organizational Behavior, Vol. 18. (pp. 1–74)
Greenwich, CT: JAI.