ArticlePDF Available

Exchange through emoting: An emotional model of leader–member resource exchanges

Authors:
Article

Exchange through emoting: An emotional model of leader–member resource exchanges

Abstract and Figures

Despite research suggesting that emotional interactions pervade daily resource exchanges between leaders and members, the leader‐member exchange (LMX) literature has predominantly focused on the interplay between general affective experiences and the overall relationship quality. Drawing upon the affect theory of social exchange, we examine why and how discrete exchange imbalance engenders distinct emotions and shapes downstream work behaviors of the members. Results from a preregistered experimental study with 247 participants and an experience sampling study with time‐lagged reports from 79 leaders and 145 members show that a positively imbalanced exchange increases members’ subsequent leader‐directed helping via gratitude (but not via shame) and that a negatively imbalanced exchange increases members’ subsequent risk taking via pride (but not via anger). Moreover, the intensity of such effects hinges upon the average level of resource contributions of leader‐member dyads. Our research casts light on the role of transient emotions in dynamic resource exchanges between leaders and members and enriches our knowledge of within‐dyad fluctuations of social exchanges. This article is protected by copyright. All rights reserved
This content is subject to copyright. Terms and conditions apply.
1
Exchange through Emoting:
An Emotional Model of Leader-Member Resource Exchanges
Zhenyu Liao
Lusi Wu
Haoyue Jack Zhang
Northeastern University
University of Electronic Science and
Technology of China
Nanyang Technological University
Zhaoli Song
Yating Wang
National University of Singapore
National University of Singapore
Forthcoming in the journal of Personnel Psychology
Despite research suggesting that emotional interactions pervade daily resource exchanges between
leaders and members, the leader-member exchange (LMX) literature has predominantly focused
on the interplay between general affective experiences and the overall relationship quality.
Drawing upon the affect theory of social exchange, we examine why and how discrete exchange
imbalance engenders distinct emotions and shapes downstream work behaviors of the members.
Results from a preregistered experimental study with 247 participants and an experience sampling
study with time-lagged reports from 79 leaders and 145 members show that a positively
imbalanced exchange increases members’ subsequent leader-directed helping via gratitude (but
not via shame) and that a negatively imbalanced exchange increases members’ subsequent risk
taking via pride (but not via anger). Moreover, the intensity of such effects hinges upon the average
level of resource contributions of leader-member dyads. Our research casts light on the role of
transient emotions in dynamic resource exchanges between leaders and members and enriches our
knowledge of within-dyad fluctuations of social exchanges.
Keywords: leader-member exchange interactions, imbalance, emotions, leader-directed helping,
risk taking
Acknowledgments: We would like to thank Associate Editor Zhen Zhang and two anonymous reviewers for their
high-quality feedback. This research benefits from the generous, constructive comments of Raymond Sparrowe, Stuart
Bunderson, Hillary Anger Elfenbein, Ashley Hardin, Andrew Knight, and other participants in the GOMERs group at
Olin Business School, Washington University. We would also like to thank Crystal Farh and Fadel Matta for their
helpful guidance on research methods and Scott Roeder for his assistance with study administration in the Olin
Behavioral Lab. This research was partially supported by the Joseph G. Riesman Research Professorship funding
awarded to the first author. All errors are our own.
Correspondence: Lusi Wu, School of Management and Economics, University of Electronic Science and Technology
of China, 2006 Xiyuan Road, Chengdu, Sichuan, China 611731. E-mail: wulusi@uestc.edu.cn
2
Leader-member exchange (LMX), which encompasses a sequence of interdependent
resource exchanges that accumulatively generate certain levels of dyadic relationship quality
(Cropanzano & Mitchell, 2005; Liden, Sparrowe, & Wayne, 1997), is an emotionally tinged
interpersonal process
1
. Despite the integral role of emotions in LMX, they have not yet received
sufficient scholarly attention (Cropanzano, Dasborough, & Weiss, 2017). Since Dienesch and
Liden’s (1986) seminal work, only a handful of studies have examined the interplay between
affect and LMX relationship formation, revealing that interpersonal differences in general
affective experiences, such as member affection toward the leader (Liden, Wayne, & Stilwell,
1993; Sears & Hackett, 2011) and leader-member affectivity (Bauer & Green, 1996), are of
particular importance for the formation of high LMX relationships. Extending this line of work,
Cropanzano and colleagues (2017) have theorized specific patterns of emotional interactions that
leaders and members experience across different stages of LMX relationships.
Although extant research has greatly informed our knowledge of the role of affect in
developing and maintaining LMX relationships, it remains elusive what discrete emotions may
emerge from daily, recurring resource exchange interactions and how these emotions shape
ensuing exchange behaviors. This is surprising, given the fundamental premise of emotional
interactions for the development of LMX relationships. Through “a series of affectively charged
exchange interactions” (Cropanzano et al., 2017, p.248), leaders and members may experience
the ebb and flow of the nature and amount of exchanged resources that accumulatively define
their relationship quality (Liao, Liu, Li & Song, 2019; Liden et al., 1997). More importantly,
1
Extant LMX research has established two distinct conceptual aspects (Cropanzano & Mitchell, 2005; Sparrowe,
2019): one conceptualizes LMX as the relationship quality maintained between leaders and members (Graen & Uhl-
Bien, 1995); the other focuses on specific resource exchanges performed by them (Liden & Maslyn,1998; Uhl-Bien
& Maslyn, 2003). In this paper, we primarily approach LMX from the resource exchange aspect to understand the
role that discrete emotions play across iterative exchange interactions.
3
they may often encounter “one exchange or a short sequence of exchanges marked by extreme
emotional content” (Ballinger & Rockmann, 2010, p.373) that may move beyond the affective
tone underlying a given LMX relationship and spur exchange behaviors unexpectedly deviating
from the habitual reciprocation pattern accounted for by that relationship. This manifests as that
members in a low LMX relationship may make extra attempts to benefit the leader, whereas
those in a high LMX relationship may engage in exploitative opportunism that likely hurts the
leader (Emerson, 1976). Nonetheless, the scant scholarly attention to momentary exchange
emotions may constrain our understanding of psychological mechanisms that underlie the
fluctuations of exchange behaviors within leader-member dyads.
The need for studying discrete emotions emanating from tangible resource exchanges
appears to be stronger if we intend to unravel the role of the self and the other party in shaping
psychological experiences of discrete exchanges. The social exchange of leader-member dyads
comprises interdependent transactions of valuable resources leveraged by two self-interested
parties to achieve individual goals that could not be accomplished alone (Cropanzano &
Mitchell, 2005; Emerson, 1976). Members may develop parallel psychological responses toward
the self and the leader based upon contributions from two parties in one exchange. The self- and
the leader-directed mechanisms are a theoretically interrelated tandem that warrants a joint
consideration for understanding why members vary contributions from time to time. Extant
research, however, has exclusively examined leader-directed mechanisms, suggesting that over-
benefiting exchanges increase members’ obligation to reciprocate (Liden et al., 1997). Such an
asymmetric theoretical account is partly due to the cognitive-based reciprocity principle as a
predominant theoretical lens, which entails mental accounting for resources received from the
leader (Sparrowe, 2019). Taking an emotional approach to study how both the self- and other-
4
directed mechanisms come into play across exchange interactions thus would enrich our
knowledge of psychological pathways that underpin the recurring exchange interactions.
We draw on the affect theory of social exchange (Lawler, 2001; Lawler & Thye, 1999) to
examine how exchange imbalance across exchange interactions dynamically elicits members’
self- and leader-directed emotions and leads to distinct work behaviors. According to this theory,
members may develop positive or negative emotions toward the self and the leader from their
interpretation of the primary causes of the imbalance in one exchange (Lawler, 2001). A positive
imbalance, an exchange condition in which members receive more resources from than what
they contribute to their leader in a resource exchange, generates member’s gratitude toward the
leader and shame toward the self. Conversely, a negative imbalance, an exchange condition in
which members contribute more resources to than what they receive from their leader, elicits
member’s pride toward the self and anger toward the leader. Considering that the extent to
which both parties contribute to resource exchanges may shape the perceived nature of the
imbalance, we further posit that the intensity of exchange imbalance effects on emotions depends
on the average contribution level of leader-member dyads. Moreover, the self- and leader-
directed positive and negative emotions emanating from exchange imbalance prompt members to
restore the exchange equity by providing their leader with “unilateral benefits without explicit
demand for reciprocity” (Lawler, 2001, p.329) or by taking exploitative actions of opportunism.
We hence study leader directed-helping and risk taking that members might engage in
subsequently. As a robust test of our theoretical framework (Figure 1), we conducted a
preregistered experimental study and a time-lagged, multi-source experience sampling study.
Our research contributes to the literature primarily in three ways. First, we enhance the
richness of the LMX literature by taking a complementary theoretical approach to understanding
5
the role of emotions in leader-member exchange interactions. Although prior research has been
informative in revealing the interplay between general affective tendencies and LMX
relationship formation (Liden et al., 1993), it has sidestepped discrete emotions that may
instantaneously emerge from iterative exchange interactions. Those ephemeral emotions are
nevertheless pivotal in capturing subtle yet meaningful nuances that may account for members’
short-lived deviation in exchange behaviors (Ballinger & Rockmann, 2010; Liao et al., 2019).
Departing from the relational lens of LMX, our work investigates how members’ transitory
emotions emerge from exchange interactions and generate ensuing behavioral outcomes. In
doing so, our research introduces a fine-grained knowledge of momentary exchange emotions,
thereby initiating a solid step toward capturing how leaders and members dynamically exchange
resources in day-to-day interactions.
Second, our research disentangles the self- and leader-directed psychological experiences
that drive exchange-balance-maintaining behaviors by examining positive and negative emotions
directed at both parties. Extending the commonly studied other-directed cognitive mechanisms in
prior research (e.g., felt obligation to the leader, Cropanzano & Mitchell, 2005), we argue that
members may concurrently develop positive and negative emotions targeted at both themselves
and their leader based upon whether the exchange is positively or negatively imbalanced. More
importantly, considering that the extent to which both parties strive to contribute resources to the
exchange may shape the nature of exchange imbalance perceived by members, we propose that
leader-member average contribution moderates the imbalance effects on emotions directed at
distinct targets and downstream behaviors. This represents an important contribution, because we
advance the understanding of the role of emotions with distinct valence and targets in explicating
psychological mechanisms underlying the exchange balance restoration process.
6
Finally, our research adds to theoretical accounts and empirical evidence of inequity
resolution behaviors by studying leader-directed helping and risk taking. We suggest that
although members would make extra efforts to help their leader when overbenefited (Liden et al.,
1997), when underbenefited, they take opportunistic actions to reinstate the exchange balance.
Such opportunistic exploitation, despite the possibility of incurring potential costs on the leader
(Adams, 1965; Hollander, 1958; Lawler, 2001), inherently differs from negative exchange
behaviors predicted by the negative reciprocity rationale such as interpersonal deviance,
aggression, or retaliation in response to leader injustice treatment or abusive supervision
(Cropanzano, Anthony, Daniels, & Hall, 2017). We argue that members may disregard possible
costs on leaders and take risky actions to retrieve deserved yet undelivered benefits accumulating
from their prior contributions. As such, our research partially answers the understudied question
of how exchange parties restore the exchange balance when they feel underbenefited.
Theoretical Development and Hypotheses
Leader-member social exchanges involve dynamic exchanges of resources between
leaders and members (Liden et al., 1997), which broadly comprise six domains: task,
information, latitude, support, attention, and influence (Graen & Scandura, 1987). Leaders
provide members with positional resources including developmental opportunities, professional
suggestions, attention, and support (i.e., leader contribution). Members, in turn, contribute work
resources such as well-accomplished tasks, information through the grapevine, and strong
interpersonal help (i.e., member contribution, Wilson, Sin, & Conlon, 2010). Akin to other
exchange parties, members are motivated to maintain the equity of exchange with their leader.
They mentally account for the amount of valuable resources they have contributed to and
received from the leader across interactions and weigh the givetake balance (Flynn, 2003;
7
Henderson & Peterson, 1992). Notably, although members are generally aware of the value of
resources provided by their leader when processing the exchange balance, due to the position
difference in organizations, leaders may contribute valuable strategic resources that are possibly
not appreciated by members in the short term (Wilson et al., 2010). Given our focus on
members’ transitory responses to exchange imbalance across a set of interactions within a certain
time frame, leader contribution captured in this research primarily revolves around resources
with the value that could be recognized by members immediately.
According to the affect theory of social exchange, when members perceive that their
gains to inputs ratio is unequal to their leader’s, they engage in an attribution process to identify
the party primarily responsible for the imbalance, which triggers specific exchange emotions and
subsequent equity-restoring behaviors (Lawler, 2001; Walster, Walster, & Berscheid, 1978).
When the leader (i.e., the other) is perceived to hold the main responsibility, members develop
leader-directed emotions: gratitude following a positive imbalance and anger following a
negative one (Lawler & Thye, 1999). When members ascribe themselves to be primarily
responsible, they experience self-directed emotions: pride following a negative imbalance and
shame following a positive one. Because exchange imbalance is a product of the relative
contribution between two parties, both members’ and leaders’ contributions could shape
emotions. The attribution process essentially reflects credit and blame assignment between the
two parties (Lawler, 2001). However, because of the interdependence inherent in exchanges,
members may not assign full credit or blame to one party. Instead, they might develop a sense of
shared responsibility for exchange outcomes. Hence, we suggest that one exchange interaction
could simultaneously trigger both leader- and self-directed emotions.
Positive Exchange Imbalance and Member Experienced Gratitude and Shame
8
A positive imbalance captures an exchange in which members receive more resources
from than what they contribute to their leader. It results from leaders’ over-contribution and
members’ under-contribution jointly, likely evoking members’ gratitude toward the leader and
shame toward themselves concurrently. Gratitude is “a feeling of appreciation in response to an
experience that is beneficial to, but not attributable to, the self” (Fehr, Fulmer, Awtrey, & Miller,
2017, p.363). Its emergence entails two factors: the acknowledgment of the receipt of gratifying
benefits, and the recognition that the other party is ascribed to the goodness (Watkins & Bell,
2017). In a positive imbalance, members receive a net gain of valuable resources that facilitate
task completion or personal development. They recognize that the received benefits derive from
their leaders’ extra efforts and contributions, so they would feel grateful toward the leader.
We expect that a positive imbalance may also make members feel ashamed of
themselves. As a self-conscious emotion, shame arises from an evaluation that one’s actions go
against widely accepted social norms or expectations and that the self is primarily responsible for
the actions (Lewis, 2007). A failure in maintaining exchange equity by providing sufficient
resources is a commonly recognized shame-inducing situation as it signals a violation of social
norms (Daniels & Robinson, 2019; Lawler, 2001). In a positively imbalanced exchange,
members may perceive that they have broken the exchange norms of equity due to their own
under contributions. They would view their shortfall in resource provision, albeit likely
unintentional, as the main cause of the leader’s underbenefiting and thus experience shame
toward themselves.
Despite the possibility that members could feel guilty following a positive imbalance, we
argue that shame demonstrates stronger conceptual coherence to our theory. First, focusing on
the global representation of the self (Lewis, 2007), shame dovetails with our overarching lens of
9
the affect theory of social exchange that parallels the self and the other party as the targets of
emotional reactions. Comparatively, guilt only involves evaluations of specific behaviors, not the
general self-concept. Second, the sense of worthlessness and uselessness associated with shame
(Lewis, 1971; Tangney, 1999) resonates with the value-based nature of leader-member resource
exchanges (Graen & Scandura, 1987). Finally, the experience of shame often involves public
exposure of personal failures, whereas guilt is a more private experience that arises from
individual processing of one’s own misbehaviors (Gehm & Scherer, 1988). Because members’
failure in meeting expectations in social exchange is observed and processed by the leader, we
expect members would experience shame instead of guilt. We thus propose:
Hypothesis 1 (H1): Members experience increased (a) gratitude and (b) shame when
their resource exchange with leaders involves a positive imbalance, compared to a
negative imbalance and an exchange balance.
Negative Exchange Imbalance and Member Experienced Anger and Pride
A negative imbalance captures an exchange in which a member contributes more to than
what they receive from a leader. We suggest that such an underbenefiting condition could be
ascribed to the leader’s under contribution and the member’s over contribution, thus
simultaneously producing members’ anger toward the leader and pride toward themselves. Anger
involves an appraisal of the responsibility for an undesirable situation caused by the other party
(Carver & Harmon-Jones, 2009). Underbenefiting exchanges for members could be anger-
triggering (Gibson & Callister, 2010), because leaders’ deficiency of resource provision may
undesirably impede members’ desired work progress, which could be avoided with leaders’
attempts to contribute equivalent resources (Adams, 1965; Walster et al., 1978). Prior studies
have well demonstrated the anger-triggering effect of under-benefiting exchanges. For example,
Guerrero et al. (2008) found that those who perceived having a “worse deal” after comparing the
10
inputs and gains of their own with those of their exchange partner reported higher anger toward
the partner. We thus expect a negative imbalance to increase members’ anger toward the leader.
Because a negative imbalance can also be caused by members’ extra resource provision,
we expect that it may evoke pride as well. Pride is a positive, self-conscious emotion emanating
from events that signal a competent and capable self (Haidt, 2003; Williams & DeSteno, 2008).
It involves a positive self-evaluation in which one credits the successful outcomes to one’s own
efforts and abilities (Kornilaki & Chlouverakis, 2004; Tracy & Robins, 2007). A negative
imbalance denotes a situation in which members benefit their leader with extra resource gains.
They could interpret it as a result of their high contributions, which would boost members’
perceptions of self-value and competence, eliciting pride in the self. Indeed, prior studies have
shown that benefit provision at the cost of one’s own resources, such as unsolicited helping,
enhances benefactors’ pride toward themselves (Kornilaki & Chlouverakis, 2004).
Hypothesis 2 (H2): Members experience increased (a) anger and (b) pride when their
resource exchange with leaders involves a negative imbalance, compared to a positive
imbalance and an exchange balance.
The Moderating Effects of Leader-Member Average Contribution
Whereas the nature of exchange imbalance and the perceived responsibility holder of the
imbalance determine the valence and targets of emotions members experience, the intensity of
such effects depends on the level of leader-member average contribution. This is because social
exchanges entail joint contributions from both leaders and members who are interdependent on
one another to achieve self-interested outcomes by following the principle of maintaining
exchange equity. In processing the extent to which each exchange benefits both parties in an
equitable fashion, members not only mentally account for and compare contributions from two
parties, but also attend to the overall amount of resources exchanged (Cropanzano & Mitchell,
11
2005). The extent to which both parties simultaneously strive to contribute may shape members’
processing of exchange imbalance with strengthened recognition of one party’s efforts or
weakened attribution to the other’s failure in fulfilling exchange responsibility (Weiner, 2014).
We hence expect that leader-member joint contribution, manifested as the level of their average
contribution, moderates the main effects of exchange imbalance on members’ emotions.
We first expect that when leader-member average contribution is high, the effect of a
positive imbalance would be stronger on gratitude and weaker on shame. In a positive imbalance
in which both parties contribute at high levels, exceeding the contribution of the other requires
more time and effort than the condition where both contribute at low levels (Weiner, 2014).
Thus, members are more likely to recognize that their gain of extra resources is derived from
leaders’ extra efforts to offer more-than-expected resources, resulting in stronger feelings of
gratitude. Simultaneously, given that members have endeavored to present high-level
contributions, they tend to attribute the positive imbalance to leaders’ unexpectedly high
contributions. They are thus less likely to perceive their failure in upholding social exchange
norms of contributing equivalent resources, leading to less experienced shame.
To illustrate, consider two distinct exchange interactions within a leader-member dyad,
Morgan and Jordan. In one scenario, leader Morgan needs to meet a client at a café at the end of
a workday but cannot get a cab. As the café happens to be on Jordan’s way home, Jordan offers
Morgan this fifteen-minute ride. On their way, Morgan shares experience and advice on work-
family balance in need by Jordan, who just got married. Comparatively, in the other scenario,
Morgan needs to attend an important business meeting thirty minutes away but in the opposite
direction of Jordan’s home. During the trip, Morgan not only offers valuable suggestions but also
connects Jordan with important colleagues from personal networks to help Jordan solve a crucial
12
task. Both exchanges seem positively imbalanced for Jordan, yet in the latter, despite both parties
putting more on the table, Morgan’s input entails a greater investment of personal resources,
leading Jordan to feel greater gratitude. Simultaneously, Jordan’s greater resource provision
might significantly reduce perceived failure in upholding exchange responsibility, producing less
shame (Weiner, 2014).
Hypothesis 3a (H3a): Leader-member average contribution moderates the relationship
between a positive imbalance and gratitude, such that the relationship is stronger when
leader-member average contribution is high (vs. low).
Hypothesis 3b (H3b): Leader-member average contribution moderates the relationship
between a positive imbalance and shame, such that the relationship is weaker when
leader-member average contribution is high (vs. low).
In a similar vein, we expect that when leader-member average contribution is high, the
effect of a negative imbalance would be stronger on pride and weaker on anger. Identical to the
above reasoning, when a negative imbalance comes with both parties contributing at high levels,
members are more likely to consider leaders’ receiving of extra resources as a result of their own
exceedingly high dedication to resource provision, experiencing stronger pride. At the same
time, as leaders also demonstrate high-level contribution, they are less likely to ascribe the
negative imbalance to leaders’ contribution deficiency, thus feeling less anger. Consider two
different exchange scenarios between Morgan and Jordan. In one exchange, member Jordan
independently prepares and presents a middle-stage report of a consulting project for clients.
Although leader Morgan is supposed to thoroughly engage in this task as well, Morgan barely
does so due to other work commitments. In the other scenario, Jordan presents a well-developed
solution for the company’s operation system to top managers after months of hard work with
helpful guidance from Morgan. This well-performed task would greatly strengthen Morgan’s
candidacy for a career award. Both exchanges seem negatively imbalanced for Jordan, but in the
second exchange, Jordan competently exerts more efforts that yield greater immediate benefits to
13
Morgan, resulting in greater pride. Concurrently, Morgan also offers great help for Jordan’s
achievement, making Jordan less likely to attribute exchange imbalance to Morgan’s insufficient
contribution and thus experience less anger.
Hypothesis 4a (H4a): Leader-member average contribution moderates the relationship
between a negative imbalance and anger, such that the relationship is weaker when
leader-member average contribution is high (vs. low).
Hypothesis 4b (H4b): Leader-member average contribution moderates the relationship
between a negative imbalance and pride, such that the relationship is stronger when
leader-member average contribution is high (vs. low).
Momentary Effects on Leader-Directed Helping Behavior and Risk-Taking Behavior
Our theoretical model thus far highlights that the intensity of exchange imbalance effects
on members’ self- and the leader-directed emotions depends on the average level of resource
contribution from both exchange parties. A more practically significant question, however,
points to what downstream work behaviors members may engage in owing to their distinct
emotional experiences. The affect theory of social exchange posits that emotions triggered by
exchange imbalance shape ensuing behaviors that aim to restore the exchange equity (Adams,
1965; Carrell & Dittrich, 1978; Lawler, 2001). Members could either alter their own resource
inputs/gains or act to adjust their leader’s inputs/gains. Specifically, when members feel
overbenefited, they may engage in leader-directed helping, which functions as an efficacious
equity-restoring strategy. Alternatively, when they feel underbenefited, they may perform risk
taking to restore exchange balance.
Leader-directed helping. Leader-directed helping refers to discretionary and affiliative
efforts that go beyond formal job responsibilities and are intended to give assistance to leaders
(Podsakoff, Whiting, Podsakoff, & Blume, 2009). This behavior represents an effective approach
for overbenefited members to restore exchange balance as it directly benefits leaders’ welfare
and task completion (Masterson, Lewis, Goldman, & Taylor, 2000). We expect that both
14
gratitude and shame elicited by a positive imbalance may drive members’ leader-directed
helping.
Gratitude is a warm and pleasant feeling that stimulates beneficiaries to act in a manner
that enhances the well-being of the benefactor (Fehr et al., 2017; McCullough et al., 2001).
When members feel grateful for the leader’s extra contribution, they would attend to the leader’s
needs and make extra efforts to provide proper facilitation without expectations for reciprocity,
as doing so could help restore the exchange equity. Echoing our arguments, prior studies have
documented that experienced gratitude enhances helping toward the benefactor (Ma, Tunney, &
Ferguson, 2017). Tsang (2006), for example, showed that gratitude induced by receiving more
financial support from another person led to increased financial support given back to that
person. Likewise, Mikulincer and Shaver (2010) found that people who were grateful for
receiving benefits would offer more help to the benefactor afterward. Thus, gratitude evoked by
a positive imbalance increases member leader-directed helping.
Shame, a negative emotion that threatens the general self-concept, would prompt
individuals to take constructive actions to repair the jeopardized self-concept (Daniels &
Robinson, 2019; Gausel, Leach, Vignoles, & Brown, 2012). For members who are ashamed of
their insufficient contribution, restoring the self-concept requires reestablishing exchange
balance with the leader. Providing extra benefits for the leader via leader-directed helping helps
members restore exchange balance, thus enabling self-concept restoration through rebuilding a
cooperative, helpful, and capable self-concept. Past studies have well demonstrated the positive
impact of shame on helping for the purpose of self-concept restoration (Daniels & Robinson,
2019; Leach & Cidam, 2015). Hence, shame elicited by a positive balance propels members to
15
perform leader-directed helping. Integrating our theoretical arguments presented above with H3a
and H3b, we propose the following moderated mediation effects of a positive imbalance.
Hypothesis 5a (H5a): Leader-member average contribution moderates the indirect effect
of a positive imbalance on leader-directed helping behavior via gratitude, such that the
indirect effect is stronger when leader-member average contribution is high (vs. low).
Hypothesis 5b (H5b): Leader-member average contribution moderates the indirect effect
of a positive imbalance on leader-directed helping behavior via shame, such that the
indirect effect is weaker when leader-member average contribution is high (vs. low).
Risk taking. Risk taking captures a set of work behaviors in pursuit of desirable benefits
yet with a probability of incurring loss or harm on the leader (Lopes, 1987; van Kleef et al.,
2021). We expect that when underbenefited, members tend to perform risk taking to obtain
deserved benefits. Although doing so may cause losses or costs on the leader, it helps restore the
exchange equity from a negative imbalance by indirectly retrieving the undelivered benefits from
the leader (Adam, 1965). Unlike reciprocating behaviors aiming to inflict harm on leaders who
engage in unfair treatments or destructive leadership (Cropanzano et al., 2017), risk taking is an
action intended to reclaim deserved yet unprovided benefits (Tversky & Kahneman, 1992;
Malhotra & Gino, 2011). Given the possibility of yielding better payoffs for both parties, this
behavior is relatively neutral in its valence nature, compared to negative reciprocity behaviors
such as leader-directed deviance.
Both anger and pride evoked by a negative imbalance may trigger member risk taking.
Anger is a high-activation negative emotion that promotes an effort to approach what “ought” to
be (Carver & Harmon-Jones, 2009). To resolve the undesired exchange inequity due to the
leader’s insufficient contribution, angry members are motivated to obtain more benefits by
propelling the leader to increase exchange inputs (Roseman et al., 1994). Directly asking for
favors and resources may leave the leader with a calculative impression and thus hurt the
exchange relationship (Flynn, 2003). Instead, angry members may take covert actions that
16
indirectly siphon resources from the leader to boost their own benefits (Fitness, 2000). Taking
certain risks at work may impose potential costs on the leader but enhance members’ task
performance or personal benefits (Adam, 1965; Lerner & Keltner, 2001). Thus, when
underbenefited, angry members may make risky attempts at work to maximize their own benefits
in resource exchanges.
Emanating from events indicating a competent self, pride motivates members to perform
risky actions at work to pursue greater self-achievements (Carver, Sinclair, & Johnson, 2010).
Having made greater contributions than their leader, proud members may feel that they deserve
the latitude from the leader to try novel, albeit risky, approaches at work to pursue better
performance, regardless of potential costs imposed on the leader (Frijda, 1986; Tracy & Robins,
2007). If the costs indeed occur, from members’ perspective, they simply offset their credits of
over contributions in prior exchanges (Lawler, 2001). Supporting our arguments, prior studies
have found that when feeling a strong sense of self competence, individuals tend to engage in
risk taking behaviors (e.g., Anderson & Galinsky, 2006; Jordan, Sivanathan, & Galinsky, 2011).
Thus, when underbenefited, the proud members would try a risky approach (e.g., use a new yet
unverified method), which may yield better work outcomes that enhance self-achievements but
expose the leader to potential costs. Integrating the above reasoning with H4a and H4b, we
propose the following moderated mediation effects of a negative imbalance.
Hypothesis 6a (H6a): Leader-member average contribution moderates the indirect effect
of a negative imbalance on risk taking behavior via anger, such that the indirect effect is
weaker when leader-member average contribution is high (vs. low).
Hypothesis 6b (H6b): Leader-member average contribution moderates the indirect effect
of a negative imbalance on risk taking behavior via pride, such that the indirect effect is
stronger when leader-member average contribution is high (vs. low).
Research Overview
17
We conducted two studies with complementary designs to test our hypotheses. Study 1
represents a preregistered experiment (https://aspredicted.org/blind.php?x=c98g6p), in which we
manipulated exchange imbalance and leader-member average contribution (Farh, Lanaj, & Ilies,
2017; Littlepage et al., 1997). The experimental design allowed us to strengthen causal inference.
In Study 2, we constructively verified and extended initial results by conducting an experience
sampling study with a time-lagged and multi-source design (Bolger et al., 2003), which allowed
us to capture within-dyad fluctuations of discrete resource exchanges and members’ momentary
emotional and behavioral responses with temporal precedence. The set of two studies with
disparate methodologies provides a robust test for the internal and external validity of our theory.
Study 1: A Preregistered Experiment
Participants and Experimental Design
We used an estimated small-to-medium effect size of the interaction effects (f2 = .05) to
determine the sample size needed (Cohen, 1988). An a priori power analysis suggests that
approximately 244 participants are required to achieve 80% power at an α of .05 (Cohen, 1992).
Considering the possible sample attrition due to some unusable responses (e.g., failing an
attention check or not understanding the task), we initially invited 274 working adults in the
United States from Prolific Academic to participate in a 3 (exchange imbalance: positive vs.
negative vs. balance) × 2 (average contribution: high vs. low) between-subjects study in
exchange for $1.00 as the base incentive and $1.50 as the bonus. Following a response screening
decision made prior to data analysis, we excluded 27 participants who provided unusable data or
knew experimental task answers. Analyses were conducted with data from 247 participants
(average age = 35.26 years, SD = 11.62; 52.2% female; 70.4% Caucasian). This data collection,
which occurred during the 2020-2021 academic year, was reviewed and approved by the
18
Institutional Review Board (Washington University in St. Louis IRB#201901100: Emotions in
leader-member exchanges).
We adapted the “lost at sea” decision-making exercise (Farh et al., 2017) to simulate
resource exchange interactions between leaders and members (please see the online
supplemental materials (Section A) for the experiment design). Participants (in the member role)
were instructed that they were part of a marketing consulting team, who would work with
another participant J. P. (a fictitious person, in the leader role) to provide advice for external
business parties on product packages targeting the general public. Following prior experimental
leadership studies (e.g., d’Adda, Darai, Pavanini, & Weber, 2017; Mayer, Nurmohamed,
Treviño, Shapiro, & Schminke, 2013), we leveraged three procedures to enhance the realism of
the leader-member hierarchical relation context in our experiment. Prior to the start of the
experimental task, participants completed a short leadership assessment survey to determine
whether they would be assigned to a leadership role. Approximately 15 seconds after the survey,
participants were informed that they were assigned into the member role based upon their own
and J. P.’s leadership assessment scores. Second, to enhance the role experience as a follower,
participants received task-related information and guidance directly from J. P. throughout the
exercise. Third, to strengthen perceptions of J. P.’s leadership position power and motivate high
engagement in the experimental exercise, participants were further informed that J. P. would
base upon their performance to determine the amount of bonus (within a range of $0.00 to $1.00)
that they could receive from the experimenter at the end of the experiment
2
.
Participants received a new assignment from J. P. to help a recreation company design an
open-water survival course as part of the sailing sports project by rank-ordering ten items
2
To ensure equal pay, all participants received a $.50 bonus regardless of their performance. We offered extra $1.00
bonus due to that the actual averaged completion time appeared longer than what we initially estimated.
19
regarding the importance of aiding survival (Littlepage et al., 1997). To create a resource
exchange context, participants were informed that the performance of two parties in this exercise
hinged on their joint efforts and they could obtain ranking hints by answering two extra task-
irrelevant questions (i.e., member contribution manipulation) or by seeking help from their
leader (i.e., leader contribution manipulation). We manipulated exchange imbalance and the
average contribution through the number of hints earned by members and provided by leaders.
When participants finished the ranking task, they were presented with their ranking accuracy
score as a performance indicator of their work (Farh et al., 2017)
3
. They then rated their
emotions, risk taking, and leader-directed helping, and responded to manipulation check
questions and demographic items, answered exercise realism questions
4
, and were debriefed and
thanked.
Experimental Materials and Measures
Exchange imbalance manipulation. We manipulated the exchange balance in the
ranking task through the number of ranking hints contributed by two parties. To obtain the
ranking hints, participants could answer two open-ended questions without standard answers
(Bushman & Baumeister, 1998). They were informed that the specific number of hints they
could receive depended upon the quality of their answers, which would be evaluated by a third-
party grading system. We varied the number of hints received by participants to manipulate
perceptions of their own contribution (i.e., member contribution). Participants were instructed
3
Robustness check analyses were conducted after controlling for each participant’s ranking accuracy score. The full
results are shown in the online supplementary materials (Sections B and C). Findings remained consistent and did
not alter our conclusions.
4
Participants answered two questions assessing the realism of the experiment scenario (Farh et al., 2017) using a 7
point scale (1 = strongly disagree to 7 = strongly agree). About 79% of participants agreed with the item “It is
realistic that I might work with a leader like J. P.” (M = 5.21, SD = 1.41) and 81% of participants agreed with the
item “At some point during my career, I will probably encounter a situation that occurred in the experimental
scenario” (M = 5.33, SD = 1.42). These checks indicated that participants found the scenario realistic.
20
that they could seek help from J. P. through emails (with two rounds), who was knowledgeable
about open-ocean survival and had connections with other experienced sailors. We varied the
content of J. P.’s responses and the number of ranking hints J. P. provided to manipulate
participants’ perceptions of leader contribution. In responses with ranking hints, J. P. offered
interpersonal support by encouraging participants and expressing confidence in their capability,
whereas in responses without ranking hints, J. P. indicated low commitment to the task and asked
participants to complete it independently. The correspondence interface was designed in a
fashion of Outlook emails, showing the subjects, sender, receiver, time sent, and quoting
participants’ prior email (see Section A of the online supplemental materials for study
procedures). Each ranking hint suggested whether a specific item was in the top or bottom halves
of the ranking along with rationale explained by Coast Guard experts (Farh et al., 2017).
Given that the task involved a total of 10 ranking items, we fixed the difference between
the number of hints contributed by both parties as 4 in two imbalance conditions (see Table 1).
Doing so enabled us to simultaneously maximize the manipulation effect of exchange imbalance
and facilitate the manipulation of average contribution. In positive imbalance conditions,
participants received more hints from J. P. than what they obtained; in the negative imbalance
conditions, participants self-obtained more hints than what they received from J. P.; in the
balance conditions, participants self-obtained and received (from J. P.) the same number of hints.
Participants responded to two respective sets of six-resource related items to verify our
exchange imbalance manipulation. Participants first compared their own and J. P.’s contributions
regarding work effort, information, support, attention, influence, and latitude using a scale
ranging from 1 = I contributed much more to J. P. than what I received in return to 7 = J. P.
contributed much more to me than what I provided in return (Flynn, 2003; α = .92). A one-way
21
analysis of variance (ANOVA) revealed a significant difference in the perception of the
contribution imbalance (F (2, 244) = 30.67, p < .001). Participants in the positive imbalance
conditions credited J. P. more (M = 3.80, SD = 1.39) than those in the balance conditions (M =
3.08, SD = 1.05, difference = .72, p = .001), who credited J. P. more than those in the negative
imbalance conditions (M = 2.22, SD = 1.43, difference = .86, p < .001). Additionally, participants
estimated how many units of effort (out of 100 units) that J. P. contributed to the task across six
resource items (α = .94). ANOVA results showed that the rating of J. P. efforts was significantly
different across three imbalance conditions (F (2, 244) = 41.32, p < .001). Participants in the
positive imbalance conditions rated more efforts from J. P. (M = 49.53, SD = 23.45) than those
in the balance conditions (M = 35.44, SD = 18.59, difference = 14.09, p < .001), who rated more
than those in the negative imbalance conditions (M = 20.00, SD = 20.10, difference = 15.44, p
< .001)
5
. These results verified that our imbalance manipulation had intended effects.
Leader-member average contribution manipulation. Within each exchange imbalance
condition, we manipulated J. P. and participants’ average level of contributions by fixing the
difference of hints contributed by both parties but varying their average contributions (see Table
1). In the negative imbalance condition with high average contribution, participants self-obtained
seven hints by extra question answering and received other three hints from J. P. (7 vs. 3); in the
negative imbalance condition with low average contribution, participants self-obtained four hints
but received zero from J. P. (4 vs. 0). In the positive imbalance condition with high average
contribution, participants self-obtained three hints and received seven hints from J. P. (3 vs. 7); in
5
We conducted a series of t-tests to check manipulations of exchange imbalance and average contribution. To
account for the family error rate, we used both Šidák’s (1967) and Bonferroni’s (1936) corrections. Šidák’s
correction suggests lowering the significance standard (alpha) from .05 to .0102; Bonferroni’s correction suggests
lowering alpha to .01. Because the p-values of all five tests are equal to or smaller than .001, they all fell below the
corrected alphas and thus were statistically significant.
22
the positive imbalance condition with low average contribution, participants self-obtained zero
hints but received four from J. P. (0 vs. 4). In the balance condition with high average
contribution, both parties contributed five hints (5 vs. 5); in the balance condition with low
average contribution, both parties equally contributed two hints (2 vs. 2).
As a manipulation check of the average contribution, participants indicated their
agreement with three items capturing the joint contribution to the decision-making exercise (e.g.,
“on average, J. P. and I contributed a lot to the task”) on a scale ranging from 1 = strongly
disagree to 7 = strongly agree (Gabriel, Diefendorff, & Erickson, 2011; α = .89). Results of a T-
test showed that participants in the high-average contribution condition rated the joint
contribution higher (M = 5.64, SD = 1.06) than those in the low-average contribution condition
(M = 3.14, SD = 1.58, t (245) = 14.64, p < .001). Our manipulation had the intended effects.
Exchange emotions measures. We measured gratitude with Emmons and McCullough’s
(2003) three-item scale (“appreciative,” “grateful,” “thankful;” α = .99), shame with Tangney et
al.’s (1996) three-item scale (“ashamed,” “humiliated,” “disgraced;” α = .95), anger with a three-
item scale (“angry,” “aggravated,” “resentful;” α = .98) adapted from Shaver et al. (1987) and
Weiss et al. (1999), and pride with a four-item scale (“proud,” “achieving,” “accomplished,”
“fulfilled;” α = .84) adapted from Dunn and Schweitzer (2005) and Tracy and Robins (2007).
Participants indicated how they felt toward J. P. (gratitude and anger), or themselves (pride and
shame) based on two parties’ task efforts (1 = not at all to 7 = extremely).
Risk taking measure. We measured risk taking using the adapted “ultimatum game”
(Andreoni et al., 2003; Güth et al., 1982). Participants were instructed to imagine that the senior
manager granted a cash award of $100 jointly shared by them and J. P. To build a more
collaborative working climate, the senior manager asked participants to make a proposal to their
23
leader, J. P., regarding how to allocate the award. If J. P. accepted the proposal, both parties
would receive the award as proposed; otherwise, neither of them would get any money. Risk
taking was measured by the amount that participants claimed for themselves larger amounts
indicated more risk taking because they carried higher possibilities of being rejected, putting
potential gains of both parties at larger risks. This behavioral measure has been widely used in
prior management studies (Larney, Rotella, & Barclay, 2019; Pillutla & Murnighan, 1995) and is
consistent with our risk taking conceptualization that is intended to maximize benefits for the self
by dissipating those for the leader (Lawler & Thye, 1999).
Leader-directed helping measure. Following Gino, Kouchaki, and Galinsky (2015), we
informed participants that J. P. was facing additional work and would like to have their help at
the close of the study. They would need to take extra 5 minutes to complete an unrelated survey
task without extra benefits. Otherwise, the study would terminate, and participants could get
paid. To reduce potential confounding from J. P.’s bonus decision power, participants were
assured that J. P. had submitted their study bonus and the decision of helping would not affect
their bonus. We used a binary variable to measure whether participants offered help (1 =
providing help and 0 = not providing help). Overall, 44.5% of participants helped J. P.
Control variables. To account for alternative cognitive mechanisms, we controlled for
felt obligation to reciprocate in testing positive imbalance effects and felt entitlement in testing
negative imbalance effects. Felt obligation to reciprocate was measured with a seven-item scale
adapted from Eisenberger et al. (2001; α = .89). Felt entitlement was measured with a five-item
scale adapted from Campbell et al. (2004; α = .86). Results of analyses with and without two
control variables yield consistent findings.
Results and Discussion
24
We conducted a series of Confirmatory Factor Analyses (CFAs) to examine the
dimensionality of the four emotion measures. A four-factor model with gratitude, pride, anger,
and shame fit the data well (χ2 (59) =96.71, p = .001, RMSEA = .05, CFI = .99, TLI = .99, SRMR
= .03), better than the alternative models (see Section D of the online supplemental materials for
detailed results), demonstrating the distinctiveness of these variables. Presented in Table 2 are
the means, standard deviations, and inter-correlations among study variables. We tested
hypotheses about the effects of exchange imbalance on emotions with ordinary least square
(OLS) regressions. Three imbalance conditions were coded with two dummy variables, one
indicating the positive imbalance conditions and the other indicating the negative imbalance
conditions (Cohen, Cohen, West, & Aiken, 2003). When testing H1, we focused on the
difference between the positive imbalance condition and the combination of other two
conditions; when testing H2, we focused on the difference between the negative imbalance
condition and the combination of other two conditions. As shown in Table 3, a positive
imbalance was positively related to gratitude (b = 1.48, t = 5.38, p < .001) and shame (b = .93, t
= 4.11, p < .001), supporting H1a and H1b. A negative imbalance was positively related to anger
(b = 1.32, t = 4.59, p < .001) and pride (b = .69, t = 4.28, p < .001), supporting H2a and H2b.
To test the hypothesized moderating effects of leader-member average contribution, we
included the binary variable of average contribution (high average = 1, low average = 0) and the
respective interaction terms into regression analyses. Results showed that the interaction term
between positive imbalance and average contribution was significantly related to gratitude (b =
1.37, t = 2.82, p = .005, ΔR2 = .17) but not shame (b = -.33, t = -.73, p = .465, ΔR2 = .04),
supporting H3a but not H3b. The interaction term between negative imbalance and average
contribution was significantly related to pride (b =.68, t = 2.12, p = .035, ΔR2 = .03) but not
25
anger (b = -.33, t = -.60, p = .550, ΔR2 = .09), supporting H4b but not H4a. Figure 2 depicts the
mean values of emotions across exchange imbalance conditions and average contribution
conditions.
Results showed that after controlling for felt obligation to reciprocate, gratitude (log odds
= .19, z = 2.07, p = .038), but not shame (log odds = -.03, z = -.35, p = .724), was positively
related to leader-directed helping; after controlling for felt entitlement, both pride (b = 2.26, t =
2.08, p = .038) and anger (b = 1.88, t = 2.96, p = .003) were positively related to risk taking. We
tested the moderated meditating effects via different emotions using the bootstrapping procedure
with 10,000 samples (Hayes, 2015). Results showed that positive imbalance had a stronger
indirect effect on leader-directed helping via gratitude when leader-member average contribution
was high (estimate = .88, SE = .28, 95%CI [.38, 1.48]), compared to when it was low (estimate
= .52, SE = .16, 95%CI [.24, .86]). The difference between the two indirect effects was
significant (estimate = .36, SE = .16, 95%CI [.10, .72]). These results support H5a. Nevertheless,
the indirect effect of positive imbalance on helping via shame was not significant when leader-
member average contribution was high (estimate = .03, SE = .07, 95%CI [-.11, .20]) or low
(estimate = .06, SE = .11, 95%CI [-.17, .27]). The difference between the two indirect effects
was not significant (estimate = -.03, SE = .07, 95%CI [-.20, .11]). Thus, H5b does not receive
support from these results.
Results showed that whereas the indirect effect of negative imbalance on risk taking via
anger was slightly stronger when leader-member average contribution was low (estimate = 3.79,
SE = 1.22, 95%CI [1.56, 6.33]) compared to when it was high (estimate = 3.47, SE = 1.52,
95%CI [.93, 6.83]), the difference between the two indirect effects was not significant (estimate
= -.32, SE = 1.11, 95%CI [-2.62, 1.82]). These results do not fully support H6a. Moreover,
26
negative imbalance had a stronger indirect effect on risk taking via pride when leader-member
average contribution was high (estimate = 6.43, SE = 2.43, 95%CI [2.20, 11.63]) compared to
when it was low (estimate = 3.80, SE = 1.25, 95%CI [1.45, 6.37]). Supporting H6b, the
difference between these two indirect effects was significant (estimate = 2.63, SE = 1.38, 95%CI
[.54, 5.87]).
Although the findings of Study 1 appear to be encouraging, this study has several
limitations. First, the between-person experimental design only allowed us to capture a single
exchange interaction, restraining us from demonstrating within-dyad fluctuations of discrete
exchanges and members’ immediate responses across time. A more rigorous approach
necessitates capturing a series of discrete exchanges between leaders and members. Second,
although the experimental design helped strengthen casual inference, it might lack the leader-
member relational context and the scope of exchanged resources might appear to be relatively
narrow, resulting in the limited external and ecological validity of our findings. A more
scrupulous test of our theory should take LMX relationship quality into account and measure a
broader scope of exchanged resources. Third, Study 1 only captured one type of behavioral
outcomes, but members may engage in various forms of leader-directed helping and risk-taking
behaviors. We sought to address these concerns in Study 2 by surveying 152 leader-follower
dyads over two weeks using the experience sampling approach.
Study 2: Leader-Member Experience Sampling Study
Sample and Procedure
The data collection of Study 2, which occurred during the 2017-2018 academic year, was
reviewed and approved by the Departmental Ethics Review (the School of Business at the
National University of Singapore #DER-18-1210: Leader-member dynamic resource
27
transactions and exchange emotions). The initial sample was comprised of 84 mid-level
managers and 169 immediate employees working in five regional branches of a medical
examination company located in Northern China. We contacted the company’s HR department
via one of the authors’ personal networks for assistance in recruiting managers and employees
who frequently interacted during the study period. Based on expressed interest, managers and
one to three of their immediate followers were invited for our research. Participants performed
various work across different departments (e.g., clinical laboratory, R&D, administration, or
marketing) and they constantly interacted on job arrangements, urgent task assignments, work
progress review, and other work-related issues in daily work, providing an appropriate site for
studying discrete resource exchanges. We collected data using the experience sampling approach
(Bolger et al., 2003) to capture the dynamic effects of discrete exchanges. Participants were
assured of the voluntariness and confidentiality of their participation prior to data collection. We
received usable survey data from 79 leaders and 145 followers, yielding a 94.0% response rate
for leaders and 85.7% for followers. Of the 79 leaders, 54.4% were female, 76.0% had college
educations or above, the mean age was 32.2 years (SD = 3.97), and their organizational tenure
averaged 63.17 months (SD = 32.2). Of the 145 followers, 69.7% were female, 51.1% had
college educations or above, their average age was 27.4 years (SD = 4.17), and their
organizational tenure averaged 34.9 months (SD = 29.3). The average dyadic tenure was 23.5
months (SD = 21.8).
Our data collection encompassed an initial baseline survey followed by two weeks of
daily surveys. Participants first completed a paper-and-pencil baseline survey that assessed
demographics at the end of the study briefing session, in which we introduced our study
procedures without disclosing specific research ideas. Members reported LMX relationship
28
quality and felt obligation to reciprocate. One week after the baseline survey, participants started
completing two daily online surveys sent via WeChat (one of the largest standalone mobile apps
for instant communication in China) over a period of 10 consecutive workdays (Monday to
Friday, two weeks). In the noon surveys (sent at 11:30 AM), members assessed the amount of
resources received from and contributed to leaders in the morning, experienced gratitude, shame,
anger, pride, and general affective states. In the afternoon surveys (sent at 5 PM), leaders rated
members’ helping behavior and members reported risk taking
6
. The time-lagged design allowed
us to capture the temporal sequence between the predictors and outcome variables, thus enabling
causal inference. Participants received 5 RMB (approximately $0.73) for each daily survey
response. We received 1088 noon surveys from members and 1113 afternoon surveys from both
leaders and members. We removed daily surveys completed beyond the scheduled time points or
those with an answer of 2 or lower on the screening question, resulting in 1042 (noon) and 876
(afternoon) valid member surveys and 877 valid leader afternoon surveys. We paired them into
845 noon-afternoon surveys (a response rate of 55.6%; a total of 1520 possible surveys) from
145 leader-follower dyads.
Measures
All survey instruments were administered in Mandarin Chinese, translated from the
original English version following standard translation-back translation procedures to ensure
meaning equivalence (Brislin, 1980).
6
To ensure participants answer daily surveys with sufficient work information, they completed daily surveys only
when they had direct interactions with their leaders/followers, including one-on-one and group meetings, informal
in-person communications, phone calls, emails, and teleconferences. Participants also responded to a screening
question to verify possessing appropriate work information by rating “How many direct interactions did you have
with your immediate leader/follower in the morning/afternoon?” on a five-point scale (1 = none, 2 = few, 3 = a
moderate amount, 4 = quite a bit, and 5 = a high amount of interactions; Liao, Yam, Johnson, Liu, & Song, 2018).
Following Barnes, Lucianetti, Bhave, and Christian (2015), we included daily surveys with a response of 3 or
greater on the scale.
29
Exchange imbalance. We operationalized exchange imbalance in daily resource
exchanges as the incongruence between leader contribution and member contribution using the
polynomial regression and response surface methodology (Edwards & Cable, 2009; Liao et al.,
2019). Work resources involve six domains (i.e., tasks, information, latitude, support, attention,
and influence). We thus measured leader and member contributions by asking members to report
how much they received from and contributed to their leaders regarding each domain of
resources during morning interactions using a scale ranging from 1 = almost none to 5 = quite a
lot. Sample items were “During the interactions that you had with your leader in this morning,
how much did you receive from him/her regarding the valuable work information?” (leader
contribution, average α = .94) and “… how much did you contribute to him/her regarding the
well-performed tasks?” (member contribution; average α = .95). We operationalized leader-
member average contribution as the mean value of leader contribution and member contribution.
Emotions. We measured gratitude (average α = .97), shame (average α = .91), and anger
(average α = .91) with the same items used in Study 1. To maintain the daily survey brevity and
increase responses, we measured pride with three items used in Study 1 (“accomplished”,
“achieving”, and “proud”; average α = .95). Members indicated the extent to which each item
captured their emotional states toward their leader (gratitude and anger) or themselves (pride and
shame) following their morning work interactions with leaders on a scale ranging from 1 = very
little or not at all to 5 = extremely.
Leader-directed helping. We measured leader-directed helping with a six-item scale
adapted from Dalal, Lam, Weiss, Welch, and Hulin (2009). Based on the afternoon interaction
with the focal member, leaders indicated their level of agreement with each item on a scale
30
ranging from 1 = strongly disagree to 5 = strongly agree (average α = .87). A sample item was
“This follower tried to help me.”
Risk taking. Risk taking was measured with a two-item scale adapted from Schilpzand,
Houston, and Cho (2018)
7
. Based upon work behavior in the afternoon, members indicated their
level of agreement with each item on a scale ranging from 1 = strongly disagree to 5 = strongly
agree (average α = .70). The two items were “I took an informed risk that might likely hurt my
leader’s performance in order to try and get better results for my work” and “I took a risk that
might likely create unexpected costs for my leader to try something different that might improve
my work.” We used the self-report measure because daily risk taking was often performed
without the knowledge of leaders (Fessler, Pillsworth, & Flamson, 2004).
Control variables. We controlled for member general affective states in the morning, so
we could provide a more robust test of the hypothesized effects. We measured member positive
and negative affect with Song, Foo, and Uy’s (2008) ten-item scale (average α = .96 for positive
affect and .88 for negative affect). At the between-dyad level, we controlled for LMX
relationship quality, which was measured with Graen and Uhl-Bien’s (1995) seven-item scale (α
= .86), because prior research has revealed that it affects members’ responses to discrete resource
exchanges with leaders (Liao et al., 2019). We also controlled for felt obligation to reciprocate as
7
The two-item scale was adapted from Schilpzand et al.’s (2018) daily risk-taking behavior measurement, which
was developed from Dewett’s (2006) eight-item scale. Results from a scale validation study with 148 full-time
working adults recruited from Prolific Academic revealed that the two-item measurement was highly correlated with
the eight-item measurement (r = .89, p = .000). Additionally, we conducted another content validation study by
following the procedure suggested by Colquitt, Sabey, Rodell, and Hill (2019). We included three orbiting
constructs that were identified as common negative exchange behaviors by the social exchange literature
(Cropanzano et al., 2017), including supervisor-directed revenge (Aquino, Tripp, & Bies, 2001), supervisor-directed
deviance (Bennett & Robinson, 2000), and supervisor-directed aggression (Douglas & Martinko, 2001). We
calculated Hinkin Tracey correspondence (htc) index to assess how well the risk taking scale corresponded to the
content of leader-directed risk taking and Hinkin Tracey distinctiveness (htd) index to assess how well the risk
taking scale distinguishes itself from the orbiting constructs (Colquitt et al., 2019). Results from 141 full-time
working adults recruited from Prolific Academic showed that the htc index was .88 (considered as “strong”) and the
htd index was .56 (considered as “very strong”). These results suggest overall strong content validity of the two-item
shortened version of the risk taking scale. Details of two studies are available from the author team.
31
it may condition the effects of exchange imbalance on momentary emotional experience. As in
Study 1, we adapted Eisenberger et al.’s (2001) seven-item scale to measure members’ general
obligation to reciprocate their leaders (α = .73). We controlled for member demographics (e.g.,
gender, age, and dyadic tenure). Analyses with and without control variables yielded similar
results that did not alter our findings.
Multilevel Confirmatory Factor Analyses (MCFAs)
We conducted a series of MCFAs to examine the dimensionality of member-rated daily
measures using Dyer, Hanges, and Hall’s (2005) approach. To maintain a sufficient sample-size-
to-parameter ratio and minimize the instability of the factor solution (Bagozzi & Edwards, 1998;
Little, Cunningham, Shahar, & Widaman, 2002), we randomly assigned items of leader and
member contributions into three-item parcels, respectively. A seven-factor baseline model
composed of leader contribution, member contribution, gratitude, pride, anger, shame, and risk
taking fit the data well (χ2 (298) = 666.86, p < .001, RMSEA = .04, CFI = .96, TLI = .95, SRMR
(Within-dyad) = .03, SRMR (Between-dyad) = .04), better than alternative models (see Section D of the
supplemental materials for detailed results). Given the relatively high correlation between leader
contribution and member contribution, we specifically examined the dimensionality of these two
measures using individual items. Results revealed that a two-factor baseline model (χ2 (106) =
444.75, p < .001, RMSEA = .06, CFI = .91, TLI = .89, SRMR (Within-dyad) = .04, SRMR (Between-dyad)
= .03) fit the data better than a one-factor model in which indicators of both constructs were set
to load on a single factor (Δχ2 (2) = 306.20, Satorra-Bentler scaled Δχ2 = 155.63, p < .001, RMSEA
= .08, CFI = .83, TLI = .80, SRMR (Within-dyad) = .07, SRMR (Between-dyad) = .04). Taken together,
these results demonstrated good discriminant validity for daily measures.
Analytical Strategy
32
Considering the multilevel structure of data (i.e., exchange variables were nested within
leader-member dyads, which were nested within leaders) and our focus on within-dyad effects of
discrete exchange imbalance, we conducted three-level polynomial regression analyses (for
detailed computations, see Edwards & Perry, 1993; Liao et al., 2019) using the random slope
approach of multilevel structural equation modeling (MSEM; Preacher, Zyphur, & Zhang, 2010;
Zhang, Zyphur, & Preacher, 2009) in Mplus 7.0. We within-dyad centered resource contribution
variables to control for between-member confounds and eliminate nonessential multicollinearity
(Enders & Tofighi, 2007; Hofmann, Griffin, & Gavin, 2000). We estimated three second-order
polynomial terms using within-dyad centered leader contribution and member contribution. To
virtually present the effects of exchange imbalance, we plotted three-dimensional response
surfaces using the corresponding five polynomial regression coefficients. We plotted leader
contribution and member contribution on the perpendicular horizontal axes and emotions or
behavioral outcomes on the vertical axis. We calculated values of pseudo-R2 as estimates of
effect sizes manifesting the amount of within-dyad variance in mediating and outcome variables
explained by proposed effects (Hofmann et al., 2000). Please see Section E in the online
supplemental materials for equations and computations details.
To examine the hypothesized main effects of exchange imbalance (H1 & H2), we
estimated MSEM models with five polynomial terms and computed the slope of the
incongruence line (i.e., member contribution = leader contribution, calculated as γ10 γ20 for
the main effects of positive imbalance; leader contribution = member contribution, calculated
as γ10 + γ20 for the main effects of negative imbalance). To reflect the proposed main effects, the
slope of the incongruence line must be significantly positive, indicating that the dependent
variables increase (or decrease) along the incongruence line from low leader contribution and
33
high member contribution to high leader contribution and low member contribution (Edwards &
Perry, 1993). To test the hypothesized contingent effects of leader-member average contribution
(H3 & H4), we examined the slope parameters of the lines parallel to the incongruence line at the
points of high and low leader-member average contribution (see Figure A in the Section E of the
online supplemental materials). Specifically, we identified the points 1 SD of leader-member
average contribution (i.e., 0.516) upward and downward Point O (0, 0) along the congruence line
as the points of high and low leader-member average contribution respectively (i.e., Points A and
B in Figure A). We then estimated the slope parameters of the lines parallel to the incongruence
line passing these two points and their difference. To test the conditional indirect effects of
exchange imbalance (H5 & H6), we used the conditional slope parameters to test H3 and H4 as
estimates of the paths between exchange imbalance and respective emotions (Edwards & Perry,
1993). We regressed two outcome variables on five polynomial terms and two relevant
emotional variables respectively to estimate the effects of emotions on behavioral outcomes. We
examined the conditional indirect effects using a Monte Carlo simulation with 20,000
replications using R (Preacher et al., 2010).
Results
Descriptive statistics and within-dyad variance. Reported in Table 4 are the means,
between- and within-dyad SDs, percentages of within-dyad variance, and inter-correlations
among study variables. Partitioning the total variance in daily variables into components at
within- and between-dyad, and between-leader levels, we found that daily variables had
significant within-dyad variances: 33.5% for leader contribution, 37.4% for member
contribution, 37.0% for gratitude, 53.3% for pride, 45.8% for anger, 44.8% for shame, 42.6% for
34
leader-directed helping, and 39.7% for risk taking. Hence, discrete exchanges and member
emotional and behavioral responses varied substantially on a daily basis.
Test of hypotheses. Presented in Table 5 are the parameter estimates of multilevel
polynomial regression analyses testing the direct effects on emotions and outcome behaviors.
Table 6 reports parameter estimates of the hypothesized main and moderating effects, and Table
7 reports those of conditional indirect effects. Results showed that the slope of the incongruence
line related to the effect on gratitude was significant (γ10 γ20 = .40, p < .001, 95%CI [.22, .57]).
We plotted the response surface of gratitude in Figure 3, which showed that member gratitude
increases as it moves along the incongruence line from low leader contribution and high member
contribution to high leader contribution and low member contribution. Nonetheless, the slope of
the incongruence line related to the effect on shame was not significant (γ10 γ20 = .10, p = .533,
95%CI [-.22, .42]). These results provide support for H1a but not for H1b. We found that the
slope of the incongruence line related to the effect on pride was significant ( γ10 + γ20 = .40, p
< .001, 95%CI [.19, .61]). Figure 4 presents the corresponding response surface, revealing that
member pride increases as it moves along the incongruence line from high leader contribution
and low member contribution to low leader contribution and high member contribution. The
slope of the incongruence line related to the effect on anger was not significant ( γ10 + γ20 = -.08,
p =.425, 95%CI [-.26, .11]). These results support H2b but not H2a.
In testing the hypothesized interactive effect on gratitude, we found that when leader-
member average contribution was high, the slope of the line parallel to the incongruence line at
the point 0.516 upward along the congruence line was positive (estimate = .53, p < .001, 95%CI
[.31, .75]). When leader-member average contribution was low, the slope of the line parallel to
the incongruence line at the point 0.516 downward along the congruence line was also positive
35
(estimate = .26, p = .016, 95%CI [.05, .47]). The difference between two slope parameters was
significant (estimate = .28, p = .028, 95%CI [.03, .52]). These results support H3a. Nevertheless,
the slope parameters for shame were not significant when leader-member average contribution
was high (estimate = .13, p = .480, 95%CI [-.23, .49]) or when it was low (estimate = .07, p
= .734, 95%CI [-.35, .50]). The difference between two slope parameters was also not significant
(estimate = .06, p = .815, 95%CI [-.41, .52]). These results do not provide support for H3b.
Results showed that when leader-member average contribution was high, the slope parameter for
pride was significantly positive (estimate = .59, p < .001, 95%CI [.34, .84]), but it was not
significant when leader-member average contribution was low (estimate = .21, p = .061, 95%CI
[-.01, .43]). The difference between two slope parameters was significant (estimate = .38, p
= .001, 95%CI [.16, .60]). These results support H4b. However, the slope parameters for anger
were not significant when leader-member average contribution was high (estimate = -.12 p
= .375, 95%CI [-.38, .14]) or when it was low (estimate = -.03, p = .731, 95%CI[-.22, .15]). The
difference between two slope parameters was also not significant (estimate = -.09, p = .516,
95%CI [-.35, .18]). These results do not support H4a.
To test H5 and H6, we initially examined the direct effects of exchange imbalance on
leader-directed helping and risk taking. Results showed that the slope of the incongruence line
related to leader-directed helping was significant (γ10 - γ20 = .17, p = .037, 95%CI [.01, .33]). As
portrayed in Figure 5, leader-directed helping appears to be increasing along the incongruence
line from low leader contribution and high member contribution to high leader contribution and
low member contribution. The slope of the incongruence line related to risk taking was
significant ( γ10 + γ20 = .24, p = .014, 95%CI [.05, .43]). As presented in Figure 6, risk taking
decreases along the incongruence line from low leader contribution and high member
36
contribution to high leader contribution and low member contribution. We then examined the
effects of respective emotions on two behavioral outcomes beyond and above those of five
polynomial regression terms. Results showed that gratitude (γ = .24, p < .001) was positively
related to leader-directed helping but not shame (γ = .006, p = .828) and that pride (γ = .17, p
= .009) was positively related to risk taking but not anger (γ = .03, p = .457).
Results of testing the moderated mediation effects showed that positive imbalance had a
significant indirect effect on leader-directed helping via gratitude when leader-member
contribution was high (estimate = .13, 95%CI [.07, .19]) and low (estimate = .06, 95%CI
[.02, .11]). The difference between these two indirect effects was significant (estimate = .07,
95%CI [.01, .12]). These results dovetail with H5a. However, the indirect effects via shame were
not significant either when leader-member contribution was high (estimate = .001, 95%CI
[-.01, .01]) or when it was low (estimate = .000, 95%CI [-.01, .01]). Neither was the difference
between the two indirect effects (estimate = .000, 95%CI [-.01, .01]). These results do not
support H5b. Further, negative imbalance had a significant indirect effect on risk taking via pride
when leader-member contribution was high (estimate = .10, 95%CI [.02, .19]), but not when it
was low (estimate = .04, 95%CI [-.0001, .09]). The difference between these two indirect effects
was significant (estimate = .06, 95%CI [.01, .14]). These results are consistent with H6b.
However, the indirect effects via anger were not significant either when leader-member
contribution was high (estimate = -.003, 95%CI [-.02, .01]) or when it was low (estimate =
- .001, 95%CI [-.01, .01]). Neither was the difference between the two indirect effects (estimate
= -.003, 95%CI [-.02, .01]). These results do not support H6a.
General Discussion
37
Drawing on the affect theory of social exchange and inequity resolution principles, we
examined the role of the self- and leader-directed emotions in leader-member discrete exchanges.
Our use of both experimental and experience sampling studies added rigor to enhancing causal
inference and capturing within-dyad fluctuations of transient emotional responses to the ebb and
flow of resource exchanges. Results of two studies corroborated our hypothesized effects related
to positive emotions but not those related to negative emotions.
Theoretical Implications
Our research first enriches the LMX literature by shifting scholarly attention from the
interplay between the general affectivity with LMX relationship quality to discrete emotions that
instantaneously arise from exchange interactions. Such a shift is of theoretical imperatives for
understanding the psychological mechanisms underpinning reiterative, ongoing discrete
exchanges. We leverage the ephemeral and fleeting nature of exchange emotions to explicate
how the waxing and waning of exchanged resources induce members’ fluctuations of unilaterally
beneficial help or exploitatively opportunistic behavior as their attempts to restore short-term
exchange balance. With LMX relationship quality controlled for in the model, results of our field
study suggest that the exchange fluctuations may move, either positively or negatively, beyond
the general, habitual reciprocation pattern. Our research thus helps to answer a long-recognized
yet understudied question of why members transitorily deviate from the primary exchange
paradigm (Ballinger & Rockmann, 2010) and enhances our knowledge of the dynamics of
discrete resource exchanges between leaders and members (Liao, Wu, Wang, & Zhang, 2019).
Our finding also moves the social exchange literature forward by empirically
disentangling the self- and the other-directed emotions as parallel mechanisms through which
exchange parties maintain exchange equity (Adams, 1965). We sought to reveal that members,
38
depending on their perceived sources of an exchange imbalance, would feel grateful toward the
leader and ashamed toward the self after a positive imbalance, and feel angry toward the leader
and proud toward the self after a negative one. That is, one exchange imbalance may
concurrently trigger multiple emotions involving different targets or distinct valence. This
represents an important attempt to understand the disparate role of the self- and other-directed
psychological pathways underlying exchange equity restoration (Adams, 1965). More
importantly, the finding on the moderating effects of leader-member average contribution
suggests that the amount of resources jointly contributed by both parties conditions the
imbalance effects on momentary emotions, shedding light on the subtle factors that might shape
the intensity of emotional experiences. Our results, when taken together, provide a thorough
illustration of how two primary theoretical lenses of behavioral research emotion and social
exchange inform each other to explicate the dynamics of leader-member exchange interactions
(Cropanzano & Mitchell, 2005; Lawler, 2001).
Intriguingly, the set of insignificant results of negative exchange emotions in our studies
appear to be seemingly opposing yet essentially complementary to the predominant theoretical
perspective that negative experiences loom larger than positive ones (Bratslavsky, Finkenauer, &
Vohs, 2001; Rozin & Royzman, 2001). Our finding suggests that both positive and negative
exchange imbalances tend to engender positive rather than negative emotions. Such asymmetric
results might primarily result from members’ self-concept enhancement motives in exchange
interactions with their leader (Adams, 1965). The joint responsibility inherent in social
exchanges between leaders and members allows two separate interpretations of the primary
causes of an outcome: one’s own over/under contribution or the other party’s under/over-
contribution. When processing exchange imbalance, members thus have the latitude to forgo the
39
self-concept threatening interpretation and take the alternative that maintains and boosts the self-
concept. This perspective implicitly echoes Rozin and Royzman’s (2001) argument that
negativity dominance only occurs when positive and negative emotional stimuli or targets are
intertwined together and are inseparable. Our finding on the salience of positive emotions
advances the social exchange literature by pointing up the credit allocation process in social
exchange. That is, to maintain exchange balance sustainably, exchange parties lean towards
paying more attention to the primary contributor in one exchange and recognize his/her deserved
credits, rather than casting blame on the one who fails to provide sufficient contributions once
(Cropanzano & Mitchell, 2005; Emerson, 1976).
Finally, our finding on risk taking adds to social exchange research regarding the
approach that dyadic parties might utilize to maintain exchange balance. We found that when
members feel underbenefited, in addition to withholding their subsequent inputs as demonstrated
in prior research (e.g., McFarlin & Sweeney, 1992), they might take exploitative actions out of
opportunism to retrieve deserved but unrelieved benefits, so that they could proactively seek
exchange balance. Such opportunistic exploitation inherently differs from the negative work
behaviors that members undertake to negatively reciprocate leaders’ unfair treatment or
destructive leadership (Cropanzano et al., 2017; Malhotra & Gino, 2011). By engaging in
opportunistic behaviors that enhance task performance, members might increase their possibility
of obtaining more resources and recognition from the leader and thus rebalance their resource
exchange (Carrell & Dittrich, 1978). By featuring the unique role of risk taking in equity
restoration, our research encourages more scholarly attention to implicit exchange behaviors.
Practical Implications
40
Our research offers important pragmatic implications. Our findings first highlight within-
dyad fluctuations of resource exchanges between managers and employees, which may deviate
from their habitual exchange pattern (Cropanzano & Mitchell, 2005). Managers may develop a
dynamic mindset about exchange interactions with employees. In addition to attending to the
general relationship quality, managers should be cognizant of the balance/imbalance of resource
exchanges on a daily basis, so that they could better forecast and manage employees’ emotions
and work behaviors. If managers want employees to be grateful and thus helpful toward
themselves, they may attempt to provide beneficial resources initially, such as offering
constructive feedback and support or delegating appropriate responsibilities (Liden et al., 1997).
The upshot of risk taking following a negative imbalance is insightful for managers to
nudge employees’ mutually beneficial behavior. When underbenefited, employees might feel
entitled to retrieve their deserved benefits by performing exploitatively opportunistic behavior
that might incur costs on managers (Adams, 1965; Wilson et al., 2010). Managers may want to
be conscious of exchanges imbalanced with employees’ extra contributions and take appropriate
actions when they occur. For example, managers could acknowledge employees’ efforts and
provide more resources in subsequent interactions. Managers may also need to attend to
employees’ emotional signals of pride so that they could use proper interventions to avoid
membersfeelings of pride developing into exploitative risk taking (Fessler et al., 2004; Isen &
Patrick, 1983). Organizations could implement leadership training programs that help managers
improve interpersonal interaction skills and thus better manage their exchanges with employees.
Our findings also speak to employees regarding how they could maintain discrete
exchanges with their managers more effectively. They may want to be aware that resource
exchanges would be imbalanced dynamically. To establish constructive and continuous
41
exchanges, when overbenefited, they should endeavor to return the favors to their managers in a
proper time frame (Cropanzano & Mitchell, 2005). When underbenefited, rather than performing
exploitative behaviors, employees could develop more sympathetic attitudes and take appropriate
opportunities to nudge managers to be more supportive, helpful, and constructive.
Limitations and Future Research
Our research has some limitations that provide tantalizing opportunities for future
research. First, although our use of both experimental and ESM studies enhanced empirical rigor,
the design might constrain the generalizability of our findings. Study 1 involved an experimental
simulation without a relational context of leader-member dyads, which might generate results
differing from field samples (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). In Study 2, due
to causal inference concern, we treated exchanges occurring in the morning as one broad
interaction unit, but leader-member dyads might experience interactions with distinct levels of
resource exchanges. We thus encourage future research to take an episodic-based design to
capture exchange dynamics (Liu, Song, Li, & Liao, 2017). Further, we tracked daily exchanges
for a relatively short time period of two weeks, raising concerns on the generalizability of our
findings to leader-member dyads with various relationship qualities. Although our control of
LMX relationship quality helps ease this concern, we encourage future research to adopt more
creative methodological approaches (e.g., using a field experiment design) to study this
fascinating phenomenon.
Second, given the daily design in Study 2, we measured afternoon risk taking with only
two items, likely restraining construct validity. Although results (see Footnote 8 for details) of
our scale validation study supported our measure, we encourage future research to measure risk
taking more thoroughly. Third, our theorizing and study design focus on the mediating role of
42
emotions on exchange imbalance and work behaviors. Although our research discusses how the
attribution process initiated by exchange imbalance evokes discrete emotions, it does not directly
hypothesize or test it (Liao, Lee, Johnson, Song, & Liu, 2021). For example, do members
consider both parties as responsibility holders for a positive imbalance, but their leaders as the
primary holder? What are the specific causes (e.g., ability, efforts) members perceive? Regarding
locus of control, if members attribute imbalances to external uncontrollable factors, how do they
react emotionally and behaviorally? Future research could find answers to these important
questions and better understand the role attribution plays.
The fourth limitation concerns our exclusive focus on members’ momentary reactions.
The resource exchanges involve leaders and thus would also impact leaders’ responses and
subsequent exchange engagement (Wilson et al., 2010). Owing to the leadership role and
responsibilities, leaders might experience different emotions. For example, when confronting
member contribution surplus, besides gratitude, leaders might experience threats as they feel
outperformed by the member (Lawler & Thye, 1999). Pride might also be evoked due to their
belief that the competence and good performance displayed by the member is a function of their
leadership (Cropanzano et al., 2017). We encourage future research to pursue this stream of
research questions and explore how leaders respond to momentary exchange imbalances.
Finally, although we focused on member work behaviors as distal outcomes, exchange
imbalances might alter members’ attachment and commitments toward exchanges with leaders.
For example, following a positive imbalance, members may have increased satisfaction,
commitment, or positive evaluations of the exchange relationship with their leader (Lawler,
2000). We thus invite future research to examine such attitudinal and relational outcomes.
Conclusion
43
Our research examines how members’ self- and leader-directed emotions arise from
discrete resource exchanges with leaders, which in turn shape their downstream work behaviors.
Our findings speak to both managers and employees regarding how they could maintain effective
work interactions and prevent unexpectedly negative deviation from their general exchange
pattern. We recognize that our work only represents an initial effort on exploring this interesting
topic and thus encourage future research to continue pursuing a more profound understanding of
the interplay between emotions and leader-member resource exchanges.
Data Availability Statement
The data that support findings of Study 1 and the syntax are available on request from the author
team. The data are not publicly available due to privacy and ethical restrictions. The raw data
support findings of Study 2 are not available due to third party restrictions. The processed data
and the syntax are available on request from the author team.
References
Adams, J. S. (1965). Inequity in social exchange. In L. Berkowitz (Ed.), Advances in
Experimental Social Psychology (pp. 267-299). New York, NY: Academic Press.
DOI:10.1016/S0065-2601(08)60108-2
Anderson, C., & Galinsky, A. D. (2006). Power, optimism, and risk‐taking. European Journal of
Social Psychology, 36, 511-536. DOI:10.1002/ejsp.324
Andreoni, J., Castillo, M., & Petrie, R. (2003). What do bargainers' preferences look like?
Experiments with a convex ultimatum game. American Economic Review, 93, 672-685.
DOI:10.1257/000282803322157034
Aquino, K., Tripp, T. M., & Bies, R. J. (2001). How employees respond to personal offense: the
effects of blame attribution, victim status, and offender status on revenge and
reconciliation in the workplace. Journal of Applied Psychology, 86, 52-59.
DOI:10.1037/0021-9010.86.1.52
Ballinger, G. & Rockman, K. (2010). Chutes versus ladders: Anchoring events and a punctuated-
equilibrium perspective on social exchange relationships. Academy of Management
Review, 35, 373-391. DOI:10.5465/amr.35.3.zok373
Barnes, C. M., Lucianetti, L., Bhave, D. P., & Christian, M. S. (2015). “You wouldn’t like me
when I’m sleepy”: Leaders’ sleep, daily abusive supervision, and work unit
engagement. Academy of Management Journal, 58, 1419-1437.
DOI:10.5465/amj.2013.1063
Bauer, T. N., & Green, S. G. (1996). Development of leader-member exchange: A longitudinal
test. Academy of Management Journal, 39, 1538-1567. DOI:10.5465/257068
Bennett, R. J., & Robinson, S. L. (2000). Development of a measure of workplace deviance.
Journal of Applied Psychology, 85, 349. DOI:10.1037/0021-9010.85.3.349
44
Bolger, N., Davis, A., & Rafaeli, E. (2003). Diary methods: Capturing life as it is lived. Annual
Review of Psychology, 54, 579-616. DOI:10.1146/annurev.psych.54.101601.145030
Bono, J. E., Foldes, H. J., Vinson, G., & Muros, J. P. (2007). Workplace emotions: The role of
supervision and leadership. Journal of Applied Psychology, 92, 1357-1367.
DOI:10.1037/0021-9010.92.5.1357
Brislin, R. W. (1980). Translation and content analysis of oral and written material. In H. C.
Triandis & J. W. Berry (Eds.), Handbook of Cross-Cultural Psychology (Vol. 2, pp. 389-
344). Boston, MA: Allyn & Bacon.
Bushman, B. J., & Baumeister, R. F. (1998). Threatened egotism, narcissism, self-esteem, and
direct and displaced aggression: Does self-love or self-hate lead to violence? Journal of
Personality and Social Psychology, 75, 219. DOI:10.1037//0022-3514.75.1.219
Campbell, W. K., Bonacci, A. M., Shelton, J., Exline, J. J., & Bushman, B. J. (2004).
Psychological entitlement: Interpersonal consequences and validation of a self-report
measure. Journal of Personality Assessment, 83, 29-45.
DOI:10.1207/s15327752jpa8301_04
Carrell, M. R., & Dittrich, J. E. (1978). Equity theory: The recent literature, methodological
considerations, and new directions. Academy of Management Review, 3, 202-210.
DOI:10.5465/amr.1978.4294844
Carver, C. S., & Harmon-Jones, E. (2009). Anger is an approach-related affect: evidence and
implications. Psychological Bulletin, 135, 183-204. DOI:10.1037/a0013965
Carver, C. S., Sinclair, S., & Johnson, S. L. (2010). Authentic and hubristic pride: Differential
relations to aspects of goal regulation, affect, and self-control. Journal of Research in
Personality, 44, 698-703. DOI:10.1016/j.jrp.2010.09.004
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd. Ed.). Lawrence
Erlbaum Associates, Inc
Cohen, J. (1992). Statistical power analysis. Current Directions in Psychological Science, 1, 98-
101. DOI:10.1111/1467-8721.ep10768783
Colquitt, J. A., Sabey, T. B., Rodell, J. B., & Hill, E. T. (2019). Content validation guidelines:
Evaluation criteria for definitional correspondence and definitional distinctiveness.
Journal of Applied Psychology, 104(10), 1243-1265. DOI:10.1037/apl0000406
Cropanzano, R., Anthony, E. L., Daniels, S. R., & Hall, A. V. (2017). Social exchange theory: A
critical review with theoretical remedies. Academy of Management Annals, 11, 479-516.
DOI:10.5465/annals.2015.0099
Cropanzano, R., & Mitchell, M. S. (2005). Social exchange theory: An interdisciplinary review.
Journal of Management, 31, 874-900. DOI:10.1177/0149206305279602
Cropanzano, R., Dasborough, M. T., & Weiss, H. M. (2017). Affective events and the
development of leader-member exchange. Academy of Management Review, 42, 233-258.
DOI:10.5465/amr.2014.0384
d’Adda, G., Darai, D., Pavanini, N., & Weber, R. A. (2017). Do leaders affect ethical conduct?
Journal of the European Economic Association, 15, 1177-1213.
DOI:10.1093/jeea/jvw027
Dalal, R. S., Lam, H., Weiss, H. M., Welch, E. R., & Hulin, C. L. (2009). A within-person
approach to work behavior and performance: Concurrent and lagged citizenship-
counterproductivity associations, and dynamic relationships with affect and overall job
performance. Academy of Management Journal, 52, 1051-1066.
DOI:10.5465/amj.2009.44636148
45
Daniels, M. A., & Robinson, S. L. (2019). The Shame of It All: A Review of Shame in
Organizational Life. Journal of Management, 45, 2448-2473.
DOI:10.1177/0149206318817604
Dewett, T. (2006). Exploring the role of risk in employee creativity. Journal of Creative
Behavior, 40, 27-45. DOI:10.1002/j.2162-6057.2006.tb01265.x
Dienesch, R. M., & Liden, R. C. (1986). Leader-member exchange model of leadership: A
critique and further development. Academy of Management Review, 11, 618-634.
DOI:10.5465/amr.1986.4306242
Douglas, S. C., & Martinko, M. J. (2001). Exploring the role of individual differences in the
prediction of workplace aggression. Journal of Applied Psychology, 86, 547-559.
DOI:10.1037/0021-9010.86.4.547
Dunn, J. R., & Schweitzer, M. E. (2005). Feeling and believing: the influence of emotion on
trust. Journal of Personality and Social Psychology, 88, 736-748. DOI:10.1037/0022-
3514.88.5.736
Dyer, N. G., Hanges, P. J., & Hall, R. J. (2005). Applying multilevel confirmatory factor analysis
techniques to the study of leadership. The Leadership Quarterly, 16, 149-167.
DOI:10.1016/j.leaqua.2004.09.009
Edwards, J. R., & Cable, D. M. (2009). The value of value congruence. Journal of Applied
Psychology, 94, 654-677. DOI:10.1037/a0014891
Edwards, J. R., & Parry, M. E. (1993). On the use of polynomial regression equations as an
alternative to difference scores in organizational research. Academy of Management
Journal, 36, 1577-1613. DOI:10.5465/256822
Eisenberger, R., Armeli, S., Rexwinkel, B., Lynch, P. D., & Rhoades, L. (2001). Reciprocation
of perceived organizational support. Journal of Applied Psychology, 86, 42-51.
DOI:10.1037/0021-9010.86.1.42
Emerson, R. M. (1976). Social exchange theory. Annual Review of Sociology, 2, 335-362.
DOI:10.1146/annurev.so.02.080176.002003
Emmons, R. A., & McCullough, M. E. (2003). Counting blessings versus burdens: An
experimental investigation of gratitude and subjective well-being in daily life. Journal of
Personality and Social Psychology, 84, 377-389. DOI:DOI:10.1037/0022-3514.84.2.377
Enders, C. K., & Tofighi, D. (2007). Centering predictor variables in cross-sectional multilevel
models: A new look at an old issue. Psychological Methods, 12, 121-138.
DOI:10.1037/1082-989X.12.2.121
Farh, C. I., Lanaj, K., & Ilies, R. (2017). Resource-based contingencies of when teammember
exchange helps member performance in teams. Academy of Management Journal, 60,
DOI:1117-1137. 10.5465/amj.2014.0261
Fehr, R., Fulmer, A., Awtrey, E., & Miller, J. A. (2017). The grateful workplace: A multilevel
model of gratitude in organizations. Academy of Management Review, 42, 361-381.
DOI:10.5465/amr.2014.0374
Fessler, D. M., Pillsworth, E. G., & Flamson, T. J. (2004). Angry men and disgusted women: An
evolutionary approach to the influence of emotions on risk taking. Organizational
Behavior and Human Decision Processes, 95, 107-123.
DOI:10.1016/j.obhdp.2004.06.006
Fitness, J. (2000). Anger in the workplace: An emotion script approach to anger episodes
between workers and their superiors, co-workers and subordinates. Journal of
46
Organizational Behavior, 21, 147-162. DOI:10.1002/(SICI)1099-
1379(200003)21:2<147::AID-JOB35>3.0.CO;2-T
Flynn, F. J. (2003). How much should I give and how often? The effects of generosity and
frequency of favor exchange on social status and productivity. Academy of Management
Journal, 46, 539-553. DOI:10.5465/30040648
Frijda, N. H. (1986). The Emotions. Cambridge, England: Cambridge University Press.
Gausel, N., Leach, C. W., Vignoles, V. L., & Brown, R. (2012). Defend or repair? Explaining
responses to in-group moral failure by disentangling feelings of shame, rejection, and
inferiority. Journal of Personality and Social Psychology, 102, 941-960.
DOI:10.1037/a0027233
Gehm, T.L., & Scherer, K.R. (1988). Relating situation evaluation to emotion differentiation:
Nonmetric analysis of cross-cultural questionnaire data. In Scherer, K.R. (Ed.), Facets of
Emotion: Recent Research (p. 6177). Lawrence Erlbaum Associates, Inc.
Gibson, D. E., & Callister, R. R. (2010). Anger in organizations: Review and integration.
Journal of Management, 36, 66-93. DOI:10.1177/0149206309348060
Gino, F., Kouchaki, M., & Galinsky, A. D. (2015). The moral virtue of authenticity: How
inauthenticity produces feelings of immorality and impurity. Psychological Science, 26.
DOI:10.1177/0956797615575277
Gouldner, A. W. (1960). The norm of reciprocity: A preliminary statement. American
Sociological Review, 25, 161-178. DOI:10.2307/2092623
Graen, G. B., & Uhl-Bien, M. (1995). Relationship-based approach to leadership: Development
of leader-member exchange (LMX) theory of leadership over 25 years: Applying a multi-
level multi-domain perspective. The Leadership Quarterly, 6, 219-247.
DOI:10.1016/1048-9843(95)90036-5
Graen, G.B., & Scandura, T. (1987). Toward a psychology of dyadic organizing. In B. Staw &
L.L. Cumming (Eds.), Research in Organizational Behavior (Vol. 9, pp. 175-208).
Greenwich, CT: JAI Press.
Guerrero, L. K., La Valley, A. G., & Farinelli, L. (2008). The experience and expression of
anger, guilt, and sadness in marriage: An equity theory explanation. Journal of Social and
Personal Relationships, 25, 699-724. DOI:10.1177/0265407508093786
Güth, W., Schmittberger, R., & Schwarze, B. (1982). An experimental analysis of ultimatum
bargaining. Journal of Economic Behavior & Organization, 3, 367-388.
DOI:10.1016/0167-2681(82)90011-7
Henderson, P. W., & Peterson, R. A. (1992). Mental accounting and categorization.
Organizational Behavior and Human Decision Processes, 51, 92-117.
DOI:10.1016/0749-5978(92)90006-S
Hofmann, D. A., Griffin, M., & Gavin, M. (2000). The application of hierarchical linear
modeling to organizational research. In K. J. Klein & S.W.J. Kozlowski (Eds.), Multilevel
Theory, Research, and Methods in Organizations. San Francisco, CA: Jossey Bass.
Hollander, E. P. (1958). Conformity, status, and idiosyncrasy credit. Psychological Review, 65.
DOI:10.1037/h0042501
Isen, A. M., & Patrick, R. (1983). The effect of positive feelings on risk taking: When the chips
are down. Organizational Behavior and Human Performance, 31, 194-202.
DOI:10.1016/0030-5073(83)90120-4
47
Jordan, J., Sivanathan, N., & Galinsky, A. D. (2011). Something to lose and nothing to gain: The
role of stress in the interactive effect of power and stability on risk taking. Administrative
Science Quarterly, 56, 530-558. DOI:10.1177/0001839212441928
Koopman, J., Lanaj, K., & Scott, B. A. (2016). Integrating the bright and dark sides of OCB: A
daily investigation of the benefits and costs of helping others. Academy of Management
Journal, 59, 414-435. DOI:10.5465/amj.2014.0262
Kornilaki, E. N., & Chlouverakis, G. (2004). The situational antecedents of pride and happiness:
Developmental and domain differences. British Journal of Developmental Psychology,
22, 605-619. DOI:10.1348/0261510042378245
Larney, A., Rotella, A., & Barclay, P. (2019). Stake size effects in ultimatum game and dictator
game offers: A meta-analysis. Organizational Behavior and Human Decision Processes,
151, 61-72. DOI:10.1016/j.obhdp.2019.01.002
Lawler, E. J. (2001). An affect theory of social exchange. American Journal of Sociology, 107,
321-352. DOI:10.1086/324071
Lawler, E. J., & Thye, S. R. (1999). Bringing emotions into social exchange theory. Annual
Review of Sociology, 25, 217-244. DOI:10.1146/annurev.soc.25.1.217
Leach, C. W., & Cidam, A. (2015). When is shame linked to constructive approach orientation?
A meta-analysis. Journal of Personality and Social Psychology, 109, 983-1002.
DOI:10.1037/pspa0000037
Lerner, J. S., & Keltner, D. (2001). Fear, anger, and risk. Journal of Personality and Social
Psychology, 81, 146-159. DOI:10.1037/0022-3514.81.1.146
Lewis, M. (1997). The self in self-conscious emotions. In J. G. Snodgrass & R. L. Thompson
(Eds.), Self across Psychology: Self-recognition, Self-awareness, and the Self-
concept (pp. 119142). New York, NY: New York Academy of Sciences.
Liao, Z., Lee, H. W., Johnson, R. E., Song, Z., & Liu, Y. (2021). Seeing from a short-term
perspective: When and why daily abusive supervisor behavior yields functional and
dysfunctional consequences. Journal of Applied Psychology, 106, 377-398.
DOI:10.1037/apl0000508Liao, Z., Liu, W., Li, X., & Song, Z. (2019). Give and take: An
episodic perspective on leader-member exchange. Journal of Applied Psychology, 104,
34-51. DOI:10.1037/apl0000371
Liao, Z., Yam, K. C., Johnson, R. E., Liu, W., & Song, Z. (2018). Cleansing my abuse: A
reparative response model of perpetrating abusive supervisor behavior. Journal of
Applied Psychology, 103, 1039-1056. DOI:10.1037/apl0000319
Liao, Z., Wu, L., Wang, Y., & Zhang, H. (2019, July). Exchanging Through Emoting: An
Emotional Model of Leader-Member Episodic Exchanges. In Academy of Management
Proceedings (Vol. 2019, No. 1, p. 14248). Briarcliff Manor, NY 10510: Academy of
Management. DOI: 10.5465/AMBPP.2019.109
Liden, R. C., Sparrowe, R. T., & Wayne, S. J. (1997). Leadermember exchange theory: The
past and potential for the future. In G. R. Ferris (Ed.), Research in Personnel and Human
Resources Management (Vol. 15, pp. 47-119). Greenwich, CT: JAI Press.
Liden, R. C., Wayne, S. J., & Stilwell, D. (1993). A longitudinal study on the early development
of leader-member exchanges. Journal of Applied Psychology, 78, 662-674.
DOI:10.1037/0021-9010.78.4.662
Little, T. D., Cunningham, W. A., Shahar, G., & Widaman, K. F. (2002). To parcel or not to
parcel: Exploring the question, weighing the merits. Structural Equation Modeling, 9,
151-173. DOI:10.1207/S15328007SEM0902_1
48
Littlepage, G., Robison, W., & Reddington, K. (1997). Effects of task experience and group
experience on group performance, member ability, and recognition of expertise.
Organizational Behavior and Human Decision Processes, 69, 133-147.
DOI:10.1006/obhd.1997.2677
Liu, W., Song, Z., Li, X., & Liao, Z. (2017). Why and when leaders’ affective states influence
employee upward voice. Academy of Management Journal, 60, 238-263. DOI:
10.5465/amj.2013.1082
Lopes, L. L. (1987). Between hope and fear: The psychology of risk. In L. Berkowitz (Ed.),
Advances in Experimental Social Psychology (pp.255-295). NY: Academic Press.
DOI:10.1016/S0065-2601(08)60416-5
Ma, L. K., Tunney, R. J., & Ferguson, E. (2017). Does gratitude enhance prosociality?: A meta-
analytic review. Psychological Bulletin, 143, 601-635. DOI:10.1037/bul0000103
Malhotra, D., & Gino, F. (2011). The pursuit of power corrupts: How investing in outside
options motivates opportunism in relationships. Administrative Science Quarterly, 56(4),
559-592. DOI:10.1177/0001839212441350
Mayer, D. M., Nurmohamed, S., Treviño, L. K., Shapiro, D. L., & Schminke, M. (2013).
Encouraging employees to report unethical conduct internally: It takes a village.
Organizational Behavior and Human Decision Processes, 121(1), 89-103.
DOI:10.1016/j.obhdp.2013.01.002
McCullough, M. E., Kilpatrick, S. D., Emmons, R. A., & Larson, D. B. (2001). Is gratitude a
moral affect? Psychological Bulletin, 127, 249-266. DOI:10.1037/0033-2909.127.2.249.
McFarlin, D. B., & Sweeney, P. D. (1992). Distributive and procedural justice as predictors of
satisfaction with personal and organizational outcomes. Academy of Management
Journal, 35, 626-637. DOI:10.5465/256489
Mikulincer, M., & Shaver, P. R. (2010). Does gratitude promote prosocial behavior? The
moderating role of attachment security. In Mikulincer, M. & Shaver, P. R. (Eds.),
Prosocial Motives, Emotions, and Behavior. Washington, DC: American Psychological
Association. DOI:10.1037/12061-014
Pillutla, M. M., & Murnighan, J. K. (1995). Being fair or appearing fair: Strategic behavior in
ultimatum bargaining. Academy of Management Journal, 38, 1408-1426.
DOI:10.5465/256863
Podsakoff, N. P., Whiting, S. W., Podsakoff, P. M., & Blume, B. D. (2009). Individual-and
organizational-level consequences of organizational citizenship behaviors: A meta-
analysis. Journal of Applied Psychology, 94, 122-141. DOI:10.1037/a0013079
Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common method
biases in behavioral research: A critical review of the literature and recommended
remedies. Journal of Applied Psychology, 88, 879-903. DOI:10.1037/0021-9010.88.5.879
Preacher, K. J., Zyphur, M. J., & Zhang, Z. (2010). A general multilevel SEM framework for
assessing multilevel mediation. Psychological Methods, 15, 209-233.
DOI:10.1037/a0020141
Roseman, I. J., Wiest, C., & Swartz, T. S. (1994). Phenomenology, behaviors, and goals
differentiate discrete emotions. Journal of Personality and Social Psychology, 67, 206.
DOI:10.1037/0022-3514.67.2.206
Schilpzand, P., Houston, L., & Cho, J. (2018). Not too tired to be proactive: Daily empowering
leadership spurs next-morning employee proactivity as moderated by nightly sleep
quality. Academy of Management Journal, 61, 2367-2387. DOI:10.5465/amj.2016.0936
49
Sears, G. J., & Hackett, R. D. (2011). The influence of role definition and affect in LMX: A
process perspective on the personalityLMX relationship. Journal of Occupational and
Organizational Psychology, 84, 544-564. DOI:10.1348/096317910X492081
Shaver, P., Schwartz, J., Kirson, D., & O’Connor, C. (1987). Emotion knowledge: Further
exploration of a prototype approach. Journal of Personality and Social Psychology, 52.
DOI:10.1037/0022-3514.52.6.1061
Song, Z., Foo, M. D., & Uy, M. A. (2008). Mood spill-over and crossover among dual-earner
couples: A cell phone event sampling study. Journal of Applied Psychology, 93, 443-452.
DOI:10.1037/0021-9010.93.2.443
Sparrowe, R. T. (2020). LMX and welfare trade-off ratios: An evolutionary perspective on
leader-member relations. The Leadership Quarterly, 31(2), 101271.
DOI:10.1016/j.leaqua.2018.11.001
Tangney, J. P. (1999). The self-conscious emotions: Shame, guilt, embarrassment and pride. In
Power, M. J., & Dalgleish, T. (Eds.). Handbook of Cognition and Emotion (pp. 541-568).
Wiley. DOI:10.1002/0470013494.ch26
Tangney, J. P., Miller, R. S., Flicker, L., & Barlow, D. H. (1996). Are shame, guilt, and
embarrassment distinct emotions? Journal of Personality and Social Psychology, 70.
DOI:10.1037/0022-3514.70.6.1256
Tepper, B. J., Dimotakis, N., Lambert, L. S., Koopman, J., Matta, F. K., Man Park, H., & Goo,
W. (2018). Examining follower responses to transformational leadership from a dynamic,
personenvironment fit perspective. Academy of Management Journal, 61(4), 1343-1368.
DOI:10.5465/amj.2014.0163
Tracy, J. L., & Robins, R. W. (2006). Appraisal antecedents of shame and guilt: Support for a
theoretical model. Personality and Social Psychology Bulletin, 32(10), 1339-1351.
DOI:10.1177/0146167206290212
Tracy, J. L., & Robins, R. W. (2007). Emerging insights into the nature and function of pride.
Current Directions in Psychological Science, 16, 147-150. DOI:10.1111/j.1467-
8721.2007.00493.x
Tsang, J. A. (2006). Gratitude and prosocial behavior: An experimental test of gratitude.
Cognition & Emotion, 20, 138-148. DOI:10.1080/02699930500172341
Tversky, A., & Kahneman, D. (1992). Advances in prospect theory: Cumulative representation
of uncertainty. Journal of Risk and Uncertainty, 5(4), 297-323.
DOI:10.1007/BF00122574
Uhl-Bien, M., & Maslyn, J. M. (2003). Reciprocity in manager-subordinate relationships:
Components, configurations, and outcomes. Journal of Management, 29, 511-532.
DOI:10.1016/S0149-2063_03_00023-0
Van Kleef, G. A., Heerdink, M. W., Cheshin, A., Stamkou, E., Wanders, F., Koning, L. F., ... &
Georgeac, O. A. (2021). No guts, no glory? How risk-taking shapes dominance, prestige,
and leadership endorsement. Journal of Applied Psychology, 106(11), 1673-1694.
DOI:10.1037/apl0000868
Walster, E., Walster, G. W., & Berscheid, E. (1978). Equity: Theory and Research. Boston:
Allyn & Bacon.
Watkins, P. C., & Bell, J. (2017). Current theories and research in the psychology of gratitude. In
M. A. Warren & S. I. Donaldson (Eds.), Scientific Advances in Positive Psychology (pp.
103-129). Santa Barbara, CA: Praeger
50
Watson, D., & Clark, L.A. (1994). Manual for the positive and negative affect schedule
(expanded form). Unpublished manuscript, University of Iowa, Iowa City, IA.
http://ir.uiowa.edu/psychology pubs/11
Weiner, B. (2014). The attribution approach to emotion and motivation: History, hypotheses,
home runs, headaches/heartaches. Emotion Review, 6(4), 353-361.
DOI:10.1177/1754073914534502
Weiss, H. M., Suckow, K., & Cropanzano, R. (1999). Effects of justice conditions on discrete
emotions. Journal of Applied Psychology, 84, 786-794. DOI:10.1037/0021-
9010.84.5.786
Williams, L. A., & DeSteno, D. (2008). Pride and perseverance: The motivational role of pride.
Journal of Personality and Social Psychology, 94, 1007-1017. DOI:10.1037/0022-
3514.94.6.1007
Wilson, K. S., Sin, H. P., & Conlon, D. E. (2010). What about the leader in leader-member
exchange? The impact of resource exchanges and substitutability on the leader. Academy
of Management Review, 35, 358-372. DOI:10.5465/amr.35.3.zok358
Zhang, Z., Zyphur, M. J., & Preacher, K. J. (2009). Testing multilevel mediation using
hierarchical linear models: Problems and solutions. Organizational Research Methods,
12(4), 695-719. DOI:10.1177/1094428108327450
51
Table 1
The Amounts of Hints Contributed by J. P. (leader) and Participants (member) Across Six Experimental Conditions in Study 1
Imbalance conditions / Average
contribution conditions
Positive imbalance
Negative imbalance
Balance
High average contributions
(averaged 5 hints)
Relatively high leader contribution
7 hints (J.P.) vs. 3 hints (participant)
Relatively high member contribution
3 hints (J.P.) vs. 7 hints (participant)
Equally high contributions
5 hints (J.P.) vs. 5 hints (participant)
Low average contributions
(averaged 2 hints)
Relatively high leader contribution
4 hints (J.P.) vs. 0 hint (participant)
Relatively high member contribution
0 hint (J.P.) vs. 4 hints (participant)
Equally low contributions
2 hints (J.P.) vs. 2 hints (participant)
Table 2
Means, Standard Deviations, and Correlations among Study 1 Variables
Variable
Mean
S.D.
1
2
3
4
5
6
7
8
9
10
1. Exchange imbalance: positive a
.33
.47
2. Exchange imbalance: negative a
.33
.47
-.50**
3. Average contribution level b
.50
.50
.00
-.03
4. Felt obligation to reciprocate
4.20
1.47
.17**
-.21**
.27**
(.89)
5. Felt entitlement
3.83
1.36
-.34**
.37**
-.23**
-.39**
(.86)
6. Gratitude
3.37
2.03
.46**
-.40**
.39**
.54**
-.44**
(.99)
7. Shame
2.05
1.55
.33**
-.24**
-.17**
.17**
-.10
.39**
(.95)
8. Anger
3.05
2.05
-.34**
.40**
-.32**
-.45**
.60**
-.51**
-.06
(.98)
9. Pride
4.87
1.24
-.50**
.45**
-.06
-.32**
.57**
-.53**
-.35**
.50**
(.84)
10. Leader-directed helping c
.45
.50
.18**
-.08
.04
.27**
-.13*
.29**
.09
-.18**
-.15*
11. Risk taking
52.93
16.51
-.18**
.23**
-.25**
-.28**
.23**
-.40**
-.21**
.32**
.30**
-.21**
Note. N = 247. a Exchange imbalance condition (dummy coding): positive imbalance (variable 1) was coded as 1, 0, and 0 and negative imbalance
(variable 2) was coded as 0, 0, and 1 for conditions of positive imbalance, balance, and negative imbalance respectively. b Average contribution
level: 1 = high and 0 = low. c Dummy variable: 1 = providing help and 0 = not providing help. Reliability coefficients are displayed in the
diagonal. *p < .05, **p < .01 (two-tailed).
52
Table 3
Regression Results for Testing Effects of Exchange Imbalance on Member Emotional and Behavioral Responses in Study 1
Variables
Gratitude
Shame
Anger
Pride
Leader-directed
helping b
Risk taking
M1
M2
M3
M4
M5
M6
M7
M8
M9
M10
M11
M12
B
SE
B
SE
B
SE
B
SE
B
SE
B
SE
B
SE
B
SE
B
SE
B
SE
B
SE
B
SE
Exchange imbalance:
positive a
1.48**
.27
.81*
.35
.93**
.23
1.09**
.32
-.83**
.29
-.64
.39
-.96**
.16
-.81**
.23
.77*
.33
.56
.35
-2.09
2.53
.28
2.59
Exchange imbalance:
negative a
-.97**
.27
-.91**
.34
-.32
.23
-.55
.31
1.32**
.29
1.44**
.38
.69**
.16
.36
.22
.28
.34
.40
.34
4.89
2.56
2.60
2.56
Average contribution:
high
1.11**
.34
-.58
.31
-1.01**
.38
-.24
.22
Positive imbalance ×
average high
1.37**
.49
-.33
.45
-.40
.55
-.31
.32
Negative imbalance ×
average high
-.02
.49
.43
.45
-.33
.55
.68*
.32
Felt obligation to
reciprocate
.38**
.10
.27*
.11
Felt entitlement
1.88*
.82
-.41
.99
Gratitude
.19*
.09
Shame
-.03
.10
Anger
1.88**
.64
Pride
2.26*
1.09
R2
.25
.42
.12
.16
.19
.28
.30
.33
.09
.11
.08
.13
ΔR2
.17
.04
.09
.03
.02
.05
Note. N = 247. a Exchange imbalance condition (dummy coding): positive imbalance (variable 1) was coded as 1, 0, and 0 and negative imbalance
(variable 2) was coded as 0, 0, and 1 for conditions of positive imbalance, balance, and negative imbalance respectively. b Dummy variable: 1 =
providing help and 0 = not providing help. R2 for leader-directed helping represents Nagelkerke R2. M = model; SE = standard error.
*p < .05, **p < .01 (two-tailed).
53
Table 4
Means, Standard Deviations, Percentages of Within-Dyad Variance, and Correlations among Variables in Study 2
Variables
Mean
Between-
dyad
S.D.
Within-
dyad
S.D.
Within-dyad
variance
/Percentage
Correlations
1
2
3
4
5
6
7
8
9
Between-dyad level
1. Age (year)
27.42
4.17
2. Gender
1.70
.46
-.18*
3. Dyadic tenure (month)
23.54
21.84
.34**
-.07
4. LMX relationship quality
3.85
.65
.18*
-.15
.14
5. Felt obligation to reciprocate
3.88
.61
.11
-.02
.08
.57**
Within-dyad level
1. Morning positive affect
4.03
.68
.47
.27**/41.5%
2. Morning negative affect
1.50
.50
.45
.23**/50.5%
-.32**
3. Leader contribution
3.23
.84
.54
.34**/33.5%
.19**
.06
4. Member contribution
3.12
.83
.57
.39**/37.4%
.19**
.02
.71**
5. Gratitude
3.97
.69
.48
.26**/37.0%
.18**
-.03
.47**
.36**
6. Shame
2.19
.96
.74
.65**/44.8%
-.06
.25**
.03
-.004
.07*
7. Anger
1.63
.65
.53
.31**/45.8%
-.05
.21**
.04
.03
-.07*
.45**
8. Pride
4.03
.61
.51
.35**/53.3%
.14**
-.04
.37**
.48**
.45**
.04
-.02
9. Leader-directed helping
3.61
.68
.46
.25**/42.6%
.06
.01
.24**
.19**
.32**
.001
-.03
.17**
10. Risk taking
2.82
.82
.55
.37**/39.7%
.17**
-.04
.31**
.38**
.19**
.05
.01
.34**
.09*
Note. Within-member level, N = 845; between-member level, N = 145; between-leader level, N = 79. Gender was coded as 1 = men, 2 = women.
Within-dyad correlations are based on within-dyad scores. The component percentage of within-dyad variance was computed as within-dyad
variance / (within-dyad variance + between-dyad variance + between-leader variance).
*p < .05, **p < .01 (two tailed).
54
Table 5
Multilevel Polynomial Regression Results for Testing Effects on Member Exchange Emotions and Behaviors in Study 2
Variables
Emotional experiences (noon)
Behavioral outcomes (afternoon)
Gratitude
Shame
Anger
Pride
Leader-directed helping
Risk taking
γ
SE
γ
SE
γ
SE
γ
SE
γ
SE
γ
SE
γ
SE
γ
SE
Control variables
Age
.002
.01
-.001
.01
-.01
.01
.000
.01
.03*
.01
.02*
.01
-.03*
.01
-.03*
.01
Gender
-.07
.11
-.08
.17
-.05
.13
.14
.10
.27**
.10
.15*
.07
-.19
.14
-.20
.15
Leader-member dyadic tenure
-.005
.003
-.002
.01
-.001
.003
-.004
.003
-.01*
.003
-.002
.002
-.003
.003
.000
.003
LMX relationship quality
.49**
.10
.04
.17
-.09
.11
.36**
.09
.36**
.12
.14*
.06
.07
.13
.07
.12
Felt obligation to reciprocate
.19
.11
.31*
.15
.11
.12
.24*
.10
-.16
.09
-.06
.06
.36*
.15
.33*
.15
Morning positive affect
.07
.04
.03
.08
-.002
.08
.07
.05
.000
.04
-.005
.04
.12**
.05
.14
.07
Morning negative affect
-.02
.04
.39**
.12
.23**
.06
-.03
.07
-.001
.03
.02
.03
-.004
.06
.03
.12
Polynomial terms
Leader contribution (LC), γ10
.40**
.05
.05
.10
.07
.06
.02
.06
.18**
.04
.09
.05
.05
.06
.04
.08
Member contribution (MC), γ20
-.001
.05
-.06
.08
-.005
.06
.42**
.06
.01
.05
.01
.05
.29**
.06
.24**
.08
LC2, γ30
.18**
.05
-.04
.16
.04
.04
-.05
.07
.09
.05
.08
.06
.02
.05
.03
.12
LC × MC, γ40
-.16*
.06
.22
.18
.02
.13
-.22
.12
-.10
.07
-.08
.09
-.17
.10
-.23
.12
MC2, γ50
-.01
.05
-.08
.08
-.02
.08
.21**
.05
.001
.05
-.03
.10
.14**
.05
.17**
.07
Mediation variables
Gratitude, β60
.24**
.04
Shame, β70
.006
.03
Anger, β80
.03
.04
Pride, β90
.17**
.07
Pseudo R2
.49
.39
.40
.53
.32
.36
AIC
5504.35
6577.70
6223.15
6019.71
6920.32
6826.14
BIC
5689.19
6762.54
6407.99
6204.54
7280.51
7195.81
Sample-Size Adjusted BIC
5565.33
6638.69
6284.14
6080.69
7039.15
6948.11
Note. Within-member level, N = 845; between-member level, N = 145; between-leader level, N = 79. SE = standard error. Table values are
unstandardized coefficients of MSEMs.
p < .10, *p < .05, **p < .01 (two tailed).
55
Table 6
Multilevel Polynomial Regression Results of Main and Moderating Effects on Emotions in Study 2
Variables & Effects
Gratitude
Shame
Estimate
SE
95% CI
Estimate
SE
95% CI
Main effects
Incongruence line slope, γ10 γ20
.40**
.09
[.22, .57]
.10
.16
[-.22, .42]
Moderating effects
High leader-member average contribution
γ10 γ20 + .730 × γ30 .730 × γ50
.53**
.11
[.31, .75]
.13
.18
[-.23, .49]
Low leader-member average contribution
γ10 γ20 .730 × γ30 + .730 × γ50
.26*
.11
[.05, .47]
.07
.22
[-.35, .50]
Estimate difference
.28*
.13
[.03, .52]
.06
.24
[-.41, .52]
Pride
Anger
Estimate
SE
95% CI
Estimate
SE
95% CI
Main effects
Incongruence line slope, γ10 + γ20
.40**
.11
[.19, .61]
-.08
.10
[-.26, .11]
Moderating effects
High leader-member average contribution
γ10 + γ20 .730 × γ30 + .730 × γ50
.59**
.13
[.34, .84]
-.12
.13
[-.38, .14]
Low leader-member average contribution
γ10 + γ20 + .730 × γ30 .730 × γ50
.21
.11
[-.01, .43]
-.03
.10
[-.22, .15]
Estimate difference
.38**
.11
[.16, .60]
-.09
.13
[-.35, .18]
Note. p < .10, *p < .05, **p < .01 (two tailed). Please see the online supplemental materials (Section E) for estimation approaches.
56
Table 7
Multilevel Polynomial Regression Results of Moderated Mediation Effects on Behavioral Outcomes in Study 2
Variables & Effects
Leader-directed helping
via gratitude
via shame
Estimate
SE
95% CI
Estimate
SE
95% CI
Conditional indirect effects
High leader-member average contribution
(γ10 γ20 + .730 × γ30 .730 × γ50) × β60 or β70
.13
.04
[.07, .19]
.001
.004
[-.01, .01]
Low leader-member average contribution
(γ10 γ20 .730 × γ30 + .730 × γ50) × β60 or β70
.06
.03
[.02, .11]
.000
.002
[-.01, .01]
Indirect effect differences
.07
.03
[.01, .12]
.000
.002
[-.01, .01]
Risk taking
via pride
via anger
Estimate
SE
95% CI
Estimate
SE
95% CI
Conditional indirect effects
High leader-member average contribution
( γ10 + γ20 .730 × γ30 + .730 × γ50) × β80 or β90
.10
.04
[.02, .19]
-.003
.01
[-.02, .01]
Low leader-member average contribution
( γ10 + γ20 + .730 × γ30 .730 × γ50) × β80 or β 90
.04
.02
[-.0001, .09]
-.001
.003
[-.01, .01]
Indirect effect differences
.06
.03
[.01, .14]
-.003
.01
[-.02, .01]
Note. p < .10, *p < .05, **p < .01 (two tailed). Please see the online supplemental materials (Section E) for estimation approaches.
57
Figure 1. An Emotional Model of Leader-Member Episodic Resource Exchanges
Figure 2. Mean Values of Emotions by Conditions and Average Contribution Levels in Study 1
Note. Error bars indicate +/- 1 standard error. Numbers above bars are the means of emotions
across conditions.
Gratitude / Shame
Leader-directed helping
Exchange imbalance
(positive vs. negative)
Pride / Anger Risk taking
Leader-member average
contribution
58
Figure 3. The Incongruence Effect on Gratitude in Study 2
Figure 4. The Incongruence Effect on Pride in Study 2
59
Figure 5. The Incongruence Effect on Leader-directed Helping in Study 2
Figure 6. The Incongruence Effect on Risk Taking in Study 2
60
Online Supplemental Materials
Section A: Study 1 Experiment Flow
Note: Explanations of the purpose of different pages are shown in [square brackets]. Page breaks are
shown with (next page) in gray color.
[Consent information]
(next page)
[Background information]
Welcome to the decision-making exercise!
During the exercise, you are selected to be part of a special highly selective marketing consulting task
force called “A-Team.” The mission of A-Team is to offer sound advice to companies on how various
product packages might “test out” with the general public.
In this exercise, you and another Prolific worker will be paired as a team and randomly assigned to
the role of “team member” or “team leader” to jointly complete this task.
(next page)
[Leadership assessment and initials]
Now please complete a short leadership assessment questionnaire. The experiment system will assign
the “leader” role based upon the assessment results of you and the other participants.
Please indicate the extent to which you agree with the following statements (from 1 = strongly disagree to
7 = strongly agree, adapted from Rubin, Munz, and Bommer’s (2005) transformational leadership scale)
I know how to inspire others with my plans for the future
I know how to foster collaboration among work groups
I am good at encouraging others to be “team players”
I know how to lead by “doing” rather than simply by “telling”
I am good at providing a good model for others to follow
Please enter your initials. This will be kept confidential and only be used in the communications between
you and your partner in the team: ____
(next page)
[Wait for role assignment]
Thanks for completing the assessments. The research system is generating the score of you and the other
participant and assigning the role of the team member and the team leader. It could take up to 2 minutes.
We appreciate your patience.
(after 15 seconds, automatically proceeds to next page; the timer was invisible to participants)
61
[Role assignment]
Based upon the leadership assessment results of you and the other participant, you have been assigned to
be the “member.” The other participant, J. P. have been assigned the role of the team leader.
Your major responsibility in this job is to work closely with your direct manager, J. P. From now on, all
task instructions will be initiated by J. P. and please follow J. P.’s instructions for the subsequent task.
At the end of the task, J. P. would evaluate your performance and decide whether and how much you may
receive a bonus for completing this task. The experimenter will grant the bonus based upon J.P.’s
recommendation.
Click “Next” to view the task.
(next page)
[General task instruction]
A-Team just received a call from Discovery Charters inquiring about a potential project. The mission of
Discovery Charters is to introduce the general public to open-ocean sailing as a recreational sport. As a
part of a marketing campaign they are hoping to launch next summer, Discovery Charters wants to design
and offer a safety course about open ocean survival. However, they want to tailor the contents of the
course to target gaps in what a layperson might know or not know about open water survival.
To facilitate the launch of Discovery Charters’ project, your leader, J. P., and you need to provide a
baseline for common knowledge about open water survival. Without consulting any outside materials, J.
P. and you need to consider a scenario and work together to rank order 10 items that could help a small
group of people stranded in mid-ocean to attract attention and aid survival until rescue arrives. The
accuracy of your ranking will affect the design of safety course and the success of Discovery Characters
project, which is an important factor shaping the overall performance of both J. P. and you. You should
try your best to provide the most appropriate ranking decision.
Before you start the task, you will receive a message from J. P. about the overall procedures of the task.
Click “Next” to view a message from your manager, J.P. You may have to wait for a short period for J.P.
to finish typing.
(move to the next page in 15 seconds; the timer was invisible to participants)
[Leader’s message]
From: J.P. [Team Leader]
To: XX [Team member] [pipe in participants’ initials]
Subject: About our task
Sent: One second ago
62
Hello. I have been assigned the role of team leader. Thank you for your participation in this task.
I have taken a quick look at the task, which I found a little bit challenging. To successfully complete the
task, we might need to get hints by answering 10 extra questions. The accuracy of our answers will
determine how many hints that you may receive.
Of course, I will also try to help you in this course. I will see if I could try to get some hints by consulting
my friends with sailing experiences and by referencing to a course that I took about survival skills. The
extent to which I could help with the task has to depend on my workload though.
As manager, I would recommend to the experimenter at the end the exercise to give you a bonus ranging
from 0 to $1, depending on your performance.
But before we work on getting hints, we can both take a few seconds to familiarize ourselves with the
items.
Good luck!
J. P.
Manager of A-Team
(next page)
[Ranking task instruction]
Now, you have about 1 min to familiarize yourselves with the requirement of the task and the items to be
ranked. Below are the instructions for the task.
You have chartered a yacht with three friends, for the holiday trip of a lifetime across the Atlantic Ocean.
Unfortunately, in mid Atlantic a fierce fire breaks out in the ship’s galley. Much of the yacht is destroyed
and is slowly sinking. Your location is unclear because vital navigational and radio equipment have been
damaged in the fire. Your best estimate is that you are many hundreds of miles from the nearest landfall.
You and your friends have salvaged a four-person rubber life craft, a box of matches, a 25 liter container
of water, and a case of army rations. In addition, you have managed to save the following 10 items,
undamaged and intact after the fire. Since you cannot take all 10 items with you on your life craft, your
task is to rank these items according to their importance in aiding you to survive, as you wait to be
rescued.
Please see the items below:
A shaving mirror
A quantity of mosquito netting
Maps of the Atlantic Ocean
A floating seat cushion
A 10 liter can of oil/petrol mixture
A small transistor radio
20 square feet of opaque plastic sheeting
A can of shark repellent
One bottle of 160 proof rum
15 feet of nylon rope
63
(next page)
Now, you may start to work on the ranking task. Throughout the task, you could try to obtain the ranking
hints by 1) seeking some help from your leader; 2) answering some questions. The order of the two
approaches will be randomly presented.
(next page)
[Example condition: Positive imbalance & High average contribution condition]
[Leader contribution manipulation]
[member seeks help from leader Round 1]
Now, to complete the task, you may seek some help from your leader, J. P. Please draft an email to
provide your initial thoughts on the items and ask J. P. for recommendations and suggestions in the space
below. Once you finish typing, click "Next" to send the email to J.P.
Email subject: Asking for your help with the ranking task
(next page)
Your email is now being sent out. Please wait a while for the response from J. P. This should take
approximately 15 seconds. Thank you for your patience.
(after 15 seconds, automatically proceeds to next page)
You just received a message from J.P. (Feel free to use your scratch paper to jot down J.P.'s
recommendations, should you choose to).
Click “Next” to view J. P.’s message.
(next page)
64
[Leader’s reply – Round 1]
From: J.P. [Team Leader]
To: XX [Team member] [pipe in participants’ initials]
Subject: RE: Asking for your help with the ranking task
Sent: One second ago
Hello, I hope you had a chance to familiarize yourself with the items. Based on the knowledge through
my network and from my course training of survival skills, I came up with three hints for you in this
round.
Items in the top half of rankings:
A 10 liter can of oil/petrol mixture Helpful for signaling; can be ignited on water using the matches
to attract attention.
Items in the bottom half of rankings:
A small transistor radio Likely out of range of any radio receiver.
A quantity of mosquito netting There are unlikely to be any mosquitoes in the middle of the ocean.
J. P.
Manager of A-Team
----------------------
From: XX [Team member] [pipe in participants’ initials]
To: J. P. [Team Leader]
Subject: Asking for your help with the ranking task
Sent time: a minute ago
[pipe in participants’ typed message in round 1 here]
(next page)
[Member seeks help from the Leader Round 2]
Now you have a chance to reply to J.P.'s message. Because J.P. has taken a survival skill course, you may
want to ask for help again to see if J.P. can recall what he learned and come up with some additional
hints.
Once you finish typing, click "Next" to send the email to J.P.
Email subject: RE: RE: Asking for your help with the ranking task
----------------------
From: J.P. [Team Manager]
To: XX [Team Member] [pipe in participants’ initials]
65
Subject: RE: Asking for your help with the ranking task
Sent: One second ago
Hello, I hope you had a chance to familiarize yourself with the items. Based on my knowledge through
my network and from my course training of survival skills, I came up with three hints for you in this
round.
Items in the top half of rankings:
A 10 liter can of oil/petrol mixture Helpful for signaling; can be ignited on water using the matches
to attract attention.
Items in the bottom half of rankings:
A small transistor radio Likely out of range of any radio receiver.
A quantity of mosquito netting There are unlikely to be any mosquitoes in the middle of the ocean.
J. P.
Manager of A-Team
----------------------
From: XX [Team member] [pipe in participants’ initials]
To: J. P. [Team Leader]
Subject: Asking for your help with the ranking task
Sent: a minute ago
[pipe in participants’ typed message in round 1 here]
(next page)
Your email is now being sent out. Please wait a while for the response from J. P. This should take
approximately 15 seconds. Thank you for your patience.
(after 15 seconds, automatically proceeds to next page)
You just received a message from J. P. (Feel free to use your scratch paper to jot down J.P.'s
recommendations, should you choose to).
Click “Next” to view J. P.’s message.
(next page)
[Leader’s reply – Round 2]
From: J.P. [Team Leader]
To: XX [Team member] [pipe in participants’ initials]
66
Subject: RE: RE: RE: Asking for your help with the ranking task
Sent: One second ago
Hello, thanks for reaching out to me again. I am happy to help. I tried to recall my knowledge related to
survival skills and here are all four additional hints I thought of.
Items in the top half of rankings:
15 feet of nylon rope Could prevent people or equipment from being washed overboard, or could be
used to build shelter.
A floating seat cushion Useful as a life preserver if someone fell overboard, but if the raft sinks,
would be unable to sustain four people.
Items in the bottom half of rankings:
Maps of the Atlantic Ocean Worthless without proper navigation equipment or landmarks to figure
out our present location.
One bottle of 160% proof rum Contains 80% alcohol, which means it can be used as an antiseptic
for any injuries, otherwise of little value. Very dangerous if drunk, as it would cause the body to
dehydrate, the opposite of what you need to survive.
J. P.
Manager of A-Team
----------------------
From: XX [Team member] [pipe in participants’ initials]
To: J.P. [Team Manager]
Subject: RE: RE: Asking for your help with the ranking task
Sent: One minute ago
[pipe in participants’ typed message in round 2 here]
(leader contribution manipulation is over here, next page)
[Member contribution manipulation]
Now you can also try to get some hints on your own by answering some extra questions.
As your manager might have communicated to you, each decently completed task will earn your team one
hint. You will have up to 60 seconds to complete each task.
Click “Next” to view your first task.
(next page)
Task 1:
Without searching online, estimate the number of dogs in U.S. and explain why.
Click “Next” to start your second task.
(next page)
Task 2:
“To understand the most important characteristics of a society, one must study its major cities.”
Write a response in which you discuss the extent to which you agree or disagree with the statement and
explain your reasoning for the position you take.
67
Click “Next” to view your performance.
(next page)
[Hints earned by members]
The grading system has evaluated the quality of your responses to the extra tasks. You have earned three
out of ten hints (feel free to use your scratch paper to jot down the hints, should you choose to). You will
be able to see all the hints you earned or provided by your manager during the ranking task.
Items in the top half of my rankings:
A shaving mirror Of all the items, the mirror is absolutely critical. It is the most powerful tool
you have for communicating your presence. In sunlight, a simple mirror can generate five to
seven million candlepower of light. The reflected sunbeam can even be seen beyond the horizon.
20 square feet of opaque plastic sheeting Could be used to collect rainwater or shelter us from
sun, wind, and waves.
Items in the bottom half of my rankings:
A can of shark repellent Sharks aren’t likely to attack, but if they do, repellent could be useful.
(next page)
[Final ranking task]
On your own, please take 2 minutes to rank-order the items below. The accuracy score of your ranking
will determine your team’s performance. Drag each item from the left side of the screen into the box on
the right, according to their importance in aiding you to survive. You should drag all 10 items into the
box, so that the item you believe to be the most important item is labeled 1, the second more important
item is labeled 2, and so forth until you have ranked all 10 items.
You may use the hints provided by your manager or earned by yourself. Scroll down the page to view
all the hints.
68
Hints provided by your manager (7):
Items in the top half of rankings:
A 10 liter can of oil/petrol mixture Helpful for signaling; can be ignited on water using the matches
to attract attention.
15 feet of nylon rope Could prevent people or equipment from being washed overboard, or could be
used to build shelter.
A floating seat cushion Useful as a life preserver if someone fell overboard, but if the raft sinks, would
be unable to sustain four people.
Items in the bottom half of rankings:
A small transistor radio Likely out of range of any radio receiver.
A quantity of mosquito netting There are unlikely to be any mosquitoes in the middle of the ocean.
Maps of the Atlantic Ocean Worthless without proper navigation equipment or landmarks to figure
out our present location.
One bottle of 160% proof rum Contains 80% alcohol, which means it can be used as an antiseptic
for any injuries, otherwise of little value. Very dangerous if drunk, as it would cause the body to
dehydrate, the opposite of what you need to survive.
Hints earned by yourself (3):
Items in the top half of my rankings:
A shaving mirror Of all the items, the mirror is absolutely critical. It is the most powerful tool you
have for communicating your presence. In sunlight, a simple mirror can generate five to seven million
candlepower of light. The reflected sunbeam can even be seen beyond the horizon.
20 square feet of opaque plastic sheeting Could be used to collect rain water or shelter us from sun,
wind, and waves.
Items in the bottom half of my rankings:
A can of shark repellent Sharks aren’t likely to attack, but if they do, repellent could be useful.
69
(next page)
[Displays the ranking score calculated based on their actual ranking on last page]
Thank you for submitting your work! Your team’s accuracy score is XX.
[After this page, participants would proceed with survey items, including manipulation check
questions, measures of exchange emotions, alternative mediators, the two behavioral tasks
measuring helping and risk taking, and the final debrief]
70
Section B: Study 1 Robustness Check of Regression Analyses with Ranking Accuracy Score as A Control Variable
Variables
Gratitude
Shame
Anger
Pride
Leader-directed
helping b
Risk taking
M1
M2
M3
M4
M5
M6
M7
M8
M9
M10
M11
M12
B
SE
B
SE
B
SE
B
SE
B
SE
B
SE
B
SE
B
SE
B
SE
B
SE
B
SE
B
SE
Exchange imbalance:
positive a
1.48**
.28
.71*
.33
.94**
.22
1.01**
.31
-.82**
.28
-.69
.39
-.96**
.16
-.82**
.23
.78*
.33
.62
.36
-2.07
2.54
.33
2.59
Exchange imbalance:
negative a
-.96**
.28
-.1.07**
.33
-.39
.22
-.66*
.31
1.25**
.28
1.37**
.39
.68**
.16
.35
.23
.23
.34
.34
.35
4.90
2.57
2.59
2.56
Average contribution:
high
1.45**
.34
-.34
.32
-.87*
.39
-.22
.23
Positive imbalance ×
average high
1.62**
.47
-.15
.44
-.30
.55
-.29
.32
Negative imbalance ×
average high
.16
.47
.56
.44
-.25
.55
.69*
.32
Felt obligation to
reciprocate
.41**
.10
.30**
.11
Felt entitlement
1.93*
.83
-.34
.99
Ranking accuracy
score
.004
.01
-.06**
.01
-.05**
.01
-.04**
.01
-.06**
.01
-.02
.02
-.01
.01
-.004
.01
-.04*
.02
-.04*
.02
.04
.13
.10
.13
Gratitude
.20*
.09
Shame
-.11
.10
Anger
1.98**
.65
Pride
2.20*
1.09
R2
.25
.47
.18
.20
.24
.29
.30
.33
.11
.13
.08
.14
ΔR2
.22
.02
.05
.03
.02
.06
Note. N = 247. a Exchange imbalance condition (dummy coding): positive imbalance (variable 1) was coded as 1, 0, and 0 and negative imbalance
(variable 2) was coded as 0, 0, and 1 for conditions of positive imbalance, balance, and negative imbalance respectively. b Dummy variable: 1 =
providing help and 0 = not providing help. R2 for leader-directed helping represents Nagelkerke R2. M = model; SE = standard error.
*p < .05, **p < .01 (two-tailed).
71
Section C: Study 1 Robustness Check of Moderated Mediation Analyses with Ranking Accuracy Score as A Control Variable
Variables & Effects
Leader-directed helping
via gratitude
via shame
Estimate
SE
95% CI
Estimate
SE
95% CI
Conditional indirect effects
High leader-member average contribution
.96*
.30
[.44, 1.61]
.01
.08
[-.16, .18]
Low leader-member average contribution
.55*
.16
[.26, .91]
.01
.12
[-.23, .22]
Indirect effect differences
.41*
.18
[.13, .80]
-.004
.06
[-.13, .15]
Risk taking
via pride
via anger
Estimate
SE
95% CI
Estimate
SE
95% CI
Conditional indirect effects
High leader-member average contribution
6.43*
2.43
[2.22, 11.62]
3.64*
1.56
[.99, 7.12]
Low leader-member average contribution
3.78*
1.25
[1.45, 6.36]
3.92*
1.20
[1.72, 6.41]
Indirect effect differences
2.65*
1.38
[.55, 5.88]
-.28
1.17
[-2.67, 1.98]
Note. *p < .05 (two tailed). Analyses were conducted using the bootstrapping procedure with 10,000 samples (Hayes, 2015), controlling for
the ranking accuracy score at both stages (mediators and dependent variables).
72
Section D: CFAs Results for Alternative Models in Studies 1 & 2
Study 1
Alternative models included a three-factor model in which items of gratitude and pride were set to load on
one factor (Δχ2 (3) = 368.23, p < .001, RMSEA = .16, CFI = .90, TLI = .87, SRMR = .13); a three-factor
model in which items of anger and shame were set to load on one factor (Δχ2 (3) = 805.81, p < .001,
RMSEA = .23, CFI = .79, TLI = .73, SRMR = .21); a three-factor model in which items of gratitude and
shame were set to load on one factor (Δχ2 (3) = 693.59, p < .001, RMSEA = .22, CFI = .82, TLI = .77,
SRMR = .15); a three-factor model in which items of pride and anger were set to load on one factor (Δχ2
(3) = 402.72, p < .001, RMSEA = .17, CFI = .89, TLI = .86, SRMR = .16); a two-factor model in which
positive emotions (gratitude and pride) and negative emotions (anger and shame) each loaded into a factor
χ2 (5) = 1157.21, p < .001, RMSEA = .27, CFI = .70, TLI = .63, SRMR = .24); and a one-factor model in
which all emotions loaded into a single factor (Δχ2 (6) = 2026.56, p < .001, RMSEA = .36, CFI = .48, TLI
= .37, SRMR = .22).
Study 2
Alternative models included a six-factor model in which indicators of leader contribution and member
contribution were set to load on a single factor (Δχ2(12) = 230.33, Satorra-Bentler scaled Δχ2 = 164.09,
p < .001, RMSEA = .05, CFI = .94, TLI = .93, SRMR (Within-dyad) = .04, SRMR (Between-dyad) = .04); a
six-factor model in which indicators of gratitude and pride were set to load on a single factor (Δχ2(12) =
1636.05, Satorra-Bentler scaled Δχ2 = 1341.17, p < .001, RMSEA = .09, CFI = .81, TLI = .76,
SRMR (Within-dyad) = .09, SRMR (Between-dyad) = .04); a six-factor model in which indicators of anger and
shame were set to load on a single factor (Δχ2(12) = 1041.35, Satorra-Bentler scaled Δχ2 = 713.30, p
< .001, RMSEA = .07, CFI = .86, TLI = .83, SRMR (Within-dyad) = .06, SRMR (Between-dyad) = .09); a
five-factor model in which indicators of leader contribution, member contribution, and gratitude were set
to load on a single factor (Δχ2 (22) = 2159.47, Satorra-Bentler scaled Δχ2 = 1125.08, p < .001, RMSEA
= .10, CFI = .76, TLI = .71, SRMR (Within-dyad) = .08, SRMR (Between-dyad) = .10); a five-factor model in
which indicators of leader contribution, member contribution, and pride were set to load on a single factor
(Δχ2 (22) = 2114.14, Satorra-Bentler scaled Δχ2 = 1361.67, p < .001, RMSEA = .10, CFI = .76, TLI
= .72, SRMR (Within-dyad) = .08, SRMR (Between-dyad) = .11); and a four-factor model in which indicators of
leader contribution, member contribution, gratitude, and pride were set to load on a single factor (Δχ2 (36)
= 3414.65, Satorra-Bentler scaled Δχ2 = 1936.68, p < .001, RMSEA = .12, CFI = .64, TLI = .59,
SRMR (Within-dyad) = .18, SRMR (Between-dyad) = .42).
73
Section E: Study 2 estimation approach using polynomial regression analysis
To control for between-dyad confounds and eliminate nonessential multicollinearity, we within-dyad
centered leader contribution and member contribution (Enders & Tofighi, 2007; Hofmann et al., 2000)
and then estimated three second-order polynomial terms. All estimates with the first- and second-order
polynomial terms were performed using the random slope approach (Preacher et al., 2010). See Figure A
for the virtual illustration of estimation procedures.
Estimating main effects. Our multilevel polynomial regression analysis first involved the estimation of
the main effects of exchange imbalance (either a positive one or a negative one) on emotions. We
estimated the following equation, in which L and M referred to leader contribution and member
contribution respectively:
Y = γ00 + γ10 L + γ20 M + γ30 L2 + γ40 L×M + γ50 M2 + e (Equation 1)
To examine the effects of positive imbalance (i.e., L M) on gratitude and shame, we estimated the slope
of the incongruence line (M + L = 0, that is, M = L) at Point O (L = M = 0). Thus, we plugged in the M
= L equation to Model 1:
Y = γ00 + (γ10 γ20) × L + (γ30 γ40 + γ50) × L2
Then, we estimated the first derivation of Y over L (L = M):
dY/dL = γ10 γ20 + (γ30 γ40 + γ50) × 2 × L
Considering L = M = 0, the slope of the incongruence line at Point O (0, 0) was equal to γ10 γ20.
Likewise, to examine the effects of negative imbalance (i.e., M L) on pride and anger, we estimated the
slope of the incongruence line (M + L = 0, that is, L = M) at Point O (L = M = 0). Thus, we plugged in
the L = M equation to Model 1:
Y = γ00 + ( γ10 + γ20) × M + (γ30 γ40 + γ50) × M2
Then, we estimated the first derivation of Y over L (L = M):
dY/dM = γ10 + γ20 + (γ30 γ40 + γ50) × 2 × M
Considering M = L = 0, the slope for the incongruence line at Point O (0, 0) was equal to γ10 + γ20.
Estimating moderating effects. To test the moderating effect of leader-member average contribution, we
examined the slope of the lines parallel to the incongruence line at the points of high and low leader-
member average contribution. We specifically identified the points 1 SD of leader-member average
contribution upward and downward Point O (0, 0) along the congruence line as the points of high and low
leader-member average contribution respectively (i.e., Points A and B in Figure A below). We first
navigated Points A and B coordination on the L-M plane. Considering:
OA2 = OF2 + FA2 = ΔM2 + ΔL2 and ΔM = ΔL,
we estimated the following equation to compute the values of ΔM = ΔL:
ΔM = ΔL = sqrt (OA2 / 2) = sqrt (+1 SD of leader-member average contribution2 / 2).
74
In Study 2, 1 SD of leader-member average contribution was 0.516. Thus, ΔM = ΔL = 0.365. Given that
L and M were within-dyad centered, we navigated:
Point A (0.365, 0.365) and Point B ( 0.365, 0.365).
Then, we followed the same procedures of testing the main effects to examine the conditional slope of the
incongruence line. We first examined the moderating effects on the relationships of positive imbalance
with gratitude and shame. When leader-member average contribution was high (Point A), the line parallel
to the incongruence line should be:
M + L = 0.516 × sqrt (2) = 0.730, and thus, M = L + 0.730
We plugged in the M = L + 0.730 equation to Model 1:
Y = γ00 + γ10 L + γ20 × ( L + 0.730) + γ30 L2 + γ40 × L × ( L + 0.730) + γ50 × ( L + 0.730)2 (Equation 2)
We then estimated the first derivation of Y over L based on Equation 2:
dY/dL = (γ10 γ20 + 0.730 × γ40 2 × 0.730 × γ50) + (γ30 γ40 + γ50) × 2 × L
Considering L = M = 0.365, the slope of the incongruence line was:
dY/dL = γ10 γ20 + 0.730 × γ30 0.730 × γ50
Using this equation, we estimated the conditional effects of positive imbalance on gratitude and shame
when leader-member average contribution was high. When leader-member average contribution was low
(Point B), the line parallel to the incongruence line should be:
M + L = 0.516 × sqrt (2) = 0.730, and thus, M = L 0.730
We plugged in the M = L 0.730 equation to Model 1:
Y = γ00 + γ10 L + γ20 × ( L 0.730) + γ30 L2 + γ40 × L × ( L 0.730) + γ50 × ( L 0.730)2 (Equation 3)
We then estimated the first derivation of Y over L based on Equation 3:
dY/dL = dY/dL = (γ10 γ20 0.730 × γ40 + 2 × 0.730 × γ50) + (γ30 γ40 + γ50) × 2 × L
Considering L = M = -0.365, the slope of the incongruence line was:
dY/dL = γ10 γ20 0.730 × γ30 + 0.730 × γ50
Using this equation, we estimated the conditional effects of positive imbalance on gratitude and shame
when leader-member average contribution was low.
Then, we examined the moderating effects on the relationships of negative imbalance with pride and
anger. When leader-member average contribution was high (Point A), When leader-member average
contribution was high (Point A), the line parallel to the incongruence line should be:
M + L = 0.516 × sqrt (2) = 0.730, and thus, L = M + 0.730
75
We plugged in the L = M + 0.730 equation to Equation 1:
Y = γ00 + γ10 × ( M + 0.730) + γ20 × M + γ30 × ( M + 0.730)2 + γ40 × ( M + 0.730) × M + γ50 M2
(Equation 4)
We then estimated the first derivation of Y over M based upon Equation 4:
dY/dM = ( γ10 + γ20 2 × 0.730 × γ30 + 0.730 × γ40) + (γ30 γ40 + γ50) × 2 × M
Considering L = M = 0.365, the slope of the incongruence line was
dY/dM = γ10 + γ20 0.730 × γ30 + 0.730 × γ50
Using this equation, we estimated the conditional effects of negative imbalance on pride and anger when
leader-member average contribution was high. When leader-member average contribution was low (Point
B), the line parallel to the incongruence line should be:
M + L = 0.516 × sqrt (2) = 0.730, and thus L = M 0.730
We plugged in the M = L 0.730 equation to Equation 1:
Y = γ00 + γ10 × ( M 0.730) + γ20 × M + γ30 × ( M 0.730)2 + γ40 × ( M 0.730) × M + γ50 M2
(Equation 5)
We then estimated the first derivation of Y over M based upon Equation 5:
dY/dM = ( γ10 + γ20 + 2 × 0.730 × γ30 0.730 × γ40) + (γ30 γ40 + γ50) × 2 × M
Considering L = M = 0.365, the slope of the incongruence line was
dY/dM = γ10 + γ20 + 0.730 × γ30 0.730 × γ50
Using this equation, we estimated the conditional effects of negative imbalance on pride and anger when
leader-member average contribution was low.
Estimating moderated mediation effects. To estimate the conditional indirect effects on two behavioral
outcomes, we estimated the following two equations separately, in which Y1 was leader-directed helping,
Y2 = risk taking, Med1 was gratitude, Med2 was shame, Med3 was anger, and Med4 was pride.
Y1 = β00 + β 10 L + β20 M + β30 L2 + β40 L×M + β50 M2 + β60 Med1+ β70 Med2 +e (Equation 6)
Y2 = β00 + β10 L + β20 M + β30 L2 + β40 L×M + β50 M2 + β80 Med3+ β90 Med4 +e (Equation 7)
We then integrated the respective estimates of the conditional main effects on emotions with those of the
effects of emotions on outcomes (in Equations 6 & 7) to compute the conditional indirect effects.
When leader-member average contribution was high, the indirect effect of positive imbalance on leader-
directed helping via gratitude or shame was (γ10 γ20 + 0.730 × γ30 0.730 × γ50) × β60 or β70. When leader-
76
member average contribution was low, the indirect effect of positive imbalance on leader-directed helping
via gratitude or shame was (γ10 γ20 0.730 × γ30 + 0.730 × γ50) × β60 or β70.
When leader-member average contribution was high, the indirect effect of negative imbalance on risk
taking via pride or anger was ( γ10 + γ20 0.730 × γ30 + 0.730 × γ50) × β80 or β90. When leader-member
average contribution was low, the indirect effect of negative imbalance on risk taking via pride or anger
was ( γ10 + γ20 + 0.730 × γ30 0.730 × γ50) × β80 or β90.
Figure A. The visual illustration of testing the hypothesized effects
Note. For the virtual illustration of the estimation approach, the minimum value of gratitude axis was extended
downward, reaching the value of -2. Leader contribution = L, Member contribution = M.
Solid Red Line PQ: Congruence Line (L = M); Dashed Red Line GH: Incongruence line (L = M). Point O (0, 0):
L = M = 0; Point A (0.365, 0.365), Point B ( 0.365, 0.365).
Given that L and M had different minimum (-2.47 vs. -2.97) and maximum (2.37 vs. 2.13) values respectively in Study
2, the congruence line PQ (L = M) in Figure A was not the 45-degree line. The starting point for the congruence line
was Point P.
... According to the affect theory of social exchange (Lawler, 2001), emotion and cognition influence how actors perceive and feel about their activities, relationships, and group relations (Lawler, 2001;Liao et al., 2022). Meanwhile, relational factors can influence the exchange process profoundly (Lawler, 2001;Liao et al., 2022). ...
... According to the affect theory of social exchange (Lawler, 2001), emotion and cognition influence how actors perceive and feel about their activities, relationships, and group relations (Lawler, 2001;Liao et al., 2022). Meanwhile, relational factors can influence the exchange process profoundly (Lawler, 2001;Liao et al., 2022). Therefore, following the affect theory of social exchange (Lawler, 2001), we explored the compensatory factors of a relational perspective. ...
... Therefore, following the affect theory of social exchange (Lawler, 2001), we explored the compensatory factors of a relational perspective. Previous research has explored the role of leader-member exchanges (LMX; Tröster & Van Quaquebeke, 2021), representing more of an instrumental exchange (Lawler, 2001;Liao et al., 2022). Guanxi is an interpersonal relationship in Chinese culture that encompasses instrumental and affective exchanges (Bian, 2017;Charoensukmongkol, 2022;Chen & Chen, 2004;Xin & Pearce, 1996). ...
Article
Full-text available
Research on abusive supervision typically considers employees as passive recipients of supervisory abuse. However, employees can respond actively. Based on the affect theory of social exchange, we proposed that employees with a good Guanxi (the closeness of a personal relationship, including tangible and emotional interactions, and perceptions of responsibility between the interacting parties) with their supervisors will have less anger and more self-blame when being abused. In comparison, those with bad Guanxi will have more anger and less self-blame. Furthermore, anger and self-blame influence employees’ turnover intention and job crafting. A three-wave paired questionnaire study of 382 employees and their colleagues supported our reasoning. We also discussed the theoretical and practical implications of these findings.
Article
Full-text available
While destructive consequences for subordinates have featured prominently in the abusive supervision literature, scholars have insinuated that supervisory abuse may temporarily yield functional results. Drawing from research on motive attribution tendencies that underlie abusive supervision and the control perspective of repetitive thought, we develop and test a multilevel theory that delineates both functional and dysfunctional subordinate responses to daily abusive supervisor behavior. We posit that when subordinates generally attribute abusive supervision to performance promotion motives, abusive supervisor behavior during the day leads to task reflexivity that night, translating into within-subordinate increases in next-day task performance. In contrast, when subordinates generally attribute abusive supervision to injury initiation motives, abusive supervisor behavior during the day instead leads to rumination that night, resulting in within-subordinate increases in next-day leader-directed deviance. Results from two experience sampling studies provide support for these predictions. By providing a more fine-grained understanding of both the adaptive and maladaptive consequences of daily abusive supervisor behavior, our research, together with prior studies, suggests that the short-lived instrumental outcomes of abusive supervisor behavior carry a substantial price, despite managers’ illusion that acting in an abusive manner could be a feasible influence tactic.
Article
Full-text available
Shame is a powerful and relevant discrete emotion in the workplace, as organizations are rife with potential to induce shame, and reactions to shame relate to important organizationally relevant outcomes. In this article, we review shame-related research from a variety of disciplines, integrating and identifying common patterns to better understand the shame process as it relates to organizational life. In doing so, we develop a framework that outlines the more internal psychological processes at the heart of the emotion of shame and situates these processes within the context of organizations. This framework highlights the organizational factors that help instigate shame in employees (i.e., unintentional shame triggers and purposeful shaming behaviors) as well as the behavioral outcomes important to organizations (i.e., prosocial, withdrawal, and aggressive behaviors) and provides insights regarding moderators that likely impact this process. In addition, we discuss future research implications as they pertain to each of the components in our framework, hoping that our article not only improves our understanding of organizational shame but encourages much-needed future research on it.
Article
Full-text available
Research on leader–member exchange (LMX) has predominantly taken a dyadic relationship perspective to understand the differences in overall exchanges across leader–member dyads, while neglecting the within-dyad exchange dynamics across a series of episodic resource transactions. Drawing from the literature on equity and reciprocity principles of social exchange, we develop and test a model of leader–member episodic resource transactions that delineates the momentary psychological mechanism and the boundary condition under which episodic resource contribution surplus generates member subsequent reciprocations. Multilevel polynomial regression analyses of 600 episodic exchange responses from 73 employees show that resource contribution surplus in an exchange episode increased state work engagement immediately following the episode and member resource contribution in the next episode by evoking member momentary sense of obligation to reciprocate. Additionally, the between-dyad LMX relationship quality attenuated these effects by reducing the likelihood to feel obligated to reciprocate owing to episodic resource contribution surplus. Our research highlights the micro dynamic transaction nature of the exchanges between leaders and members and provides insight into how leader–member dyads exchange resources in episodic interactions.
Article
Risk-taking can fuel innovation and growth, but it can also have devastating consequences for individuals and organizations. Here we examine whether risk-taking affords social-hierarchical benefits to risk-takers. Specifically, we investigate how risk-taking influences perceived dominance, prestige, and the willingness to endorse risk-takers' leadership. Integrating insights from costly signaling theory and the dominance/prestige framework of social rank, we theorized that risk-taking increases leadership endorsement to the degree that it fuels perceptions of prestige, but decreases leadership endorsement to the degree that it fuels perceptions of dominance. However, we also hypothesized that risk-induced perceptions of dominance do translate into leadership endorsement in competitive (rather than cooperative) intergroup settings. We tested these hypotheses in four studies involving different samples, methods, and operationalizations. In Study 1, participants performed an implicit association test (IAT) that revealed that people associate risk with leader positions, and safety with follower positions. Study 2 was a longitudinal field survey conducted during the September 2019 Israeli elections, which showed that voters' perceptions of politicians' risk-taking propensities prior to the elections positively predicted perceived dominance and prestige as well as voting behavior during the elections. Finally, Studies 3 and 4 demonstrated that people are willing to support risk-takers as leaders in the context of competitive (as opposed to cooperative) intergroup situations, because perceived dominance positively predicts leadership endorsement in competitive (but not cooperative) intergroup settings. We discuss implications for understanding the social dynamics of organizational rank and the perpetuation of risky behavior in organizations, politics, and society at large. (PsycInfo Database Record (c) 2021 APA, all rights reserved).
Article
Several reviews have been critical of the degree to which scales in industrial/organizational psychology and organizational behavior adequately reflect the content of their construct. One potential reason for that circumstance is a tendency for scholars to focus less on content validation than on other validation methods (e.g., establishing reliability, performing convergent, discriminant, and criterion-related validation, and examining factor structure). We provide clear evaluation criteria for 2 commonly used content validation approaches: Anderson and Gerbing (1991) and Hinkin and Tracey (1999). To create those guidelines, we gathered all new scales introduced in Journal of Applied Psychology, Academy of Management Journal, Personnel Psychology, and Organizational Behavior and Human Decision Processes from 2010 to 2016. We then subjected those 112 scales to Anderson and Gerbing's (1991) and Hinkin and Tracey's (1999) approaches using 6,240 participants from Amazon's Mechanical Turk with detailed, transparent, and replicable instructions. For both approaches, our results provide evaluation criteria for definitional correspondence-the degree to which a scale's items correspond to the construct's definition-and definitional distinctiveness-the degree to which a scale's items correspond more to the construct's definition than to the definitions of other orbiting constructs.
Article
A growing stream of theory and research in evolutionary psychology proposes that people mentally represent dyadic relationships through an internal regulatory variable, the “Welfare Tradeoff Ratio.” I introduce welfare tradeoff theory and indicate how it speaks to the problematic distinction between social exchanges and relationship quality in Leader-Member Exchange (LMX) theory and research. I then address issues related to leader-member agreement, LMX development, upward impression management, emotions in leader-member relationships and LMX differentiation from the perspective of welfare tradeoffs. Implications and consideration of measurement also are discussed.