Organizational Behavior and Human Decision Processes 141 (2017) 57–73

Contents lists available at ScienceDirect

Organizational Behavior and Human Decision Processes

journal homepage: www.elsevier.com/locate/obhdp

Lack of sleep and the development of leader-follower relationships over

time
Cristiano L. Guarana a,⇑, Christopher M. Barnes b
a
Kelley School of Business, Indiana University, 1309 E. Tenth Street, Bloomington, IN 47405, United States
b
Michael G. Foster School of Business, University of Washington, 585 Paccar Hall, Seattle, WA 98195-3226, United States

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

Article history: Drawing from the sleep and emotion regulation model, and attribution theory, we argue that sleep can
Received 4 September 2015 influence the quality of the relationship between leaders and their followers. Specifically, we examined
Revised 17 April 2017 the effects of lack of sleep on leader-follower relationship development at the beginning of their dyad
Accepted 24 April 2017
tenure. We hypothesized that the negative effects of lack of sleep on relationships are mediated by hos-
tility. Results based on 86 new dyads (first three days of their work relationship) showed support for our
hypotheses (Study 1). Results based on 40 leaders and 120 followers over three months (five waves) also
Keywords:
showed that lack of sleep influences perceptions of relationship quality via hostility for both leaders and
Sleep
Leader-follower relationship
followers (Study 2). Moreover, we found that the direct effects of follower lack of sleep affect leader per-
Leader-member exchange ceptions of relationship quality in the first month of their dyad tenure but decreasingly so over time; the
Hostility direct effects of a leader lack of sleep on follower perceptions of relationship quality did not vary based on
dyad tenure. Results revealed that individuals are not aware of the impact of their own lack of sleep on
other people’s perceptions of relationship quality, suggesting that leaders and followers may be damag-
ing their relationship without realizing it.
Ó 2017 Elsevier Inc. All rights reserved.

1. Introduction
Organizations often benefit when leaders and followers develop
high-quality relationships (Sparrowe & Liden, 1997). Meta-
analyses on leader-member exchange theory (LMX) have consis-
tently found that the quality of relationships is positively related
to organizational commitment, perceptions of justice, organiza-
tional citizenship behaviors (OCB), and job performance
(Dulebohn, Bommer, Liden, Brouer, & Ferris, 2012; Martin, Yues,
Thomas, Lee, & Epitropaki, 2015). Leaders develop unique relation-
ships with each of their subordinates over time (Graen, 1976;
Graen & Scandura, 1987; Nahrgang, Morgeson, & Ilies, 2009), and
initial interactions have a disproportionate influence on the forma-
tion of these work relationships (Graen & Cashman, 1975; Graen &
Scandura, 1987; Liden & Graen, 1980). Nevertheless, relationship
development is influenced by many contextual and relational cues
(Bartz & Lydon, 2006; Fraley & Shaver, 2000). Considering that a
person’s sleep tends to vary over time (e.g., Barnes, Schaubroeck,
⇑ Corresponding author.
E-mail addresses: cguarana@iu.edu (C.L. Guarana), chris24b@u.washington.edu
(C.M. Barnes).
Huth, & Ghumman, 2011; Barnes, Wagner, & Ghumman, 2012),
influences the manner that people interact (Cartwright & Knight,
1987; Hasler & Troxel, 2010; Strawbridge, Shema, & Roberts,
2004; Troxel, Buysse, Hall, & Matthews, 2009), and affects the
experience and suppression of negative emotions (e.g., Barnes,
2012), we focus on sleep as a potential dynamic characteristic that
influences the formation of leader and follower relationship via
negative emotions. We thus argue that sleep can influence the
quality of future exchanges (Sparrowe & Liden, 1997) via projected
hostile cues (Cropanzano, Dasborough, & Weiss, 2016; Todorov &
Porter, 2014).
We draw from the sleep and emotion regulation model (Barnes,
2012), and attribution theory (Jones & Davis, 1965; Kelley, 1973;
Ross & Fletcher, 1985) to extend LMX theory. In integrating the lit-
erature on sleep to emotions, Barnes (2012) theorized that sleep
deprivation augments the experience of negative affect and under-
mines the sleep-deprived person’s ability to suppress it. Thus,
sleep-deprived people tend to be more emotionally reactive,
especially with negative emotions (Gujar, Yoo, Hu, & Walker,
2011). We posit that sleep influences the quality of the relationship
between leaders and followers via these harmful effects on
emotions. Specifically, we argue that sleep-deprived leaders and
followers expose each other to hostility, a discrete negative and

http://dx.doi.org/10.1016/j.obhdp.2017.04.003
0749-5978/Ó 2017 Elsevier Inc. All rights reserved.
58 C.L. Guarana, C.M. Barnes / Organizational Behavior and Human Decision Processes 141 (2017) 57–73
other-oriented emotion (Major, Kaiser, & Mccoy, 2003), which then
diminishes their relationship quality.
Researchers on attribution theory posit that individuals naively
interpret events based on limited information and attribute
observed cues either to the situation or to the actor (Jones &
Davis, 1965; Kelley, 1973; see Ross & Fletcher, 1985 for a review).
Individuals, nevertheless, tend to make internal attributions to
observed cues (i.e., fundamental attribution error; see Bradbury
& Fincham, 1990 for a review). We argue that the negative effects
of a lack of sleep on leader–follower relationships should be espe-
cially important at the beginning of the relationship when leaders
and followers are getting to know each other and do not have a
broad set of observations to evaluate the stability of emotional
cues. Later in their tenure, leaders and followers can integrate a
wider set of observations when attributing the causal relationship
to the individual or situation. Leaders and followers, therefore, are
more likely to make internal attributions early in their dyadic
tenure.
Moreover, previous research highlights that sleep-deprived
people are often unaware of or unwilling to acknowledge how a
lack of sleep can negatively influence their behavior (Banks &
Dinges, 2007; Mullins, Cortina, Drake, & Dalal, 2014; Van
Dongen, Maislin, Mullington, & Dinges, 2003). It is possible, there-
fore, that the hostility cues triggered by lack of sleep are sent to the
other party without the actor’s awareness. This is potentially of
interest given that leaders and followers often do not agree in their
perceptions of relationship quality (Sin, Nahrgang, & Morgeson,
2009). Thus, sleepy followers can undermine the way that their
leaders perceive their relationships, but the followers can remain
unaware of this effect. By the same token, sleepy leaders can
unwittingly harm their followers’ perceptions of their relation-
ships. Therefore, we focus on both leader and follower evaluations
of the quality of their relationship, which is different from the gen-
eral stream of research on leader-follower relationship that is one-
sided and has focused mainly on the follower perspective
(Scandura & Schriesheim, 1994) (see Fig. 1).
We begin with a review of the literature on leader-follower
relationships. We then explain how sleep influences the develop-
ment of leader-follower relationships. Next we integrate hostility
as a causal mechanism connecting sleep quantity and perceptions
of leader-follower relationships. Finally, we expound that sleep
affects self-awareness having detrimental effects on how the other
party evaluates the quality of their relationship. We test our model
in two field studies in which we collected data from leaders and
their followers. In Study 1 we collected data from 86 dyads. In
study 2 we collected data from 40 leaders and 120 followers over
five waves at 15-day intervals.
2. A dynamic approach to the development of leader-follower
relationships
The development of leader and follower relationships is
dynamic and evolves over time (Dienesch & Liden, 1986). Leaders
Fig. 1. Conceptual model.
and followers start developing unique relationships through for-
mal and informal processes that occur when they are being assim-
ilated into new positions (Graen, 1976; Graen & Scandura, 1987;
Nahrgang et al., 2009). Uhl-Bien and Graen (1993) suggest that
leaders and followers enter the dyadic relationships as strangers,
not knowing if their effort to create good quality relationships will
be reciprocated. Given that new dyads are embedded in an uncer-
tain context, leaders and followers are more vulnerable to social
and contextual cues (Guarana & Hernandez, 2015) and update
their perceptions of the relationship quality over time as they accu-
mulate more information.
Consistent with the embedded dynamism in social interactions,
LMX theorists proposed that leader-follower relationships develop
over a series of interactions between leaders and followers
(Dienesch & Liden, 1986). Relationships develop through role tak-
ing, role making, and role routinization (Graen, 1976; Graen &
Scandura, 1987). Role taking involves one or more formal episodes
in which the leaders communicate and assign roles to their follow-
ers. Followers, then, respond and react to leaders’ formal request
providing feedback on what they are capable of. This is the most
dynamic stage in relationship formation, where leaders and follow-
ers constantly update their initial evaluations based on the saliency
and vividness of new cues (Graen & Scandura, 1987; Hamilton,
Katz, & Leirer, 1980).
In role making, leaders and followers take part in unstructured
and informal negotiations and requests, and implicit benefits for
dedication and loyalty are conferred. Therefore, this process
defines whether relationships will be based on economic or social
exchanges creating out-groups and in-groups. Relationships, in this
stage, start to become stable and less susceptible to new observ-
able cues. Whereas out-group members experience low-quality
relationship with low trust, interaction, support, and reward, in-
group members enjoy high-quality relationships with high trust,
interaction, support, and reward (Dienesch & Liden, 1986).
In role routinization, leaders and followers develop an under-
standing and expectations that result from collaborating in struc-
tured and unstructured tasks. This is characterized by perceived
obligations on each part to reciprocate, especially in high-quality
relationships founded upon social exchanges. The nature of the
relationship, then, becomes clear and typically remains stable from
this point (Graen & Scandura, 1987).
2.1. Lack of sleep, hostility, and leader-follower relationships
As noted above, the development of leader-follower relation-
ship evolves over time until reaching a predictable state of stabil-
ity. Although research on LMX to date has mainly focused on the
stable characteristics of the leader and follower sending consistent
signals to the other party (e.g., extroversion, agreeableness, and
conscientiousness) (Dulebohn et al., 2012; Nahrgang et al., 2009),
there is initial evidence that leader and follower relationship
development is likely to be influenced by dynamic and less pre-
dictable observable cues sent early in their social interactions
 C.L. Guarana, C.M. Barnes / Organizational Behavior and Human Decision Processes 141 (2017) 57–73
(e.g., Kangas, 2013). Some of these dynamic and observable cues
can be transmitted via emotions (Cropanzano et al., 2016).
Cropanzano et al. (2016) suggest that leaders and followers
react to observed emotions, which inform the development of their
relationship. They argue that ‘‘affective expressions serve as an
anchoring event that influence the course of subsequent LMX
development” (Cropanzano et al., 2016; p. 14). Indeed, emotional
expressions act as a type of affective event that provides informa-
tion to the observer about the actor, thus informing future relation-
ship development (Côté, 2005). Others have shown that expressed
emotion is relevant to perceptions of leadership. For instance,
Johnson (2009) found that leaders who expressed negative affect
influenced their followers’ evaluations of their leaders’ charisma.
Barnes, Guarana, Nauman, and Kong (2016) found that leaders
who were able to suppress their negative emotions through deep
acting were perceived as more charismatic by their followers.
Although these studies did not measure LMX, they are consistent
with our position that emotions can send observable emotional
signals that inform evaluations of relationship quality.
In pursuing theoretical precision (Ashforth, Kreiner, & Fugate,
2000; Harrison, Newman, & Roth, 2006) for the effects of emotions
on LMX, Cropanzano et al. (2016) encouraged researchers to focus
on discrete emotions. Researchers have shown that discrete emo-
tions have distinct antecedents and structures (Weiss &
Cropanzano, 1996), and predict changes in cognition, judgment,
experience, and physiology (Lench, Flores, & Bench, 2011).
Shaver, Schwartz, Kirson, and O’Conner (1987) came up with a list
of 200 discrete emotions. A theoretical challenge, therefore, is to
identify the most relevant discrete emotions. For example, in their
theoretical manuscript, Cropanzano et al. (2016) focused on happi-
ness/joy, anger, sadness, fear, and gratitude as antecedents of
leader-follower relationship quality. For this initial theoretical
extension we focus on hostility, because it is a discrete negative
and other-oriented emotion (Major, Kaiser & MacCoy, 2003) that
is easily observable by the other party (Hugenberg &
Bodenhausen, 2003; Tepper, 2000). In addition, hostility fluctuates
within individuals (Judge, Scott, & Ilies, 2006).
Several bodies of literature outside LMX provide some evidence
for the negative relationship between hostility and relationship
quality. For example, hostile clients are unable to develop positive
relationships with counselors (Gomes-Schwartz, 1978; Marziali,
Marmar, & Krupnick, 1981; Strupp, 1980). Customers placing com-
plaints evaluate interactions with hostile employees as unfair and
are less satisfied with the outcome (Maxham & Netemeyer, 2003).
Parents of siblings in hostile relationships report less love and sat-
isfaction in the spousal relationship (McGuire, McHale, &
Updegraff, 1996). In addition, hostility is the primary presenting
problem for marital therapy (O’Leary et al., 1992). All in all, hostile
partners tend to send intimidating signals (Spielberger, 1988) lead-
ing to low satisfaction levels in dyadic relationships (Huston &
Vangelisti, 1991; Roberts, 2000).
In short, the development of leader and follower relationship is
dynamic and likely influenced by discrete emotions. However, the
LMX literature is silent on the topic of sleep, which is a basic and
dynamic physiological and psychological need that has restorative
effects on the brain (Barnes, Lucianetti, Bhave, & Christian, 2015;
Hobson, 2005; Saper, Csammell, & Lu, 2005). Researchers have
found that lack of sleep can be both an acute and a chronic problem
(Barnes, 2012), but that within-individual variation exceeds the
between-individual variation in sleep (Knutson, Rathouz, Yan,
Liu, & Lauderdale, 2007). Recent work has shown that leaders’
and followers’ sleep can be an important antecedent of leadership
attributions (Barnes et al., 2016) and behaviors (Barnes et al.,
2015). Specifically, Barnes et al. (2016) showed that leaders’ sleep
influence displayed emotions, which then influenced followers’
59
evaluations of leadership charisma. Followers’ sleep also had an
indirect effect on evaluations of leadership charisma via positive
affect. In a diary design, Barnes et al. (2015) found that leaders’
sleep deprivation is positively related to abusive behaviors.
A growing body of research outside of management also indi-
cates that a lack of sleep is harmful to relationships. Strawbridge
et al. (2004) conducted a study with 405 marital couples and
showed that a partner’s sleep problems (e.g., sleep disorders and
sleep disturbances) were associated with self-ratings of marital
unhappiness even after controlling for one’s own sleep problems.
This is in line with Spanier and Thompson’s (1982) findings that
parents’ lack of sleep is negatively related to marital satisfaction.
Moreover, intrapersonal adjustment and friendship quality also
suffer from lack of sleep. Tavernier and Willoughby (2014) fol-
lowed 942 students over three years and showed that better sleep
predicted better interpersonal adjustment and friendship quality
over time.
We build on the sleep and emotion regulation model to suggest
that sleep influences the quality of leader-follower relationship via
hostility. The sleep and emotion regulation model notes that sleep
deprivation augments the experience of negative affect and under-
mines the ability for individuals to suppress it (Barnes, 2012;
Davidson, Putnam, & Larson, 2000; Gross, 1998). This summary
and integration of previous research on sleep deprivation and emo-
tion regulation highlights how sleep deprivation affects brain areas
(amygdala and prefrontal cortex) that are heavily activated when
individuals try to proactively exert control over their emotions
(Banks, Eddy, Angstadt, Nathan, & Phan, 2007). Sleep-deprived
individuals, therefore, tend to be more emotionally reactive, espe-
cially with negative emotions (Gailliot, Schmeichel, & Baumeister,
2006; Gujar et al., 2011; Muraven & Baumeister, 2000). Research-
ers also showed that sleep dynamically covaries with emotions
(Scott & Judge, 2006; Sonnentag, Binnewies, & Mojza, 2008). Build-
ing on this literature, we propose that sleep-deprived leaders (and
followers) can expose followers (and leaders) to hostility, which
affects the quality of their relationship.
In integrating hostility into the sleep literature, researchers
argue that lack of sleep contributes to loss of emotional control
and suppression of aggressive signals (Kamphuis, Meerlo,
Koolhass, & Lancel, 2012). Empirical evidence showing the rela-
tionship between lack of sleep and hostility has started to accumu-
late. For example, researchers have found that sleep-deprived
individuals signal hostility more often than individuals who are
not sleep deprived (e.g., Kamphuis et al., 2012). In two groups of
healthy men, with one group sleeping seven to eight hours per
night and the other 9.5–10.5 h per night, the group with the
shorter sleep duration scored significantly higher on hostility on
a mood scale (Taub, 1977). Adolescent offenders’ lack of sleep
was associated with increased hostility (Ireland & Culpin, 2006).
This pattern of results was replicated in very different samples,
research designs and settings. Indeed, in a review of the literature
on sleep, Durmer and Dinges (2005) found that sleep deprivation
affects the experience and expression of hostility in a variety of
contexts. In organizational settings, sleep-deprived nurses, and
junior and senior business students reported high levels of hostility
(Christian & Ellis, 2011), administrative employees experiencing
insomnia also reported high levels of hostility (Scott & Judge,
2006). Accordingly, we hypothesize:
Hypothesis 1a. Leader lack of sleep will be related to his/her
follower perceptions of low relationship quality.
Hypothesis 1b. Follower lack of sleep will be related to his/her
leader perceptions of low relationship quality.
60 C.L. Guarana, C.M. Barnes / Organizational Behavior and Human Decision Processes 141 (2017) 57–73
Hypothesis 2a. Leader hostility will mediate the relationship
between leader lack of sleep and follower perceptions of relation-
ship quality.
Hypothesis 2b. Follower hostility will mediate the relationship
between follower lack of sleep and leader perceptions of relation-
ship quality.
These hypotheses predict that lack of sleep, an individual
dynamic characteristic, influences leader-follower relationship
quality via hostility independently of the developmental stage of
the leader and follower relationship. In the next section, we inte-
grate time (i.e., dyad tenure) as an important contextual character-
istic (Mitchell & James, 2001) that can influence the strength of the
relationship between lack of sleep and relationship quality.
2.2. The impact of dyad tenure on sleep quantity and leader-follower
relationships
Leaders’ and followers’ early interactions are characterized by
attributional processes (Hamilton et al., 1980) whereby a variety
of relational and emotional cues influence the quality of their
exchanges (Cropanzano et al., 2016). Individuals try to interpret
events based on limited information and what is cognitively acces-
sible, changing the way they think and behave (Heider, 1958).
Attribution theory suggests that observers have an inherent need
to make sense of other people’s behavior by attributing the
observed cues either to the situation or to the actor (Jones &
Davis, 1965; Kelley, 1973; see Ross & Fletcher, 1985 for a review).
Jones & Nisbett (1971) posit that observers tend to attribute an
actor’s behaviors to stable personal dispositions, not to situational
requirements. This pervasive tendency to make internal attribu-
tions to observed behaviors, known as fundamental attribution
error, is also present in relationships (see Bradbury & Fincham,
1990 for a review).
In expanding the understanding of attribution theory, Weiner
(1985) proposed three causal dimensions: locus (internal versus
external), controllability, and stability. Considering our focus on
time and the development of leader and follower relationship qual-
ity and that sleep is dynamic (e.g., Barnes et al., 2011), we focus on
the dimension of stability. The stability dimension refers to
whether the cause of the behavior is stable or unstable across time
and situations.
Consistent with Martinko, Harvey, and Douglas (2007), who
suggest that new, negative, and unexpected situations can trigger
attributional processes, we posit that early in their relationship,
leaders (or followers) can attribute negative cues triggered by lack
of sleep to personal dispositions of followers (or leaders) and gen-
eralize that the follower is a bad person to relate and even interact
with. This cognitive process that leads to internal attributions, we
argue, is more pervasive at the beginning of the dyadic tenure
because leaders and followers do not have a broad set of observa-
tions to evaluate the stability of the negative cues more precisely.
Like individuals who rely on heuristics to make fairness assess-
ments when having limited information (Van den Bos, 2001), lead-
ers and followers can use shortcuts to make generalized
assumptions and react on the basis of those initial impressions.
Late in their tenure, leaders and followers have accumulated more
observable data to falsify their naïve assessment (Van den Bos,
2001) and potentially make situational attributions. Taken
together, this is consistent with the argument that in face of failure
(e.g., hostility cues), attributions to stable causes can result in low
expectations for the future (e.g., development of relationships).
Turner and West (2010) argue that one of the primary concerns
of new dyads is to increase predictability in the relationship; there-
fore, individuals rely more heavily on salient and distinct cues to
minimize uncertainty (Guarana & Hernandez, 2015). The literature
on first impressions also indicates that individuals rely on limited
and distinct information to make judgmental assessments about
the target person which influence their future interactions
(Hamilton et al., 1980). Also relevant to our reasoning, this limited
information is not encoded in a vacuum; individuals use past expe-
riences to make generalized statements about the target person
(implicit personality theory, Schneider, 1973). It seems plausible
to conclude that leaders and followers will make generalized eval-
uations of the other party character when observing limited but
salient cues.
Indeed, researchers have shown that stable individual charac-
teristics predicted relationship quality at initial interactions
(Nahrgang et al., 2009). Nahrgang et al. (2009) showed that extro-
verted leaders more successfully build relationships with their fol-
lowers because of their enhanced social skills. The authors also
demonstrated that agreeable followers initially develop high-
quality relationship with leaders because of their high levels of
cooperation and collaboration. Individuals, therefore, attribute
observed behaviors to potential signals that capture their attention
(Lassiter, Geers, Munhall, Handley, & Beers, 2001; Smith & Miller,
1979) and generalize those attributions to individuals’ general
character. We therefore hypothesize:
Hypothesis 3a. The relationship between leader lack of sleep and
follower perceptions of relationship quality will be moderated by
dyad tenure, such as the proposed relationship will be stronger at
initial interactions than at late interactions.
Hypothesis 3b. The relationship between follower lack of sleep
and leader perceptions of relationship quality will be moderated
by dyad tenure, such as the proposed relationship will be stronger
at initial interactions than at late interactions.
Finally, as we note in Hypothesis 2, hostility should mediate the
relationship between lack of sleep and perceived leader-follower
relationship quality. In the preceding paragraphs, we outlined that
dyadic tenure should influence the magnitude of the relationship
between lack of sleep and leader-follower relationship quality.
An integration of these hypotheses and accompanied line of rea-
soning serves the basis for a moderated mediation hypothesis, in
which the effects of lack of sleep on leader-follower relationship
quality via hostility is stronger at initial interaction than at late
interactions.
We argue that followers’ (or leaders’) hostility cues triggered by
lack of sleep will become highly salient to leaders, capturing their
attention and creating a general perception of followers’ disposi-
tions early in their tenure. Leaders and followers therefore can
attribute the causes of the counterparty’s hostility to internal fac-
tors and ignore or minimize the impact of external variables (i.e.,
fundamental attribution error) (Ross, 1977). Nevertheless, as lead-
ers and followers progress in their tenure, initial attributions of
hostility can change based on new relevant information. Indeed,
researchers have shown that internal/external attributions change
over time when observers accumulate new information (Burger,
1991; Truchot, Maure, & Patte, 2003). In general, actors’ hostility
cues are salient, vivid, and easily accessed by the observers
(Hugenberg & Bodenhausen, 2003; Tepper, 2000), and when
encoded, observed hostility cues influences future interactions
(Tepper, 2000). Nevertheless, as stated above, individuals’ sleep
quantity tends to vary over time, making hostility cues less stable
across situations and over time. These emotional inconsistencies
give leaders and followers new relevant observations. As such,
leaders (or followers) can be more likely to make external attribu-
tions for followers (or leaders) hostility. We propose the following
hypotheses:
 C.L. Guarana, C.M. Barnes / Organizational Behavior and Human Decision Processes 141 (2017) 57–73
Hypothesis 4a. The indirect effect of leader lack of sleep on
follower perceptions of relationship quality via leader hostility will
be stronger in initial interactions than in later interactions.
Hypothesis 4b. The indirect effect of follower lack of sleep on lea-
der perceptions of relationship quality via follower hostility will be
stronger in initial interactions than in later interactions.
Up to this point, we predicted that a lack of sleep can
influence the quality of the relationship between leaders and
followers via hostility, especially early in their relationship;
however, sleep-deprived individuals may not be aware of the
consequences of their sleep loss on the counterparty, as such
lack of sleep influences relational awareness. This is important
because the development of the relationship is an interdepen-
dent process and there may be an asymmetry, such that sleep
deprivation harms the other party’s view of the relationship,
but not the one’s own view of that relationship. In the next
section, we build on attribution theory to propose that sleepy
leaders (or followers) can unknowingly damage their work
relationships.
2.3. The effects of sleep on Leader-Follower relational awareness
Salient relational cues influence the development of the rela-
tionship between leaders and followers (Dienesch & Liden, 1986;
Nahrgang et al., 2009). Nonetheless, individuals are usually una-
ware of the impact of sleep on their social interactions and fre-
quently underestimate the emotional impact of sleep deprivation
on their relationships (Banks & Dinges, 2007). Relevant to the cur-
rent topic, sleep loss can be related to individuals’ ability to gauge
their own emotions accurately. Gordon and Chen (2014) asked par-
ticipants to identify the degree to which their partners experienced
10 different emotions. They compared self-reported to partner-
reported emotions. Considering that emotional inaccuracy may
be damaging to relationships (Bissonette, Rusbult, & Kilpatrick,
1997), the authors concluded that a lack of sleep is detrimental
to couple relationship because sleep-deprived individuals were
less able to assess their own emotions accurately. Indirect evidence
for poor emotional accuracy was presented by Killgore et al.
(2008), who showed that sleep deprivation is negatively related
to emotional intelligence. Therefore, sleep can influence the ability
to monitor one’s own and others’ feelings and emotions, affecting
the other’s party perceptions of relationship quality. This is critical
because relationship development is an interdependent process
(Linden et al., 1997).
In integrating the effects of sleep to attributions, we argue that
sleep deprivation intensifies the interpretative distance between
the actor and the observer. The literature on sleep has shown that
sleep deprivation affects individuals’ ability to accurately assess
their levels of tiredness (Akram, Ellis, Myachykov, & Barclay,
2016), sleepiness (Kaplan, Itoi, & Dmenet, 2007), and emotional
expressions (Gordon & Chen, 2014). As such, sleep-deprived indi-
viduals are less able to suppress their emotions, sending poten-
tially unintended signals to the observer. The observer, therefore,
tries to interpret the signals by making assessments about the sta-
bility of the actors’ behavior over time. Considering that the sleep-
deprived actors may be unknowingly sending hostile signals,
observers may interpret them as stable characteristics of the
actors.
We draw from these previous findings to posit the asymmetry
noted above. Given that sleep-deprived leaders (and followers)
may not be fully aware of the negative signals that are being sent
to followers (and leaders), we propose that a lack of sleep will
have a stronger effect on the other party’s perceptions of the
61
relationship quality than on the self’s perceptions of relationship
quality. We hypothesize:
Hypothesis 5a. Leader lack of sleep will have a stronger influence
on follower perceptions of the relationship quality than on the
leader perceptions of the relationship quality.
Hypothesis 5b. Follower lack of sleep will have a stronger influ-
ence on leader perceptions of the relationship quality than on
the follower perceptions of the relationship quality.
3. Overview of empirical studies
We tested our hypotheses across two large field studies in Bra-
zil. In Study 1, we collected data from 86 new dyads (leaders and
followers) in a variety of industries to test for the mono-level
effects of sleep quantity on perceptions of relationship quality
via state hostility. In Study 2, we collected data from 60 leaders
and 180 followers over three months to test for the evolving effects
of sleep quantity on perceptions of relationship quality via state
hostility. Across our studies, we consistently find that sleep quan-
tity influences perceptions of relationship quality through state
hostility.
3.1. Study 1
3.1.1. Sample and procedure
We collected data from new dyads (leaders and followers) in a
wide range of industries (e.g., banking, computer software, health
care, insurance, food and beverage, gas, and manufacturing) in Bra-
zil. After contacting the human resources department and leaders
of several companies, we identified 123 new dyads that had just
started working together (i.e., first three days of their work
relationship).
The survey was administered to Portuguese-language speak-
ers. Accordingly, we followed Brislin’s (1996) back-translation
procedure: the questionnaire was initially formulated in English,
translated into Portuguese by a researcher who is a native
Portuguese speaker, and then back-translated into English by
another individual. Three human resources managers reviewed
a pilot survey. We also administered the pilot survey to five ran-
domly selected leaders to evaluate the accuracy of the translation
(these data were eliminated from our final data collection). They
found the English and Portuguese versions of the questionnaire to
be consistent.
We sent an email inviting the 123 leaders to participate in a
short online survey. We explained that their new hires (followers)
would also fill out an online short survey and asked for their fol-
lowers’ name and contact information. One hundred and three
agreed to participate and sent information for new followers. We
limited the data collection to one hire (follower) per direct leader
to avoid nesting. If the direct leader had more than one new hire,
we randomly selected one of the followers. After getting the direct
leader’s consent, we sent out the survey to both parties of the dyad.
The surveys were sent on the second or third official day of new
hires employment and participants were told to fill out the survey
at the end of the work day. The survey included measures of sleep
quantity, perceptions of LMX, self-reports of state hostility and
anxiety, perceptions job performance and organizational citizen-
ship behaviors toward individuals (OCBI) of the other party, per-
sonality traits (agreeableness and extraversion), age, gender,
work experience before joining the firm, and tenure in the
company (leaders).
62 C.L. Guarana, C.M. Barnes / Organizational Behavior and Human Decision Processes 141 (2017) 57–73
Eighty-six new dyads completed the surveys, which resulted in
a response rate of 70% (leaders: 43% female, average tenure was
5.64 years [s.d. = 2.05], mean age was 35.17 [s.d. = 4.98]; follower:
56% female, average prior work experience was 7.97 years [s.d.
= 6.46], mean age was 29.57 [s.d. = 8.51]). We conducted t-tests
comparing the mean levels of sleep quantity, perceptions of LMX,
self-reports of state hostility between participants who completed
the survey on the second and third days of their work relationship
and found no significant differences.
3.1.2. Measures
3.1.2.1. Sleep quantity. Leaders and followers were asked to report
how much they had slept the night before completing the survey
(see Barnes et al., 2011). Participants were given options that ran-
ged from one to twelve hours with thirty-minute intervals.
3.1.2.2. Leader-member exchange (LMX). The quality of leader-
member relationship was measured using the LMX-7 scale
(Graen & Uhl-Bien, 1995). Participants were asked to evaluate
the quality of their relationship with the other party based on
the interactions they had on that particular day. Examples of items
are ‘‘Based on today’s interactions I know where I stand with my
leader (follower)” and ‘‘Based on today’s interactions I know how
satisfied my leader (followers) is with me.” Coefficient alpha was
0.91 for leaders and 0.85 for followers.
3.1.2.3. State hostility. We adopted four items from the PANAS-X
(Watson & Clark, 1994) to measure state hostility (i.e., angry, irri-
table, hostile, and scornful). These items have been used by others
(e.g., Scott & Judge, 2006; Vaidya et al., 2008; Watson & Clark,
1992). Similar to LMX we asked participants to consider that par-
ticular day when evaluating their own level of hostility. Coefficient
alpha was 0.92 for leaders and 0.86 for followers.
3.1.2.4. Control variables. We controlled for state- and trait-like
constructs. We controlled for state anxiety because it can influence
both sleep and LMX evaluations (e.g., Caldwell, Caldwell, Brown, &
Smith, 2004; Killgore, Kamimori, & Balkin, 2011). State anxiety was
measured with four items in the PANAS-X scale (Watson & Clark,
1994). The items were nervous, jittery, calm (reverse coded), and
fidgety. Leader’s coefficient alpha was 0.82. Follower’s coefficient
alpha was 0.79. We also controlled for leader and follower job per-
formance with the scale developed by Liden, Wayne, and Stilwell
(1993) to avoid halo effects of performance on relationship quality.
Nahrgang et al. (2009) found that job performance is a predictor of
LMX later in the work relationship. We measured leader and fol-
lower job performance on that particular day. The items were
‘‘My leader’s (follower’s) performance was very high today,” ‘‘My
leader (follower) was very effective today,” ‘‘My leader (follower)
performed very well today,” and ‘‘My leader’s (follower’s) overall
effectiveness was excellent today.” Leader’s coefficient alpha was
0.80 and follower’s coefficient alpha was 0.86. We also controlled
for OCBI because LMX is related to individual-targeted behavior
(Ilies, Nahrgang, & Morgeson, 2007). Moreover, Hui, Law, and
Chen (1999) found that follower negative affectivity had an indi-
rect relationship with OCB via relationship quality. OCBI was mea-
sured using four items of the OCB scale developed by Lee and Allen
(2002). Leaders reported their perceptions of followers’ OCBI and
followers reported their perceptions of leaders’ OCBI. Examples of
items are ‘‘My leader (follower) willingly gave his/her time to help
others who have work-related problems today” and ‘‘Today, my
leader (follower) showed genuine concern and courtesy toward
coworkers, even under the most trying business or personal situa-
tions.” Coefficient alpha was 0.86 to 0.88 for leaders and 0.84 for
follower. We controlled for extroversion and agreeableness based
on the findings by Nahrgang et al. (2009). Extraversion and agree-
ableness were measured using 20 items (10 for each trait) taken
from the revised NEO personality inventory for each trait (NEO-
PI-R; Costa & McCrae, 1992). The coefficient alpha reliability for
extraversion was 0.81 and 0.93 for leaders and followers, respec-
tively. Coefficient alpha for agreeableness was 0.83 for leaders
and 0.95 for followers. We controlled for leader and follower age,
gender, work experience prior joining the company, and tenure
in the company (leaders only). Moreover, to minimize concerns
about interdependencies, we controlled for the other party LMX
evaluations.
3.1.3. Results
Means, standard deviations, and correlations for the variables of
interest appear in Table 1. Consistent with prior work, we found
that leaders and followers do not always agree on their perceptions
of relationship quality (r = 0.14). We also found that hostility is
negatively related to relationship quality. Sleep quantity was not
related to self-evaluations of relationship quality, but was related
to the other party perceptions of relationship quality.
We conducted ordinary least square (OLS) regression analysis to
examine the relationship between leader sleep quantity and fol-
lower perceptions of relationship quality (Hypothesis 1a) and the
relationship between follower sleep quantity and leader percep-
tions of relationship quality (Hypothesis 1b). Models 1 and 3
(Table 2) included all the control variables and showed that follow-
ers’ ratings of leader performance (B = 0.27, p < 0.01) and followers
rating of leader OCBI (B = 0.23, p < 0.05) were related to follower
rating of relationship quality, and that leader ratings of follower
performance (B = 0.33, p < 0.01) was related to leader ratings of
relationship quality. Consistent with Nahrgang et al. (2009), we
also found that leader agreeableness (B = 0.39, p < 0.05) was related
to follower ratings of relationship quality, and follower extrover-
sion (B = 0.33, p < 0.05) was related to leader ratings of relationship
quality. We tested for the main effect of sleep quantity on
perceptions of relationship quality on Models 2 and 4. After con-
trolling for the same set of variables, we found that leader sleep
quantity (B = 0.003, p < 0.05) was positively related to follower rat-
ings of relationship quality, and that follower sleep quantity
(B = 0.008, p < 0.01) was positively related to leader ratings of
relationship quality. Therefore, we found support for Hypotheses
1a and 1b.
Next we sought to provide evidence for our mediation hypothe-
ses suggesting that sleep quantity influences perceptions of rela-
tionship quality via hostility (Hypotheses 2a and 2b). To this end,
we conducted mediation analysis using PROCESS (Hayes, 2013).
Results showed that the path from leader sleep quantity to leader
hostility was significant (B = -0.005, p < 0.05), as was the path
between leader hostility and follower ratings of relationship qual-
ity (B = -0.21, p < 0.05). Follower sleep quantity (B = -0.008,
p < 0.01) predicted follower hostility and follower hostility (B = -
0.41, p < 0.01) was significantly related to leader ratings of rela-
tionship quality. Bootstrapping procedures using 1000 resamples
revealed significant indirect effects of leader sleep quantity on fol-
lower ratings of relationship quality through leader hostility (indi-
rect effect = 0.001; 95% CI [0.001, 0.003]) and of follower sleep
quantity on leader ratings of relationship quality through follower
hostility (indirect effect = 0.003; 95% CI [0.001, 0.007]). See Table 3
for the results. Because all variables were measured at one point in
time, we did not test for Hypotheses 3 and 4.
To test for Hypotheses 5a and 5b which suggest that lack of
sleep will have a stronger influence on the other party perceptions
of the relationship quality than on self-perceptions of the relation-
ship quality we adopted beta difference tests. Leader sleep
Table 1
Descriptive statistics and correlations among study 1 variables.a

C.L. Guarana, C.M. Barnes / Organizational Behavior and Human Decision Processes 141 (2017) 57–73
Mean SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Follower sleep quantity 383.02 35.25
Follower report LMX 3.66 0.50 0.08 (0.85)
Follower hostility 1.80 0.55 0.59** 0.11 (0.86)
Follower anxiety 1.90 0.53 0.09 0.01 0.01 (.79)
Follower report of leader 3.43 0.61 0.06 0.42** 0.13 0.06 (0.86)
performance
* *
Follower report of leader 3.37 0.60 0.02 0.22 0.06 0.26 0.01 (0.84)
OCBI
Follower agreeableness 3.71 0.70 0.00 .12 0.05 0.13 00.09 .11 (0.95)
Follower extraversion 3.61 0.58 0.01 0.07 0.15 0.07 0.09 0.03 0.13 (0.93)
Follower age 29.57 8.51 0.03 0.10 0.08 0.10 0.10 0.08 0.03 0.27*
Follower work 7.97 6.46 0.01 0.12 0.04 0.02 0.09 0.12 0.12 0.23* 0.88**
experience
Follower gender 1.56 0.50 0.14 0.08 0.19 0.02 0.17 0.03 0.03 0.04 0.16 0.21
Leader sleep quantity 397.56 41.07 0.17 0.51** 0.19 0.09 0.36** 0.37** 0.17 0.14 0.16 0.08 0.03
Leader report LMX 3.39 0.67 0.52** 0.14 0.58** 0.03 0.14 0.04 0.01 0.23 *
0.00 0.03 0.18 0.12 (0.91)
Leader hostility 1.73 0.65 0.02 0.54** 0.06 0.09 0.27* 0.24* 0.16 0.25* 0.05 0.07 0.12 0.45** 0.09 (0.92)
Leader anxiety 1.71 0.34 0.07 0.13 0.01 0.05 0.14 0.02 0.07 0.07 0.05 0.12 0.01 0.09 0.10 0.11 (.82)
Leader report of follower 3.37 0.52 0.25* 0.03 0.37** 0.00 0.16 0.02 0.05 0.17 0.16 0.10 0.06 0.04 0.32** 0.00 0.03 (.80)
performance
*
Leader report of follower 2.37 0.53 0.06 0.20 0.07 0.08 0.14 0.13 0.05 0.07 0.06 0.11 0.04 0.22 0.04 0.04 0.05 0.16 (0.86)
OCBI
** **
Leader agreeableness 4.03 0.46 0.10 0.35 0.11 0.01 0.17 0.02 0.07 0.13 0.10 .17 0.00 0.21 0.11 0.43 0.07 0.03 0.02 (0.83)
Leader extraversion 3.84 0.36 0.09 0.02 0.00 0.11 0.08 0.03 0.02 0.20 0.03 0.06 0.11 0.02 0.03 0.26* 0.13 0.01 0.13 0.36** (.81)
Leader age 35.17 4.98 0.03 0.06 0.08 0.04 0.17 0.03 0.05 0.09 0.12 0.04 0.09 0.06 0.03 0.13 0.03 0.06 0.10 0.01 0.12
Leader work experience 6.79 4.71 0.13 0.04 0.10 0.05 0.35** 0.02 0.06 0.05 0.01 0.10 0.04 0.09 0.00 0.04 0.13 0.14 0.01 0.02 0.20 0.64**
Leader tenure 5.64 2.05 0.00 0.13 0.12 0.11 0.12 0.28** 0.08 0.03 0.04 0.06 0.13 0.01 0.00 0.12 0.03 0.07 0.08 0.10 0.05 0.22* 0.06
Leader gender 1.43 0.50 0.11 0.06 0.14 0.08 0.01 0.28** 0.02 0.18 0.10 0.06 0.02 0.22* 0.02 0.01 0.06 0.05 0.12 0.09 0.09 0.07 0.06 0.26*

Note. Reliabilities are in parentheses on the diagonal.

a
N = 86.
*
p < .05.
**
p< .01.

63
64 C.L. Guarana, C.M. Barnes / Organizational Behavior and Human Decision Processes 141 (2017) 57–73

Table 2
Regression results for main effects in study 1.a

Follower ratings of relationship quality Leader ratings of relationship quality

Model 1 - controls Model 2 - main effect Model 3 - controls Model 4 - main effect
B SE B SE B SE B SE
Intercept 1.28 1.48 0.48 1.47 0.78 2.16 2.50 2.00
Follower anxiety 0.03 0.10 0.02 0.10 0.12 0.16 0.03 0.15
Leader anxiety 0.24 0.15 0.21 0.14 0.36 0.23 0.30 0.21
Follower ratings of leader performance 0.27** 0.09 0.21* 0.10 0.07 0.15 0.06 0.14
Leader ratings of follower performance 0.05 0.11 0.03 0.10 0.33** 0.15 0.20 0.14
Follower ratings of leader OCBI 0.23* 0.09 0.13 0.10 0.20 0.16 0.14 0.14
Leader ratings of follower OCBI 0.18 0.10 0.14 0.10 0.14 0.15 0.08 0.14
Follower gender 0.08 0.10 0.06 0.10 0.26 0.15 0.17 0.14
Leader gender 0.11 0.11 0.13 0.10 0.16 0.16 0.10 0.15
Follower age 0.01 0.02 0.01 0.01 0.02 0.02 -0.01 0.02
Leader age 0.01 0.01 0.01 0.01 0.01 0.02 0.01 0.02
Follower work experience 0.03 0.02 0.03 0.02 0.03 0.03 0.02 0.03
Leader work experience 0.01 0.02 0.01 0.02 0.02 0.02 0.01 0.02
Leader tenure 0.01 0.03 0.02 0.03 0.02 0.05 0.01 0.04
Follower extroversion 0.02 0.10 0.01 0.09 0.33** 0.14 0.32* 0.13
Leader extroversion 0.14 0.15 0.11 0.15 0.01 0.23 0.07 0.21
Follower agreeableness 0.04 0.08 0.01 0.07 0.02 0.11 0.03 0.10
Leader agreeableness 0.39** 0.12 0.33** 0.12 0.01 0.20 0.04 0.18
Follower hostility 0.01 0.12 0.03 0.12
Leader hostility 0.08 0.16 0.12 0.14
Follower ratings of relationship quality 0.15 0.20 0.14 0.18
Leader ratings of relationship quality 0.11 0.10 0.10 0.09
Follower sleep quantity 0.00 0.00 0.01 0.01 0.008** 0.002
Leader sleep quantity 0.003* 0.001 0.02 0.002 0.001 0.002
R2 0.43 0.48 0.28 0.42

Note. Unstandardized regression coefficients are reported.

a
N = 86.
*
p < 0.05.
**
p < 0.01.

Table 3
Regression results for mediation in study 1.a

Leader hostility Follower ratings of Follower hostility Leader ratings of relationship

relationship quality quality
B SE B SE B SE B SE
Intercept 6.75 1.90 1.88 0.156 6.89** 1.54 0.35 2.19
Follower anxiety 0.09 0.13 0.01 0.10 0.04 0.11 0.04 0.14
Leader anxiety 0.23 0.19 0.16 0.14 0.01 0.15 0.30 0.20
Follower ratings of leader performance 0.08 0.12 0.19* 0.09 0.05 0.10 0.03 0.13
Leader ratings of follower performance 0.01 0.13 0.03 0.10 0.20 0.11 0.11 0.14
Follower ratings of leader OCBI 0.17 0.13 0.09 0.10 0.02 0.11 0.14 0.14
Leader ratings of follower OCBI 0.06 0.12 0.12 0.09 0.07 0.11 0.05 0.13
Follower gender 0.09 0.13 0.07 0.10 0.11 0.11 0.12 0.13
Leader gender 0.26* 0.13 0.07 0.10 0.09 0.12 0.06 0.14
Follower age 0.02 0.02 0.01 0.01 0.01 0.02 0.01 0.02
Leader age 0.03 0.02 0.01 0.01 0.01 0.02 0.01 0.02
Follower work experience 0.03 0.03 0.02 0.02 0.01 0.02 0.02 0.03
Leader work experience 0.02 0.02 0.01 0.01 0.01 0.02 0.01 0.02
Leader tenure 0.01 0.04 0.02 0.03 0.03 0.03 0.02 0.04
Follower extroversion 0.21 0.12 0.05 0.09 0.18 0.10 0.24 0.13
Leader extroversion 0.25 0.19 0.16 0.14 0.14 0.16 0.01 0.20
Follower agreeableness 0.06 0.09 0.01 0.07 0.09 0.08 0.01 0.10
Leader agreeableness 0.43** 0.15 0.25* 0.12 0.02 0.14 0.03 0.17
Follower hostility 0.02 0.16 0.02 0.12 0.41** 0.15
Leader hostility 0.21* 0.09 0.02 0.10 0.11 0.14
Follower ratings of relationship quality 0.01 0.14 0.13 0.17
Leader ratings of relationship quality 0.13 0.12 0.07 0.09
Follower sleep quantity 0.01 0.01 0.01 0.01 0.008** 0.002 0.005 0.002*
Leader sleep quantity 0.005* 0.002 0.002 0.001 0.01 0.01 0.01 0.01
R2 0.49 0.52 0.49 0.46
Bootstrapping effect SE 95% CI (LL, UL) Bootstrapping effect SE 95% CI (LL, UL)
Indirect effect 0.001 0.001 0.001, 0.003 0.003 0.002 0.001, 0.007

Note.; Unstandardized regression coefficients are reported; bootstrap sample size = 1000.
a
N = 86.
*
p < 0.05.
**
p < 0.01.
 C.L. Guarana, C.M. Barnes / Organizational Behavior and Human Decision Processes 141 (2017) 57–73
quantity had a stronger effect on follower ratings of the relation-
ship quality (z = 1.96, p = 0.05) and follower sleep quantity had a
stronger effect on leader perceptions of relationship quality than
on follower perceptions (z = 3.12, p = 0.01). Hypotheses 5a and 5b
were supported.
These findings suggest that sleep quantity influences percep-
tions of relationship quality through hostility, and that lack of sleep
influences the other party perceptions of relationship quality more
so than self-perceptions of relationship quality. Next, we
conducted Study 2 to adopt a within-individual design and collect-
ing data from leaders and followers through 3 months. This allows
us to not only test our hypotheses in a different sample, but also to
test the hypotheses that include changes across dyad tenure
(Hypotheses 3 and 4).
3.2. Study 2
3.2.1. Sample and procedure
The sample was comprised of new dyads (leaders and follow-
ers) at a large legal service firm in Brazil. The CEO and the human
resource director emphasized that their aggressive strategy would
require the rapid hiring of new employees (the company had hired
more than 200 employees in the prior fiscal year). The firm had
more than 4000 employees and 1000 lawyers in all of Brazil’s main
cities with clients in industries ranging from baking to retail. After
a few meetings with executives and employees’ representatives, it
was clear that leaders and followers worked long hours and inter-
acted frequently. We were given access to new hires and their
respective direct leaders. To avoid leader fatigue, we were granted
survey access to three random new hires for each leader.
Consistent with Study 1, we followed Brislin’s (1996) back-
translation procedure and also sent a pilot questionnaire to the
company’s human resources manager. In addition, we adminis-
tered the pilot questionnaire to five randomly selected employees
to evaluate the accuracy of the translation. They found no discrep-
ancies between the questionnaires.
The online data collection effort took place over a year and con-
sisted of five phases. Immediately after hiring new employees, the
human resource manager sent the researchers their names, lead-
ers, and contact information. We sent the first questionnaire to fol-
lowers and leaders in the first official day of their employment.
Phase 1, in which we administered to 78 leaders and 234 followers,
included measures of personality traits (agreeableness and
extraversion), age, gender, work experience before joining the firm,
and tenure in the company (leaders). Of the leaders and followers
solicited, 76 leaders and 215 followers filled out the question-
naires. Fifteen days after our first data collection, we gathered data
for Phase 2. The survey comprised measures of sleep quantity, per-
ceptions of LMX, self-reports of state hostility and anxiety, and per-
ceptions job performance of the other party. Phases 3, 4, and 5
consisted of similar measures and were collected 15 days apart.
We obtained 600 matched surveys (40 leaders and 120 follow-
ers), which converts to a response rate of 51.28% (leaders: 47.5%
female, average tenure was 5.50 years [s.d. = 1.91], mean age was
33.30 [s.d. = 3.70]; follower: 56.7% female, average prior work
experience was 8.45 years [s.d. = 6.67], mean age was 29.94 [s.d.
= 8.89]). Thus, there were 120 newly formed leader-follower dyads.
Independent sample t-tests comparing the mean levels of gender
and age between participants (leaders and followers) who com-
pleted all phases and those who did not revealed no significant dif-
ferences. Moreover, we found no significant effect of the starting
date (quarter) on the main variables.
3.2.2. Measures
3.2.2.1. Sleep quantity. Consistent with Study 1, we asked leaders
and followers to report their sleep quantity (see Barnes et al.,
65
2011). However, we told them to consider a typical night in the last
15 days. Participants were also given options that ranged from one
to twelve hours with thirty-minute intervals.1
3.2.2.2. Leader-member exchange. The quality of leader-member
relationship was measured similar to Study 1. We adopted the
LMX-7 scale (Graen & Uhl-Bien, 1995). Both leaders and members
rated the quality of the relationship at waves 2, 3, 4, and 5. Consis-
tent with the sleep quantity measure, participants were asked to
consider the last 15 days when evaluating their leader-member
relationship quality. Example or items are ‘‘I usually know where
I stand with my leader (follower)” and ‘‘I usually know how satis-
fied my leader (follower) is with me.” Coefficient alpha ranged
from 0.94 to 0.97 for leaders and 0.92 to 0.93 for followers.
3.2.2.3. State hostility. Consistent with Study 1, we adopted four
items from the PANAS-X (Watson & Clark, 1994) to measure state
hostility. Similar to sleep quantity and LMX we asked participants
to consider the last 15 days when evaluating their own hostility
level. Coefficient alpha ranged from 0.75 to 0.90 for leaders and
from 0.77 to 0.83 for followers.
3.2.2.4. Dyad tenure. We operationalized dyad tenure by the wave
that leaders and followers filled out the questionnaire. By adopting
this approach we kept LMX formation time constant across partic-
ipants. Prior research has shown that LMX tends to stabilize
between two weeks and two months from the start of the relation-
ship (Dansereau, Graen, & Haga, 1975; Liden et al., 1993; Nahrgang
et al., 2009).
3.2.2.5. Control variables. We controlled for the same state- and
trait-like constructs adopted in Study 1. State anxiety was mea-
sured from wave 2 to wave 5. Leader’s coefficient alpha ranged
from 0.80 to 0.83. Follower’s coefficient alpha ranged from 0.79
to 0.83. We measured leader and follower job performance every
15 days. Leader’s coefficient alpha ranged from 0.78 to 0.83 and
follower’s coefficient alpha ranged from 0.81 to 0.83. OCBI was also
measured every 15 days from wave 2 to wave 5. Leader’s coeffi-
cient alpha ranged from 0.78 to 0.88 and follower’s coefficient
alpha ranged from 0.86 to 0.90. Extraversion and agreeableness
were measured at wave 1. Coefficient alpha reliability for extraver-
sion was 0.85 for leaders and 0.94 and followers. Coefficient alpha
for agreeableness was 0.87 for leaders and 0.94 for followers. We
also controlled for leader and follower age, gender, work experi-
ence prior joining the company, and tenure in the company (lead-
ers only). Similar to Study 1, we controlled for the other party LMX
evaluations to minimize concerns about interdependencies.
1
We conducted a diary study on MTurk to assess the degree to which the average
of 14 daily ratings of sleep, hostility, and LMX correlate with a single survey
retrospectively measuring these same constructs over the same time period. We
recruited 130 participants to complete 15 daily online surveys. The first 14 surveys
asked participants to report their sleep from the previous night, their experienced
daily hostility at that moment, and their current evaluation of their LMX. After the 14
daily surveys, participants completed one final survey the next day in which we asked
them to report how many hours they had slept in the past 15 nights on average, the
extent they experienced hostility in the past 15 days on average, and the relationship
quality with their leaders in the past 15 days on average. We adopted the same scales
used in this study. Eighty-one participants filled out all surveys. We calculated the
within-person average of the daily reported scales and correlated them with the
participants’ reported average over the past 15 days. The correlation between the
calculated average sleep quantity and retrospectively reported measure of sleep over
the past 15 days was 0.91. The correlation between the calculated average hostility
and retrospectively reported state hostility over the past 15 days was 0.91. The
correlation between the calculated average LMX and retrospectively reported LMX
over the past 15 days was 0.90. We next ran multilevel models to take into
consideration the nesting structure of our data. Retrospective ratings predicted daily
ratings for sleep quantity (c = 0.84, p < 0.01), hostility (c = 0.74, p < 0.01), and LMX
(c = 0.74, p < 0.01).
66 C.L. Guarana, C.M. Barnes / Organizational Behavior and Human Decision Processes 141 (2017) 57–73
3.2.3. Analysis
Given that the data collected to test our model violate assump-
tions of independence of observation required for ordinary least
squares (OLS) regression analyses (biweekly assessment nested
within leaders and followers, and followers nested within leaders),
we conducted our analyses in a multilevel format using Mplus
(Muthen & Muthen, 1998–2010). In addition, in order to test the
proposed mediation, and moderated mediation hypotheses, we
performed multilevel path analysis (MacKinnon, 2008; Preacher,
Zyphur, & Zhang, 2010) and adopted Monte Carlo simulator to gen-
erate the confidence intervals for the indirect effects (Preacher &
Selig, 2010). Following Wang, Liao, Zhan, and Shi (2011), control
variables were modeled with fixed slopes and allowed to correlate
at different levels. The data consisted of three levels. Level 1 com-
prised biweekly leaders and followers rating of sleep quantity,
relationship quality, state hostility, anxiety, and job performance.
Level 2 contained followers’ age, gender, work experience, extro-
version, and agreeableness. Level 3 included leaders’ age, gender,
work experience, tenure in the company, extroversion, and
agreeableness.
To test the hypothesized within-individual relationships among
sleep quantity, state hostility, and relationship quality, all Level 1
variables were group-mean-centered at the individual level. It is
worth noting that because follower hypotheses were all within-
person analyses (Level 1), any between-person differences were
accounted for by group mean centering. Thus, any between-
individual differences in follower need for sleep is parsed out in
these analyses. Variables at Level 2 were group-mean-centered at
the leader level, and variables at Level 3 were grand-mean-
centered (Hoffman, Griffin, & Gavin, 2000). The leader-level
hypotheses were also within-leaders, meaning that between-
individual differences in leaders (such as individual need for sleep)
were parsed out for leaders as well. To test for the hypothesized
moderating effects of dyadic tenure on the within-individual rela-
tionship, we used the group-mean-centered variables to create the
multiplicative term. The same multiplicative term was used in
testing the moderated mediation hypotheses.
3.2.4. Results
Supplementary data reports the descriptive statistics and the
correlations at the within-individual level. Consistent with the
LMX theory and prior findings, the agreement between leaders’
and followers’ evaluations of their relationship strengthens over
time (r1 = 0.07, r 2 = 0.19, r3 = 0.22, r4 = 0.25). Followers sleep quan-
tity was positively correlated with leaders’ evaluations of relation-
ship quality at Times 1 and 2. Leaders’ sleep quantity was also
positively related to followers’ evaluations of relationship quality
at Times 1 and 2. Supplementary data reports the correlations at
the between-individual levels. Between-individual correlations
aggregated at the follower level are shown below the diagonal
(n = 120), and between-individual correlations aggregated at the
Table 4
Parameter Estimates and Variance Components of Null Models for Level 1 Endogenous Variables in Study 2.a
Variables Intercept b00 Within-individual variance (e2 )
Follower sleep quantity 390.81 631.88
Follower hostility 1.59 0.11
Follower ratings of relationship quality 3.76 0.04
Leader sleep quantity 390.19 676.53
Leader hostility 1.60 0.15
Leader ratings of relationship quality 3.54 0.49
Note.
a
N = 600; b00 is the pooled intercept representing average level of variable across individuals; e2 is the within-individual variance in a variable; and r 2 is the between-
individual variance in the variable.
*
p < 0.05.
leader level are shown above the diagonal (n = 40). As a prelimi-
nary step in the analysis, we ran latent growth models to investi-
gate if leader–follower relationships tend to stabilize after the
initial developmental phases. The results showed that the models
with quadratic terms fit the data better than the models with lin-
ear terms only (leader evaluations: D v2 with Satorra-Bentler cor-
rection = 35.81, df = 4; p < 0.01; follower evaluations: leader
evaluations: D v2 with Satorra-Bentler correction = 14.45, df = 4;
p < 0.01) (leader quadratic slope = 0.02, p < 0.01; follower quad-
ratic slope = 0.02, p < 0.01) suggesting that leader-follower rela-
tionships develop over a few weeks quickly and then tend to
remain stable after initial phases. These findings are consistent
with those of Nahrgang et al. (2009).
Before testing our hypotheses, we sought to provide some ini-
tial evidence for the amount of within-individual variance in our
models. The results of null models in HLM for each endogenous
variable showed that 69.83% of the variance in follower sleep
quantity is accounted by individuals, and 60.37% of the variance
in leader sleep quantity is accounted by individuals. Also the anal-
yses showed that individual accounted for 61.11% of the variance
in follower hostility, 50.00% of variance in leader hostility, 12.50%
of the variance in follower ratings of relationship quality, and
89.10% of the variance in leader ratings of relationship quality
(see Table 4 for details). The amount of within-individual
explained variance was similar to those presented elsewhere
(e.g., Scott & Barnes, 2011).
We proposed that leader sleep quantity is positively related to
follower perceptions of relationship quality (Hypothesis 1a), and
that follower sleep quantity is positively related to leader percep-
tions of relationship quality (Hypothesis 1b). Results of model 1 in
Table 5 provided support for Hypothesis 1a (c = 0.001, p < 0.05).
Within individuals, leader sleep quantity had a positive and signif-
icant relationship with follower perceptions of relationship quality
even after accounting for leader’s and follower’s stable characteris-
tics (agreeableness and extroversion), state anxiety, follower eval-
uations of leader job performance and OCBI, leader perceptions of
relationship quality, and demographics. Model 5 demonstrated no
significant relationship between follower sleep quantity and leader
perceptions of relationship quality (c = 0.0001, ns), therefore
Hypothesis 1b was not supported.
Turning to the indirect effects of sleep quantity on relationship
quality through hostility, we predicted that leader sleep quantity
will have an indirect effect on follower perceptions of relationship
quality via leader hostility (Hypothesis 2a) and that follower sleep
quantity will have an indirect effect on leader perceptions of rela-
tionship quality via follower hostility (Hypothesis 2b). In support
of Hypothesis 2a, leader sleep quantity was negatively related to
leader hostility (c = 0.01, p < 0.01), and leader hostility was
negatively related to follower perceptions of relationship quality
(c = 0.08, p < 0.05). Leader hostility mediated the relationship
between leader sleep quantity and follower perceptions of
Between-individual variance (r 2 ) Percentage of within-individual variance
272.93 69.84%
0.07* 61.11%
0.28 12.5%
444.07 60.37%
0.15 50.00%
0.06 89.10%
Table 5
Conditional Indirect Effects of Sleep Quantity and Dyadic Tenure on Relationship Quality via Hostility in Study 2.a

Follower ratings of relationship quality Leader ratings of relationship quality

Variables Model 1 – Model 2 – Model 3 -interaction Model 4 – Model 5 – Model 6 – Model 7 -interaction Model 8 –
direct effect indirect effect cond. ind. effect direct effect indirect effect cond. ind. effect
Intercept 4.23** 4.24** 4.29** 4.27** 0.21 0.13 0.18 0.20
DV = Ratings of hostility
Level 1 – Variables
Follower sleep quantity 0.009** 0.009**
Leader sleep quantity 0.01** 0.01**
DV = Rating of relationship quality

C.L. Guarana, C.M. Barnes / Organizational Behavior and Human Decision Processes 141 (2017) 57–73
Level 1 - Control variables
Dyadic tenure 0.05** 0.04** 0.05** 0.04** 0.04* 0.04** 0.04* 0.04**
Follower hostility 0.04 0.03 0.04 0.03
Leader hostility 0.03 0.03 0.02 0.02
Follower anxiety 0.04 0.03 0.04 0.03 0.07 0.09 0.07 0.07
Leader anxiety 0.02 0.02 0.03 0.03 0.06 0.07 0.07 0.04
Follower sleep quantity 0.001 0.001 0.001 0.001
Leader sleep quantity 0.001 0.001 0.001 0.001
Follower ratings of leader performance 0.16** 0.16** 0.16** 0.16** 0.06 0.05 0.06 0.05
Leader ratings of follower performance 0.02 0.02 0.02 0.02 0.22 0.21* 0.22** 0.21**
Follower ratings of leader OCBI 0.11** 0.12** 0.11** 0.12** 0.09 0.09 0.08 0.10*
Leader ratings of follower OCBI 0.02 0.03 0.02 0.03 0.09 0.09 0.09 0.09
Follower ratings of relationship quality 0.10 0.09 0.09 0.08
Leader ratings of relationship quality 0.06 0.05 0.06 0.05
Level 2 – Control variables
Follower gender 0.10 0.10 0.09 0.09 0.11 0.02 0.11 0.10
Follower age 0.04** 0.04** 0.04** 0.04** 0.01 0.01 0.01 0.01
Follower work experience 0.04* 0.05* 0.04* 0.04* 0.02 0.02 0.01 0.02
Follower extroversion 0.05 0.06 0.05 0.05 0.08 0.09 0.08 0.09
Follower agreeableness 0.02 0.02 0.02 0.02 0.02 0.01 0.01 0.02
Level 3 – Control variables
Leader gender 0.15 0.15 0.17 0.17 0.10 0.10 0.08 0.11
Leader age 0.002 0.002 0.002 0.002 0.01 0.01 0.01 0.01
Leader work experience 0.001 0.001 0.002 0.001 0.005 0.005 0.001 0.01
Leader tenure 0.04 0.04 0.04 0.04 0.03 0.03 0.02 0.03
Leader extroversion 0.10 0.11 0.19 0.19 0.07 0.07 0.03 0.07
Leader agreeableness 0.08 0.08 0.04 0.03 0.04 0.04 0.04 0.04
Level 1 – Variables
Follower hostility 0.16** 0.16**
Leader hostility 0.08* 0.08*
Follower sleep quantity 0.001 0.001 0.001 0.04
Dyadic tenure  Follower sleep quantity 0.001* 0.04
Leader sleep quantity 0.001* 0.001 0.001* 0.001
Dyadic tenure  Leader sleep quantity 0.001 0.001
Loglikelihood 7325.19 7188.06 7321.46 7184.37 7334.84 7210.83 11759.52 7199.29
Parameter (95% CI) Parameter (95% CI) Parameter (95% CI) Parameter (95% CI)
Indirect effect 0.001* (0.0001; 0.002) 0.001* (0.0001; 0.002)
Conditional indirect effect
Early (Dyadic tenure = -1.5) 0.001 (.0001; 0.002) 0.002* (.001; 0.003)
Late (Dyadic tenure = 1.5) 0.001 (.0001;0.001) 0.000 (.0001; 0.001)
Difference 0.000 ( 0.001; 0.001) 0.001 ( 0.001; 0.003)

Note. N = 600 at the within-individual level (Level 1), n = 120 at the follower level (Level 2), n = 40 at the leader level (Level 3).
*
p < 0.05.
**
p < 0.01.

67
68 C.L. Guarana, C.M. Barnes / Organizational Behavior and Human Decision Processes 141 (2017) 57–73

relationship quality (indirect effect = 0.001, p < 0.05; 95% CI 0.0001, Table 6
0.002). Results of Monte Carlo simulation with 20,000 Monte Carlo Results of gamma difference tests in study 2.a

replications showed that the 95% bias-corrected bootstrap CI Time DV = Leader ratings DV = Follower ratings Gamma difference
excluded zero [0.0002; 0.002], providing further support for our of relationship quality of relationship quality test
Hypothesis 2a. In addition, follower sleep quantity was negatively IV = Leader sleep IV = Leader sleep
related to follower hostility (c = 0.009, p < 0.01), and follower quantity quantity
hostility was negatively related to leader perceptions of relation- 1 c = 0.001, p = 0.99 c = 0.007, p = 0.00 z = 3.46, p = 0.001
2 c = 0.001, p = 0.74 c = 0.007, p = 0.00 z = 2.25, p = 0.01
ship quality (c = 0.16, p < 0.01). Follower hostility mediated the 3 c = 0.001, p = 0.87 c = 0.003, p = 0.17 z = 0.51, p = 0.31
relationship between follower sleep quantity and leader percep- 4 c = 0.001, p = 0.84 c = 0.004, p = 0.24 z = 1.30, p = 0.20
tions of relationship quality (indirect effect = 0.001, p < 0.05; 95% IV = Follower sleep IV = Follower sleep
CI 0.0001, 0.002). Results of 20,000 Monte Carlo replications also quantity quantity
showed evidence for the proposed indirect effect (95% CI 0.0003, 1 c = 0.01, p = 0.00 c = 0.001, p = 0.74 z = 3.02, p = 0.001
0.002); therefore, although we did not find support for Hypothesis 2 c = 0.01, p = 0.02 c = 0.001, p = 0.70 z = 2.42, p = 0.02
1, which stated that follower sleep quantity was related to leader 3 c = 0.003, p = 0.56 c = 0.001, p = 0.64 z = 0.83, p = 0.20
4 c = 0.002, p = 0.53 c = 0.002, p = 0.17 z = 0.65, p = 0.26
perceptions of relationship quality, Hypothesis 2b was supported
(see models 3 and 7 in Table 5). Note.
a
N = 600 at the within-individual level (Level 1), n = 120 at the follower level
Hypothesis 3a stated that the effects of leader sleep quantity on
(Level 2), n = 40 at the leader level (Level 3).
follower perceptions of relationship quality are stronger at early
stages of the relationship. The results of model 3 in Table 5 showed
that interaction term was not significant (c = 0.000, ns). Hypothesis are stronger on the counterparty perceptions of the relationship
3b posited that follower sleep quantity has a stronger effect on lea- quality than on the self-perceptions of the relationship quality.
der perceived relationship quality at the beginning of their tenure. We adopted gamma difference tests to analyze these hypotheses.
As expected, results of model 7 indicated that dyadic tenure has a First, we got the standardized coefficients by multiplying the
negative interactive effect with sleep quantity to predict relation- unstandardized coefficient by the division of the standard devia-
ship quality (c=.-001, p < 0.05). Following procedures by Aiken tion of the predictor (sleep quantity) by the standard deviation of
and West (1991), we plotted the interaction (see Fig. 2). We also the outcome (relationship quality). Second, considering that stan-
conducted multilevel simple slope analysis (Preacher, Curran, & dardized coefficients can be interpreted as correlations, we used
Bauer, 2006). These analyses provided additional support for the a Fisher r-to-z transformation, and calculate a z-value that can be
findings. The effects of follower sleep quantity on leader percep- applied to assess the significance of the difference between two
tions of relationship quality was significant at the beginning of correlation coefficients, r leader and rfollower found in two independent
their tenure (c = 0.005, p < 0.01); however, it was not significant samples. Both hypotheses were supported for dyadic relationships
late in their tenure (c = 0.002, p = 0.76). at the beginning of their tenure, as indicated by the results of
Hypotheses 4a and 4b predicted that dyad tenure influenced gamma differences on Table 6. Leader sleep quantity had a stronger
the strength of the indirect relationship between sleep quantity effect on follower ratings of the relationship quality at Time 1
and perceptions of relationship quality via hostility. As demon- (z = 3.46, p = 0.001) and Time 2 (z = 2.25, p = 0.01), but not later
strated by Table 5, model 4, the indirect effect of leader sleep quan- in their tenure (Time 3: z = 0.51, p = 0.31; Time 4: z = 1.30,
tity on follower perceptions of relationship quality via leader p = 0.20). Similar results were found for follower sleep quantity
hostility did not vary based on dyadic tenure (D c = 0.000; 95% and leader ratings of relationship quality. Follower sleep quantity
CI 0.001, 0.001). We found similar results for the conditional indi- had a stronger effect on leader perceptions of relationship quality
rect effect of follower sleep quantity on leader perceptions of rela- than on follower perceptions at the beginning of their tenure (Time
tionship quality via follower hostility. The strength or direction of 1: z = 3.02, p = 0.001; Time 2: z = 2.42, p = 0.02) but not later in
the effects did not change based on dyadic tenure (D c = 0.001; 95% their tenure (Time 3: z = 0.83, p = 0.20; Time 4: z = 0.65, p = 0.26).
CI 0.001, 0.003). Thus, Hypotheses 4a and 4b were not supported
(see models 4 and 8 in Table 5).
Finally, we proposed in hypotheses 5a (leader sleep quantity) 4. Discussion
and 5b (follower sleep quantity) that the effects of sleep quantity
In their review of the LMX literature, Anand, Hu, Liden, and
Vidyarthi (2011) noted that ‘‘relationship quality develops quite
early and remains generally stable through the life of the dyad”
(p. 311). It is perhaps surprising that so few papers have investi-
gated early leader-follower relationship formations. In one of these
few papers, Nahrgang et al. (2009) called for researchers to go
beyond theorizing about established dyad relationships and exam-
ine individual characteristics that may affect the quality of the
relationship at initial interactions, especially in organizational set-
tings. In addition, current work on LMX has encouraged work on
the effect of discrete emotions on LMX development (Cropanzano
et al., 2016). Towards those ends, we conducted two studies with
newly formed leader–follower dyads. In Study 1, we collected data
from leaders and followers in their first three days of work rela-
tionship. In Study 2, we adopted a longitudinal design and col-
lected data from new dyads for over three months. We examined
the effects of a dynamic individual characteristic (i.e. lack of sleep)
via discrete emotions (i.e., hostility) on LMX.
Fig. 2. Interactive effect of follower sleep quantity and dyadic tenure on leader. We found that leader and follower lack of sleep influenced the
Perceptions of relationship quality for Study 2. other party’s perceptions of relationship quality via hostility in
 C.L. Guarana, C.M. Barnes / Organizational Behavior and Human Decision Processes 141 (2017) 57–73
both studies. We also found that leader sleep was related to fol-
lower perceptions of relationship quality regardless of dyad tenure,
whereas follower sleep quantity was positively related to leader
perceptions of the relationship only at the beginning of their rela-
tionship in Study 2. These findings were slightly different from the
ones in Study 1, in which we found that both leader and follower
sleep quantity were related to the other party’s perceptions of rela-
tionship quality. Considering that leaders and followers were in
their first three days of the work relationship in Study 1, this is
consistent to the findings of the moderating effect in Study 2. Also,
in Study 1, we showed that people can accurately recall their aver-
age sleep quantity over periods of two weeks. Nevertheless, we did
not find support for our second-stage moderated mediation model,
suggesting that the effects of sleep quantity on perceptions of rela-
tionship quality via hostility should not vary based on dyad tenure
(Study 2). Finally, we found that lack of sleep influenced relational
awareness in both studies. Specifically, leader sleep quantity had a
stronger effect on follower ratings of relationship quality than self-
ratings of relationship quality, and vice versa for followers. Our
study has several theoretical and practical implications.
4.1. Theoretical implications
Our findings challenge some of the received wisdom of the LMX
literature where researchers have largely emphasized the effects of
stable individual characteristics on relationship quality (Dulebohn
et al., 2012; Martin et al., in press; Nahrgang et al., 2009). By exam-
ining the effects of lack of sleep on early LMX development, we
advance a more nuanced understanding of leader-follower rela-
tionship formation than proposed by prior research. That is, our
findings show not only that a dynamic individual characteristic
(i.e., sleep) can influence an individual’s perceptions of relationship
quality over and above stable levels of extraversion and agreeable-
ness, but also demonstrate that dyadic tenure is of particular
importance for the relationship between follower sleep and leader
perceptions of relationship quality. These findings, therefore, lend
support to the role of dynamic variables on leader-follower rela-
tionship formation where relational and contextual cues can
inform individuals’ perceptions that affect their social interactions.
This manuscript deepens our understanding of sleep depriva-
tion’s role in triggering hostility, which affects the development
of leader-follower relationship quality. We found that sleep-
deprived leaders and followers experience hostility, which influ-
ences the other party’s perception of relationship quality. This is
consistent with the sleep and emotion regulation model (Barnes,
2012), which suggests that sleep deprivation increases the experi-
ence of negative emotions and the individual’s ability to suppress
it. Our results also imply that LMX is influenced by emotional cues
that are triggered by factors potentially unrelated to the organiza-
tional context. Therefore, we advance the LMX field by incorporat-
ing a distal antecedent (sleep) and a discrete emotion (hostility) as
antecedents influencing the development of leader-follower rela-
tionship quality. This is consistent with the work by Cropanzano
and colleagues that developed a theoretical model explaining
how discrete emotions fluctuate and influence LMX development.
Although we predicted that dyad tenure would have a moderat-
ing effect on the relationship between leader sleep and follower
perceptions of relationship quality, such as the effects would be
stronger at initial interactions than at late interactions, we did
not find support for this hypothesis in Study 2. Perhaps this is sim-
ply the result of a small sample size for our level 3 variable (n = 40),
which makes it difficult to find cross-level interactions (Scherbaum
& Ferreter, 2009). An alternative explanation is that it takes longer
for followers to attribute leaders’ negative emotions and conflict-
ing behaviors to external factors than it takes for leaders to do
so. One of the typical roles of leaders is to monitor their followers
69
(Morgeson, DeRue, & Karam, 2010), which can increase the number
of behavioral observations over time. Leaders, therefore, can have
more information about the uniqueness and variability of follow-
ers’ emotional and behavioral responses to situations, and attribute
followers’ behaviors to contextual factors. Such accumulative
information to change attribution processes, may take longer for
followers.
We also predicted that the indirect effect of sleep quantity on
perceptions of relationship quality via hostility would be stronger
at the beginning of the leader-follower dyad tenure. We did not
find support for these hypotheses. One potential explanation for
these null effects is that hostility cues are extremely salient and
easily accessed by the observers, making them difficult to over-
come even when individuals gather more data updating their attri-
butions. Researchers suggest that negative information is
remembered more clearly than neutral or positive information
(Baumeister, Bratslavsky, Finkenauer, & Vohs, 2001). In addition,
expressions of hostility may not match the cognitive representa-
tions individuals have of leaders (implicit leadership theory; e.g.,
Epitropaki & Martin, 2005) and followers (implicit followership
theory; e.g., Sy, 2010).
In addition, our investigation deepens our understanding of the
asymmetrical effect of sleep deprivation on relational awareness
(Gordon et al., 2014; Killgore et al., 2008). We found that sleepy
leaders and followers underestimate the effects of lack of sleep
on the perceptions of the relationship quality by the other party.
That is, sleep-deprived leaders and followers are not completely
aware of the affective and behavioral signals that are perceived
by their counterparty. The presence of relational asymmetry in
dyads is consistent with recent work on emotional inaccuracy
reported by sleep-deprived couples (Gordon et al., 2014). Our
results also imply that the low level of agreement between leader
and follower ratings of LMX (Sin et al., 2009) may be the result of
impaired self-awareness caused by lack of sleep. Therefore, we also
advance the LMX field by examining the perceptual factors that
influence agreement beyond objective relationship tenure and
number of dyadic interactions (Sin et al., 2009).
4.2. Practical implications
Our study makes several contributions to practice. Sleep depri-
vation is a prevalent phenomenon in organizations. Luckhaupt,
Tak, and Calvert (2010) found that 40.5% of employees get less than
six hours of sleep on average, even though the National Sleep
Foundation (2009) recommends seven to nine hours of sleep per
night for adults. Swanson et al. (2011) found that 29% of Americans
experienced extreme sleepiness or had even fallen asleep at work
in the past month. Such sleep deprivation is likely causing wide-
spread decrements in LMX. Even if those decrements to each dyad
may be relatively small, the number of potential dyads affected by
sleep deprivation may result in a huge aggregate effect.
Our results add to the body of research showing that sleep
deprivation has negative effects on organizations, this time extend-
ing to the effects on the behavior of others. The underlines the
importance of recommendations for companies adopting strategies
for minimizing sleep loss (see Barnes, 2012 for a review). However,
our findings supplement this recommendation by showing that
lack of sleep can be harmful to leader-follower relationship forma-
tion and to voluntary supportive behaviors towards co-workers. In
other words, although sleepy leaders and followers may try to
develop high-quality relationships, lack of sleep can make them
unpleasant to work with. Even worse, sleepy people may not be
aware of the negative signs that they send to the counterparty.
High-quality relationships between leaders and followers are crit-
ical for organizational success. The relationship between LMX and
job performance (in-role and extra-role behaviors) has been con-
70 C.L. Guarana, C.M. Barnes / Organizational Behavior and Human Decision Processes 141 (2017) 57–73
sistently found in several studies (Dubeblohn et al. 2012; Ilies
et al., 2007). Consequently, managers may want to avoid infringing
upon employees’ sleep by providing the right amount of resources
and organizational support to get work done during work hours.
For instance, simply by not using their smartphones at night, lead-
ers and followers can sleep better (Lanaj, Johnson, & Barnes, 2014).
The finding that leaders and followers are unaware of the neg-
ative consequences of lack of sleep on their professional relation-
ship suggests that relational feedback mechanisms (VanderDrift
& Agnew, 2014) could be implemented early in their dyad tenure.
Because the quality of the relationship is highly and disproportion-
ally influenced by early interactions, such feedback mechanisms
may increase relational awareness, and provide opportunities for
sleepy leaders and followers to regulate their behaviors after a
sleepless night. Regulating sporadic negative emotions and con-
flicting behaviors, therefore, can allow initial leader-follower rela-
tionships to develop into high-quality exchanges. In addition,
relational feedback mechanisms can allow leaders and followers
to overcome initial superficial categorization processes (Harrison,
Price, & Bell, 1998). Therefore, relational feedback may not only
help leaders and followers align their perceptions of their relation-
ship quality, but also minimize the effects of implicit stereotypes.
4.3. Limitations and future research
Although our studies possess several strengths (e.g., biweekly
measures of sleep and LMX, field context, and new work dyads),
there are also some limitations. First, our data were collected from
leaders and followers in a Brazilian organizations. Some Brazilian
cultural differences (e.g., low in individualism) (Hofstede, 1980)
may limit the generalizability of the results. Brazilians may be
more sensitive to relationships and more closely monitor and eval-
uate their partners. Consequently, relational and contextual cues
may be more salient and easily activated in cultures high on collec-
tivism. Nevertheless, researchers have shown that relationships
are important in individualistic and collectivistic cultures
(Triandis, Bontempo, Villareal, Asai, & Lucca, 1988) and that differ-
ences in the effects of LMX on task performance and affective com-
mitment are not based on culture (Rockstuhl, Dulebohn, Ang, &
Shore, 2012). Future research can investigate whether the effects
of lack of sleep on LMX depend on whether a culture is individual-
istic or collectivistic.
Second, our research investigates the effects of sleep in work
relationships by testing our hypotheses with new leader-follower
dyads. Although the field study increases our confidence about
generalizability, we cannot say the same about causality. Never-
theless, we have adopted a within-individual longitudinal design,
controlled for theoretical relevant variables (agreeableness, extro-
version, job performance, OCBI, and state anxiety), and used differ-
ent sources of information to minimize the presence of potential
alternative explanations and common method biases. Despite our
effort, some potential unobserved variables could be causing the
effects on sleep and relationship quality, creating an endogeneity
problem. For example, followers may have accepted their new jobs
because of a lack of better options. Although this may be a plausi-
ble explanation for some followers, we do not expect the same for
leaders. To better understand the role of person-environment and
person-organizational fit in the relationship between lack of sleep
and LMX, future researchers might measure fit development and
how it correlates with lack of sleep over time. Perhaps sleep-
deprived individuals have a tougher time perceiving that the new
job is a good fit for them (objective vs. subjective fit).
Third, we asked participants in Study 2 to think about their last
15 days to report their sleep quantity, state hostility and anxiety,
perceptions of relationship quality, job performance, and OCBI.
Although we have adopted this time frame based on the work of
others (e.g., Nahrgang et al., 2009), asked participants to report
their sleep quantity, state hostility, and anxiety in a particular
day (Study 1), and conducted a pilot test on MTurk to minimize
concerns about our operationalizations, we may have missed some
of the more granular and important variance in sleep, should we
had adopted a shorter window. Future research, therefore, should
consider other time frames and operationalizations of sleep (e.g.,
sleep quality) to investigate if more frequent changes in sleep
influence the development of leader-follower relationship quality.
For example, scholars can adopt a diary design that collects data on
sleep and LMX every day. Another potential future direction is to
investigate the effects of sleep debt on the development of LMX.
Although researchers have not arrived at a consensus on how far
back in time sleep debt reaches and how to separate out individual
differences in the need for sleep, it seems likely that sleep debt can
also affect LMX ratings. Future research can follow individuals for
longer periods of time and investigate the amount of sleep debt
necessary to change LMX evaluations.
Fourth, despite the large body of research outside management
showing that lack of sleep influences relationship quality via emo-
tions (Bernerth, Armenakis, Field, Giles, & Walker, 2007; Gordon &
Chen, 2013; Hui et al., 1999; Spanier & Thompson, 1982) we mea-
sured only two discrete emotions (hostility and anxiety). Because
we collected data over five waves with real leaders and followers
(Study 2), we kept the survey at a minimum size to prevent partic-
ipant fatigue and attrition. Future research might benefit from col-
lecting cross-sectional data and from adding other potential and
theoretically relevant discrete emotions and workload-related
variables. For example, gratitude has been theoretically related to
LMX (Cropanzano et al., 2016) and can be a predictor of LMX.
Fifth, our research showed that leaders and followers may not
be completely aware of the impact of their sleep deprivation in
the development of relationships. Although prior research has
shown that LMX agreement is low among leaders and followers
(Sin et al., 2009) and many factors can influence LMX agreement,
we have shown that sleep may be one of these factors and can
intensify the interpretative distance between leaders and follow-
ers. Future scholars should investigate if work experience, for
example, influences LMX agreement. It could be that more experi-
enced workers are better at interpreting the quality of their rela-
tionship with other parties.
Finally, although our theory suggests that lack of sleep causes
hostility, there may be a recursive effect in which anger leads to
sleep loss (e.g., Brissette & Cohen, 2002). Therefore, low levels of
relationship quality may cause sleep loss. Nevertheless, our studies
offer an important first glance at the directionality of sleep on LMX.
To better understand the causal relationship between sleep and
LMX, future research might use laboratory experiments that com-
plement our survey findings.
5. Conclusions
The development of a leader-follower relationship is a complex
process. Although stable personality characteristics can influence
initial judgments in relationship formation, our study shows that
dynamic individual differences can also send hostility cues that
affect the type of exchanges in which leaders and followers engage.
The relationship between leaders and followers suffers when lead-
ers and/or followers are sleep-deprived. Our study shows that fol-
lowers’ lack of sleep is particularly important at the beginning of
their relationship with leaders. Perhaps even more surprising,
sleepy leaders and followers seem to be unaware of how unpleas-
ant they are. We conclude that organizations that pay attention to
the sleep patterns of their employees may enjoy not only a more
productive, less hostile, but also one that is more pleasant.
 C.L. Guarana, C.M. Barnes / Organizational Behavior and Human Decision Processes 141 (2017) 57–73 71

Appendix A. Supplementary data Dienesch, R. M., & Liden, R. C. (1986). Leader–member exchange model of
leadership: A critique and further development. Academy of Management
Review, 11(3), 618–634.
Supplementary data associated with this article can be found, in Dulebohn, J. H., Bommer, W. H., Liden, R. C., Brouer, R. L., & Ferris, G. R. (2012). A
the online version, at http://dx.doi.org/10.1016/j.obhdp.2017.04. meta-analysis of antecedents and consequences of leader-member exchange:
Integrating the past with an eye toward the future. Journal of Management, 38
003.
(6), 1715–1759.
Durmer, J. S., & Dinges, D. F. (2005). Neurocognitive consequences of sleep
deprivation. Seminars in Neurology, 25, 117–129.
References Epitropaki, O., & Martin, R. (2005). From ideal to real: A longitudinal study of the
role of implicit leadership theories on leader-member exchanges and employee
outcomes. Journal of Applied Psychology, 90, 659–676.
Aiken, L. S., & West, S. G. (1991). Testing and interpreting interactions in multiple
Fraley, R. C., & Shaver, P. R. (2000). Adult romantic attachment: Theoretical
regression. Sage Publications.
developments, emerging controversies, and unanswered questions. Review of
Akram, U., Ellis, J. G., Myachykov, A., & Barclay, N. L. (2016). Misperception of
General Psychology, 4, 132–154.
tiredness in young adults with insomnia. Journal of Sleep Research. http://dx.doi.
Gailliot, M. T., Schmeichel, B. J., & Baumeister, R. F. (2006). Self-regulatory processes
org/10.1111/jsr.12395.
defend against the threat of death: Effects of self-control depletion and trait
Anand, S., Hu, J., Liden, R. C., & Vidyarthi, P. R. (2011). Leader-member exchange:
self-control on thoughts and fears of dying. Journal of Personality and Social
Recent research findings and prospects for the future. In A. Bryman, D.
Psychology, 91, 49–62.
Collinson, K. Grint, B. Jackson, & M. Uhl-Bien (Eds.), The sage handbook of
Gomes-Schwartz, B. (1978). Effective ingredients in psychotherapy: Prediction of
leadership. Thousand Oaks, CA: Sage.
outcome from process variables. Journal of Consulting and Clinical Psychology, 46
Ashforth, B. E., Kreiner, G. E., & Fugate, M. (2000). All in a day’s work:
(5), 1023–1035.
Boundaries and micro role transitions. Academy of Management Review, 25(3),
Gordon, A. M., & Chen, S. (2014). The role of sleep in interpersonal conflict: Do
472–491.
sleepless nights mean worse fights? Social Psychological and Personality Science,
Banks, S., & Dinges, D. F. (2007). Behavioral and physiological consequences of sleep
5(2), 168–175.
restriction in humans. Journal of Clinical Sleep Medicine, 3(5), 519–528.
Graen, G., & Cashman, J. R. (1975). A role making model in formal organizations: A
Banks, S. J., Eddy, K. T., Angstadt, M., Nathan, P. J., & Phan, K. L. (2007). Amygdala–
developmental approach. In J. G. Hunt & L. L. Larson (Eds.), Leadership frontiers
frontal connectivity during emotion regulation. Social Cognitive and Affective
(pp. 143–165). Kent, OH: Kent State Press.
Neuroscience, 2(4), 303–312.
Graen, G. (1976). Role-making processes of leadership development. In M. D.
Barnes, C. M. (2012). Working in our sleep: Sleep and self-regulation in
Dunnette (Ed.), Handbook of industrial and organizational psychology
organizations. Organizational Psychology Review, 2(3), 234–257.
(pp. 1201–1245). Chicago: Rand McNally.
Barnes, C. M., Guarana, C. L., Nauman, S., & Kong, D. T. (2016). Too tired to inspire or
Graen, G., & Scandura, T. A. (1987). Toward a psychology of dyadic organizing. In L.
be inspired: sleep deprivation and charismatic leadership. Journal of Applied
L. Cummings & B. M. Staw (Eds.), Research in organizational behavior
Psychology. http://dx.doi.org/10.1037/apl0000123.
(pp. 175–208). Greenwich, CT: JAI Press.
Barnes, C. M., Lucianetti, L., Bhave, D., & Christian, M. (2015). You wouldn’t like me
Graen, G., & Uhl-Bien, M. (1995). The Relationship-based approach to leadership:
when I’m sleepy: Leader sleep, daily abusive supervision, and work unit
Development of LMX theory of leadership over 25 years: Applying a multi-level,
engagement. Academy of Management Journal, 58(5), 1419–1437.
multi-domain perspective. Leadership Quarterly, 6(2), 219–247.
Barnes, C. M., Schaubroeck, J., Huth, M., & Ghumman, S. (2011). Lack of sleep and
Gross, J. J. (1998). Antecedent- and response-focused emotion regulation: Divergent
unethical conduct. Organizational Behavior & Human Decision Processes, 115(2),
consequences for experience, expression, and physiology. Journal of Personality
169–180.
and Social Psychology, 74, 224–237.
Barnes, C. M., Wagner, D. T., & Ghumman, S. (2012). Borrowing from sleep to pay
Guarana, C. L., & Hernandez, M. (2015). Building sense out of situational
work and family: Expanding time-based conflict to the broader nonwork
complexity: The role of ambivalence in creating functional leadership
domain. Personnel Psychology, 65(4), 789–819.
processes. Organizational Psychology Review, 5(1), 50–73.
Bartz, J. A., & Lydon, J. E. (2006). Navigating the interdependence dilemma:
Gujar, N., Yoo, S. S., Hu, P., & Walker, M. P. (2011). Sleep deprivation amplifies
Attachment goals and the use of communal norms with potential close others.
reactivity of brain reward networks, biasing the appraisal of positive emotional
Journal of Personality and Social Psychology, 91, 77–96.
experiences. Journal of Neuroscience, 31(12), 4466–4474.
Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001). Bad is stronger
Hamilton, D. L., Katz, L. B., & Leirer, V. O. (1980). Organizational processes in
than good. Review of General Psychology, 5(4), 323–370.
impression formation. In R. Hastie, T. M. Ostrom, E. B. Ebbesen, R. S. Wyer, Jr., D.
Bernerth, J. B., Armenakis, A. A., Field, H. S., Giles, W. F., & Walker, H. J. (2007). Is
L. Hamilton, & D. E. Carlston (Eds.), Person memory: The cognitive basis of social
personality associated with perceptions of LMX? An empirical study. Leadership
perception (pp. 121–153). Hillsdale, NJ: Erlbaum.
& Organization Development Journal, 28(7), 613–631.
Harrison, D. A., Newman, D. A., & Roth, P. L. (2006). How important are job
Bissonette, V. L., Rusbult, C. E., & Kilpatrick, S. D. (1997). Empathic accuracy and
attitudes? Meta-analytic comparisons of integrative behavioral outcomes and
marital conflict resolution. In W. Ickes (Ed.), Empathic accuracy (pp. 251–281).
time sequences. Academy of Management Journal, 49, 305–325.
New York, NY: The Guilford Press.
Harrison, D. A., Price, K. H., & Bell, M. P. (1998). Beyond relational demography:
Bradbury, T. N., & Fincham, F. D. (1990). Attributions in marriage: Review and
Time and the effects of surface- and deep-level diversity on work group
critique. Psychological Bulletin, 107, 3–33.
cohesion. Academy of Management Journal, 41(1), 96–107.
Brislin, R. W. (1986). The wording and translation of research instruments. In W. J.
Hasler, B. P., & Troxel, W. M. (2010). Couples’ nighttime sleep efficiency and
Loner & J. W. Berry (Eds.), Field methods in cross-cultural research (pp. 137–164).
concordance: Evidence for bidirectional associations with daytime relationship
Newbury Park, CA: Sage.
functioning. Psychosomatic Medicine, 72(8), 794–801.
Brissette, I., & Cohen, S. (2002). The contribution of individual differences in
Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process
hostility to the associations between daily interpersonal conflict, affect, and
analysis. New York: The Guilford Press.
sleep. Personality and Social Psychology Bulletin, 28(9), 1265–1274.
Heider, F. (1958). The psychology of interpersonal relations. New York, NY: Wiley.
Burger, J. M. (1991). Changes in attributions over time: The ephemeral fundamental
Hobson, J. A. (2005). Sleep is of the brain, by the brain and for the brain. Nature, 437
attribution error. Social Cognition, 9(2), 182–193.
(7063), 1254–1256.
Caldwell, J. A., Caldwell, J. L., Brown, D. L., & Smith, J. K. (2004). The effects of 37
Hoffmann, S. A., Griffin, M. A., & Gavin, M. B. (2000). The application of hierarchical
hours of continuous wakefulness on the physiological arousal, cognitive
linear modeling to organizational research. In K. Klein & S. Kozlowski (Eds.),
performance, self-reported mood, and simulator flight performance of F-117A
Multilevel theory, research, and methods in organizations: Foundations, extensions,
pilots. Military Psychology, 16(3), 163–181.
and new directions (pp. 75–170). San Francisco, CA: Jossey-Bass.
Cartwright, R. D., & Knight, S. (1987). Silent partners: The wives of sleep apneic
Hofstede, G. (1980). Culture’s consequences: International differences in work-related
patients. Sleep, 10(3), 244–248.
values. London: Sage Publications.
Christian, M. S., & Ellis, A. P. (2011). Examining the effects of sleep deprivation on
Hugenberg, K., & Bodenhausen, G. V. (2003). Facing prejudice: Implicit prejudice
workplace deviance: A self-regulatory perspective. Academy of Management
and the perception of facial threat. Psychological Science, 14(6), 640–643.
Journal, 54(5), 913–934.
Hui, C., Law, K. S., & Chen, Z. X. (1999). A structural equation model of the effects of
Costa, P. T., Jr., & McCrae, R. R. (1992). Revised NEO personality inventory (NEO-PI-R)
negative affectivity, leader-member exchange, and perceived job mobility on
and NEO five-factor inventory (NEO-FFI) professional manual. Odessa, FL:
in-role and extra-role performance: A Chinese case. Organizational Behavior and
Psychological Assessment Resources.
Human Decision Processes, 77(1), 3–21.
Côté, S. (2005). A social interaction model of the effects of emotion regulation on
Huston, T. L., & Vangelisti, A. L. (1991). Socioemotional behavior and satisfaction in
work strain. The Academy of Management Review, 30(3), 509–530.
marital relationships. Journal of Personality and Social Psychology, 61, 721–733.
Cropanzano, R., Dasborough, M., & Weiss, H. (2016). Affective events and the
Ilies, R., Nahrgang, J. D., & Morgeson, F. P. (2007). Leader-member exchange and
development of leader-member exchange. Academy of Management Review.
citizenship behaviors: A meta-analysis. The Journal of Applied Psychology, 92(1),
http://dx.doi.org/10.5465/amr.2014.0384.
269–277.
Dansereau, F., Graen, G., & Haga, W. J. (1975). A vertical dyad approach to leadership
Ireland, J., & Culpin, V. (2006). The relationship between sleeping problems and
within formal organizations. Organizational Behavior and Human Performance,
aggression, anger, and impulsivity in a population of juvenile and young
13(1), 46–78.
offenders. Journal of Adolescent Health, 38(6), 649–655.
Davidson, R. J., Putnam, K. M., & Larson, C. L. (2000). Dysfunction in the neural
Johnson, S. K. (2009). Do you feel what I feel? Mood contagion and leadership
circuitry of emotion regulation—A possible prelude to violence. Science, 289,
outcomes. The Leadership Quarterly, 20, 814–827.
591–594.
72 C.L. Guarana, C.M. Barnes / Organizational Behavior and Human Decision Processes 141 (2017) 57–73
Jones, E. E., & Davis, K. E. (1965). From acts to dispositions: The attribution process
in social psychology. In L. Berkowitz (Ed.), Advances in experimental social
psychology, 2 (pp. 219–266). New York: Academic Press.
Jones, E. E., & Nisbett, R. E. (1971). The actor and the observer: Divergent perceptions of
the causes of behavior. Morristown, NJ: General Learning Press.
Judge, T. A., Scott, B. A., & Ilies, R. (2006). Hostility, job attitudes, and workplace
deviance: Test of a multilevel model. Journal of Applied Psychology, 91(1),
126–138. http://dx.doi.org/10.1037/0021-9010.91.1.126.
Kamphuis, J., Meerlo, P., Koolhaas, J. M., & Lancel, M. (2012). Poor sleep as a
potential causal factor in aggression and violence. Sleep Medicine, 13(4),
327–334.
Kangas, H. M. (2013). The development of the LMX relationships after a newly
appointed leader enters an organization. Human Resource Development
International, 16(5), 575–589.
Kaplan, K. A., Itoi, A., & Dement, W. C. (2007). Awareness of sleepiness and ability to
predict sleep onset: Can drivers avoid falling asleep at the wheel? Sleep
Medicine, 9(1), 71–79.
Kelley, H. H. (1973). The processes of causal attribution. American Psychologist, 28
(2), 107–128.
Killgore, W. D. S., Kahn-Greene, E. T., Lipizzi, E. L., Newman, R. A., Kamimori, G. H., &
Balkin, T. J. (2008). Sleep deprivation reduces perceived emotional intelligence
and constructive thinking skills. Sleep Medicine, 9(5), 517–526.
Killgore, W. D. S., Kamimori, G. H., & Balkin, T. J. (2011). Caffeine protects against
increased risk-taking propensity during severe sleep deprivation. Journal of
Sleep Research, 20(3), 395–403.
Knutson, K. L., Rathouz, P. J., Yan, L. J. L., Liu, K., & Lauderdale, D. S. (2007). Intra-
individual daily and yearly variability in actigraphically recorded sleep
measures: The CARDIA study. Sleep, 30, 793–796.
Lanaj, K., Johnson, R., & Barnes, C. M. (2014). Beginning the workday yet already
depleted? Consequences of late-night smartphone use and sleep. Organizational
Behavior and Human Decision Processes, 124(1), 11–23.
Lassiter, G. D., Geers, A. L., Munhall, P. J., Handley, I. M., & Beers, M. J. (2001).
Videotaped confessions: Is guilt in the eye of the camera? In M. P. Zanna (Ed.).
Advances in experimental social psychology (Vol. 33, pp. 189–254). New York:
Academic Press.
Lee, K., & Allen, N. J. (2002). Organizational citizenship behavior and workplace
deviance: The role of affect and cognition. Journal of Applied Psychology, 87,
131–142.
Lench, H. C., Flores, S. A., & Bench, S. W. (2011). Discrete emotions predict changes in
cognition, judgment, experience, behavior, and physiology: A meta-analysis of
experimental emotion elicitations. Psychological Bulletin, 137(5), 834–855.
Liden, R. C., & Graen, G. (1980). Generalizability of the vertical dyad linkage model
of leadership. Academy of Management Journal, 23(3), 451–465.
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(4),
662–674.
Luckhaupt, S. E., Tak, S., & Calvert, G. M. (2010). The prevalence of short sleep
duration by industry and occupation in the National Health Interview Survey.
Sleep, 33, 149–159.
MacKinnon, D. P. (2008). Introduction to statistical mediation analysis. Mahwah, NJ:
Lawrence Erlbaum Associates Inc.
Major, B., Kaiser, C. R., & Mccoy, S. K. (2003). It’s not my fault: When and why
attributions to prejudice protect self-esteem. Personality and Social Psychology
Bulletin, 29(6), 772–781.
Martin, R., Yues, G., Thomas, G., Lee, A., & Epitropaki, O. (2015). Leader-member
exchange (LMX) and performance: A meta-analytic review. Personnel
Psychology, 69(1), 67–121.
Martinko, M. J., Harvey, P., & Douglas, S. C. (2007). The role, function, and
contributions of attribution theory to leadership: A review. Leadership Quarterly,
18(6), 561–585.
Marziali, E., Marmar, C., & Krupnick, J. (1981). Therapeutic alliance scales:
Development and relationship to psychotherapy outcome. American Journal of
Psychiatry, 138, 361–364.
Maxham, J. G. I., & Netemeyer, R. G. (2003). Firms reap what they sow: The effects of
shared values and perceived organizational justice on customers’ evaluation of
complaint handling. Journal of Marketing, 67(1), 29–45.
McGuire, S., McHale, S. M., & Updegraff, K. (1996). Children’s perceptions of the
sibling relationship in middle childhood: Connections within and between
family relationships. Personal Relationships, 3(3), 229–239.
Mitchell, T. R., & James, L. R. (2001). Building better theory: Time and the
specification of when things happen. Academy of Management Review, 26(4),
530–547.
Morgeson, F. P., DeRue, D. S., & Karam, E. P. (2010). Leadership in teams: A
functional approach to understanding leadership structures and processes.
Journal of Management, 36(1), 5–39.
Mullins, H. M., Cortina, J. M., Drake, C. L., & Dalal, R. S. (2014). Sleepiness at work: A
review and framework of how the physiology of sleepiness impacts the
workplace. The Journal of Applied Psychology, 99(6), 1096–1112.
Muraven, M., & Baumeister, R. F. (2000). Self-regulation and depletion of limited
resources: Does self-control resemble a muscle? Psychological Bulletin, 126,
247–259.
Muthén, L. K., & Muthén, B. O. (1998-2010). Mplus user’s guide (Sixth Edition). Los
Angeles, CA: Muthén & Muthén.
Nahrgang, J. D., Morgeson, F. P., & Ilies, R. (2009). The development of leader-
member exchanges: Exploring how personality and performance influence
leader and member relationships over time. Organizational Behavior and Human
Decision Processes, 108(2), 256–266.
National Sleep Foundation (2009). Sleep in america poll. Washington, DC: Author.
O’Leary, K. D., Malone, J., & Tyree, A. (1994). Physical aggression in early marriage:
Prerelationship and relationship effects. Journal of Consulting and Clinical
Psychology, 62(3), 594–602.
Preacher, K. J., & Selig, J. P. (2010). Monte Carlo method for assessing multilevel
Mediation: An interactive tool for creating confidence intervals for indirect effects in
1–1-1 multilevel models [Computer software]. <http://quantpsy.org/>.
Preacher, K. J., Curran, P. J., & Bauer, D. J. (2006). Computational tools for probing
interaction effects in multiple linear regression, multilevel modeling, and latent
curve analysis. Journal of Educational and Behavioral Statistics, 31, 437–448.
Preacher, K. J., Zyphur, M. J., & Zhang, Z. (2010). A general multilevel SEM framework
for assessing multilevel mediation. Psychological Methods, 15, 209–233.
Roberts, J. H. (2000). Developing new rules for new markets. Journal of the Academy
of Marketing Science, 28, 31–44.
Rockstuhl, T., Dulebohn, J. H., Ang, S., & Shore, L. M. (2012). Leader-member
exchange (LMX) and culture: A meta-analysis of correlates of LMX across 23
countries. Journal of Applied Psychology, 97(6), 1097–1130.
Ross, L. (1977). The intuitive psychologist and his shortcomings: Distortions in the
attribution process. In L. Berkowitz (Ed.). Advances in experimental social
psychology (Vol. 10, pp. 173–220). New York: Academic Press.
Ross, M., & Fletcher, G. J. O. (1985). Attribution and social perception. In G. Lindzey &
E. Aronson (Eds.), The handbook of social psychology (pp. 73–122). New York:
Random House.
Saper, C. B., Scammell, T. E., & Liu, J. (2005). Hypothalamic regulation of sleep and
circadian rhythms. Nature, 437, 1257–1263.
Scandura, T. A., & Schriesheim, C. A. (1994). Leader–member exchange and
supervisor career mentoring as complementary constructs in leadership
research. Academy of Management Journal, 37(6), 1588–1602.
Scherbaum, C. A., & Ferreter, J. M. (2009). Estimating statistical power and required
sample sizes for organizational research using multilevel modeling.
Organizational Research Methods, 12(2), 347–367.
Schneider, D. J. (1973). Implicit personality theory: A review. Psychological Bulletin,
79(5), 294–309.
Scott, B. A., & Barnes, C. M. (2011). A multilevel field investigation of emotional
labor, affect, work withdrawal, and gender. Academy of Management Journal, 54,
116–136.
Scott, B. A., & Judge, T. A. (2006). Insomnia, emotions, and job satisfaction: A
multilevel study. Journal of Management, 32(5), 622–645.
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, 1061–1086.
Sin, H. P., Nahrgang, J. D., & Morgeson, F. P. (2009). Understanding why they don’t
see eye to eye: An examination of leader-member exchange (LMX) agreement.
The Journal of Applied Psychology, 94(4), 1048–1057.
Smith, E. R., & Miller, F. D. (1979). Salience and the cognitive mediation of
attribution. Journal of Personality and Social Psychology, 37(12), 2240–2252.
Sonnentag, S., Binnewies, C., & Mojza, E. J. (2008). ‘‘Did you have a nice evening?” A
day-level study on recovery experiences, sleep, and affect. Journal of Applied
Psychology, 93, 674–684.
Spanier, G. B., & Thompson, L. (1982). A confirmatory analysis of the Dyadic
Adjustment Scale. Journal of Marriage and the Family, 44(3), 731–738.
Sparrowe, R. T., & Liden, R. C. (1997). Process and structure in leader-member
exchange. Academy of Management Review, 22(2), 522–552.
Spielberger, C. D. (1988). State-trait-anger-expression-inventory. Palo Alto, CA:
Consulting Psychologist Press.
Strawbridge, W. J., Shema, S. J., & Roberts, R. E. (2004). Impact of spouses’ sleep
problems on partners. Sleep, 27(3), 527–531.
Strupp, H. H. (1980). Success and failure in time-limited psychotherapy. Archives of
General Psychiatry, 37, 595–603.
Swanson, L. M., Arnedt, J. T., Rosekind, M. R., Belenky, G., Balkin, T. J., & Drake, C.
(2011). Sleep disorders and work performance: Findings from the 2008 National
Sleep Foundation Sleep in America Poll. Journal of Sleep Research, 20, 487–494.
Sy, T. (2010). What do you think of followers? Examining the content, structure, and
consequences of implicit followership theories. Organizational Behavior and
Human Decision Processes, 113, 73–84.
Taub, J. M. (1977). Napping behavior, activation and sleep function. Waking Sleeping,
1, 281–290.
Tavernier, R., & Willoughby, T. (2014). Bidirectional associations between sleep
(Quality and Duration) and psychosocial functioning across the university
years. Developmental Psychology, 50(3), 674–682.
Tepper, B. J. (2000). Consequences of abusive supervision. Academy of Management
Journal, 43(2), 178–190.
Todorov, A., & Porter, J. (2014). Misleading first impressions: Different for different
images of the same person. Psychological Science, 25, 1404–1417.
Triandis, H. C., Bontempo, R., Villareal, M., Asai, M., & Lucca, N. (1988). Individualism
and collectivism: Cross-cultural perspectives on self-ingroup relationships.
Journal of Personality and Social Psychology, 54(2), 323–338.
Troxel, W. M., Buysse, D. J., Hall, M., & Matthews, K. A. (2009). Marital happiness and
sleep disturbances in a multi-ethnic sample of middle-aged women. Behavioral
Sleep Medicine, 7(1), 2–19.
Truchot, D., Maure, G., & Patte, S. (2003). Do attributions change over time when the
actor’s behavior is hedonically relevant to the perceiver? The Journal of Social
Psychology, 143(2), 202–208.
 C.L. Guarana, C.M. Barnes / Organizational Behavior and Human Decision Processes 141 (2017) 57–73
Turner, L. H., & West, R. L. (2010). Introducing communication theory: Analysis and
application. New York, NY: McGraw-Hill.
Uhl-Bien, M., & Graen, G. (1993). Leadership-making in self-managing professional
work teams: An empirical investigation. In K. E. Clark, M. B. Clark, & D. P.
Campbell (Eds.), The impact of leadership (pp. 379–387). West Orange, NJ:
Leadership Library of America.
Vaidya, J. G., Gray, E. K., Haig, J. R., Mroczek, D. K., & Watson, D. (2008). Differential
Stability and Individual Growth Trajectories of Big Five and Affective Traits
During Young Adulthood. Journal of Personality, 76(2), 267–304.
Van den Bos, K. (2001). Fairness heuristic theory: Assessing the information to
which people are reacting has a pivotal role in understanding organizational
justice. In S. Gilliland, D. Steiner, & D. Skarlicki (Eds.), Theoretical and cultural
perspectives on organizational justice (pp. 63–84). Greenwich, CT: Information
Age.
Van Dongen, H. P., Maislin, G., Mullington, J. M., & Dinges, D. F. (2003). The
cumulative cost of additional wakefulness: Dose-response effects on
73
neurobehavioral functions and sleep physiology from chronic sleep restriction
and total sleep deprivation. Sleep, 26(2), 117–126.
VanderDrift, L. E., & Agnew, C. R. (2014). Relational consequences of personal goal
pursuits. Journal of Personality and Social Psychology, 106(6), 927–940.
Wang, M., Liao, H., Zhan, Y., & Shi, J. (2011). Daily customer mistreatment and
employee sabotage against customers: Examining emotion and resource
perspectives. Academy of Management Journal, 54(2), 312–334.
Watson, D., & Clark, L. A. (1994). The PANAS-X: Manual for the positive and negative
affect schedule-expanded form. Ames: The University of Iowa.
Weiner, B. (1985). An attributional theory of achievement motivation and emotion.
Psychological Review, 92(4), 548–573.
Weiss, H. M., & Cropanzano, R. (1996). An affective events approach to job
satisfaction. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational
behavior, Vol.18 (pp. 1–74). Greenwich, CT: JAI Press.