You Wouldn't Like Me When I'm Sleepy: Leader Sleep, Daily Abusive Supervision, and Work Unit Engagement

Abstract

We examine daily leader sleep as an antecedent to daily abusive supervisory behavior and work unit engagement. Drawing from ego depletion theory, our theoretical extension includes a serial mediation model of nightly sleep quantity and quality as predictors of abusive supervision. We argue that poor nightly sleep influences leaders to enact daily abusive behaviors via ego depletion, and these abusive behaviors ultimately result in decreased daily subordinate unit work engagement. We test this model through an experience sampling study spread over ten work days with data from both supervisors and their subordinates. Our study supports the role of the indirect effects of sleep quality (but not sleep quantity) via leader ego depletion and daily abusive supervisor behavior on daily subordinate unit work engagement.

Full text

You Wouldn’t Like Me W
hen I’m Sleepy: Leader Sleep, Daily
Abusive Supervision, and Work Unit Engagement
Journal:
Academy of Management Journal
Manuscript ID:
AMJ-2013-1063.R3
Manuscript Type:
Revision
Keywords:
Leadership < Organizational Behavior < Topic Areas, Mood and emotions <
Attitudes, Cognitions, and Affect < Organizational Behavior < Topic Areas,
Organizational behavior (General) < Organizational Behavior < Topic Areas
Abstract:
We examine daily leader sleep as an antecedent to daily abusive
supervisory behavior and work unit engagement. Drawing from ego
depletion theory, our theoretical extension includes a serial mediation
model of nightly sleep quantity and quality as predictors of abusive
supervision. We argue that poor nightly sleep influences leaders to enact
daily abusive behaviors via ego depletion, and these abusive behaviors
ultimately result in decreased daily subordinate unit work engagement. We
test this model through an experience sampling study spread over ten
work days with data from both supervisors and their subordinates. Our
study supports the role of the indirect effects of sleep quality (but not sleep
quantity) via leader ego depletion and daily abusive supervisor behavior on
daily subordinate unit work engagement.
Academy of Management Journal

You Wouldn’t Like Me When I’m Sleepy:
Leader Sleep, Daily Abusive Supervision, and
Work Unit Engagement
Christopher M. Barnes
University of Washington
585 Paccar Hall
Seattle, WA 98195
Phone: 1-517-214-0438
Email: chris24b@uw.edu
Lorenzo Lucianetti
University of Chieti and Pescara
Viale Pindaro 42
65127 – Pescara (ITALY)
Phone: +39 085 4537900
Email: llucianetti@unich.it
Devasheesh P. Bhave
Singapore Management University
Phone: +65-6808 5371
Email: dbhave@smu.edu.sg
Michael S. Christian
University of North Carolina at Chapel Hill
Chapel Hill, NC, 27510
Phone: 919.962.2983
Fax: 919.962.4266
Email: mike_christian@unc.edu
Page 1 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Abstract: We examine daily leader sleep as an antecedent to daily abusive supervisory behavior
and work unit engagement. Drawing from ego depletion theory, our theoretical extension
includes a serial mediation model of nightly sleep quantity and quality as predictors of abusive
supervision. We argue that poor nightly sleep influences leaders to enact daily abusive behaviors
via ego depletion, and these abusive behaviors ultimately result in decreased daily subordinate
unit work engagement. We test this model through an experience sampling study spread over ten
work days with data from both supervisors and their subordinates. Our study supports the role of
the indirect effects of sleep quality (but not sleep quantity) via leader ego depletion and daily
abusive supervisor behavior on daily subordinate unit work engagement.
Page 2 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Abusive supervision – the “sustained display of hostile verbal and nonverbal behavior,
excluding physical contact” of supervisors toward subordinates, as perceived by subordinates
(Tepper, 2000: 178) – has pervasive and negative effects on employees, their work outcomes,
and organizations. Over the past decade and a half, researchers have investigated the deleterious
effects of abusive supervision, particularly on subordinate affect, attitudes, motivation, and job
performance (for a recent meta-analysis, see Schyns & Schilling, 2013). Thus, understanding
why and under what circumstances supervisors might be abusive is paramount for researchers
and practitioners interested in improving a variety of organizationally relevant outcomes.
However, researchers have recently noted two important limitations to theory explaining
abusive supervision. First, as noted by Tepper (2007) and again in Tepper, Moss, and Duffy
(2011), theory and research on abusive supervision has focused much more on outcomes of
abusive supervision than antecedents. Although the outcomes of abusive supervision are
important, a sound understanding of its causes is necessary to enable management theory to
guide managers towards reducing abusive supervision.
A second limitation is that research on abusive supervision has typically taken a static
approach, implicitly assuming that some supervisors engage in abusive supervision and some do
not, rather than examining whether this behavior fluctuates within a given supervisor. This
assumption is highlighted by the word “sustained” in the definition of abusive supervision.
Tepper (2007: 265) explicitly notes that “abusive supervision involves continuing exposure to
hierarchical mistreatment—a boss who has a bad day and takes it out on his or her subordinates
by exploding at them would not be considered an abusive supervisor unless such behavior
became a regular feature of his or her repertoire.” Thus, although research has confirmed the
proposition that some supervisors are often abusive, whereas others are usually not abusive, this
Page 3 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

definition is restricted to a leader’s “style,” or behaviors on average. Thus, much of the research
on abusive supervision has developed around the examination of “abusive supervisors,”
precluding the possibility that any leader could be high in abusive supervision on one day and
low on abusive supervision on another day.
With our research, we offer a complementary perspective to the between-persons
paradigm of abusive supervisors by examining abusive supervisory behaviors, which we argue
fluctuate within-person on a daily basis. Emerging evidence suggests that leaders might be more
(or less) abusive on some days than on others. Johnson, Venus, Lanaj, Mao, and Chang (2012)
found that abusive supervisory behavior varied more within-supervisors than it did between
supervisors. In other words, supervisors exhibited more within-person variation in abusive
behavior than was observed for comparisons between supervisors. Building from this research,
we expand Tepper’s (2000) definition of abusive supervision to examine abusive supervisory
behaviors, defined as any display of hostile verbal and nonverbal behavior, excluding physical
contact. With our research, we examine how these behaviors are likely to vary on a day-to-day
basis, and refer to them henceforth as daily abusive supervision. We posit that not only is there
potentially more predictive power within individuals than between individuals, but a less static
view of abusive supervision allows for interventions that can potentially apply to a broad set of
employees. This opens options beyond staffing for managing abusive supervision. Interventions
aimed at improving daily self-control and mood, such as breaks, positive events, or even a
mindfulness exercise, could potentially set the stage for low abusive supervision on a given day.
However, an important question remains unanswered: What factors that were previously
assumed to be simply noise may account for daily abusive supervision?
Page 4 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Accordingly, the purpose of this paper is to take a within-person approach to extend
theory on abusive supervision by examining daily antecedents and outcomes. Specifically, we
draw from theory on ego depletion to examine nightly sleep quantity and quality as antecedents
of daily abusive supervisor behavior. Moreover, our conceptual framework suggests that when
supervisors are depleted and thus abusive, there will be regulatory consequences that “trickle-
down,” to the work unit, sapping their collective work engagement, or the willingness of the
members of the unit to self-regulate by investing energy in their work tasks. Thus, we expand the
abusive supervision literature by hypothesizing that daily abusive supervision reduces unit work
engagement. We move beyond traditional static approaches to studying the antecedents of
abusive supervision by proposing that daily abusive supervisor behavior varies in part on the
quantity and quality of sleep the night before. Moreover, this includes a crossover view, in that
leader sleep influences work unit engagement. Consistent with our theorizing, we test our model
of sleep and daily abusive supervisor behavior using a sample of supervisor-led work units and
an experience sampling method research design.
ABUSIVE SUPERVISION: MOVING TO A DAILY APPROACH
As noted by Tepper (2007) and Tepper, Moss, and Duffy (2011), theory and research on
abusive supervision has focused much more on outcomes of abusive supervision than
antecedents. However, this nascent area of research has been helpful in beginning to explore
important antecedents such as justice, subordinate characteristics, and diversity (Aquino, Grover,
Bradfield, & Allen, 1999; Aryee, Chen, Sun, & Debrah, 2007; Tepper, Duffy, Henle, & Lambert,
2006; Tepper, Moss, & Duffy, 2011). This research has begun to open the topic of antecedents to
abusive supervision, although from a relatively static, cross-sectional point of view. However, no
studies have considered factors that (a) vary on a daily basis such as sleep and (b) are proximally
Page 5 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

aligned with self-regulatory models, which we argue can expand our understanding of what
causes abusive supervision. Models of sleep and work require a daily focus in order to be
properly specified at the correct level of analysis (Klein & Kozlowski, 2000). Although there are
clearly important relationships between abusive supervision and other constructs at the between-
person level of analysis (c.f. Tepper, 2007), the baseline assumption that there is nothing of
importance in the domain of abusive supervision occurring at the within-level of analysis may be
a model misspecification. In the specific case of leader sleep and work unit engagement, both the
conceptual development and the data are consistent with daily variance as the focus.
Leadership research has long contended that leadership occurs within a specific context,
and a specific set of circumstances (c.f. DeRue, 2011). For example, leaders can switch from a
directive set of behaviors to an empowering set of behaviors (Lorinkova, Pearsall, & Sims,
2013). Although the leadership literature has focused on long term changes in leader behavior,
human behavior in general also varies on a much shorter time scale, based on dynamic factors
such as mood, self-control, salient goals, and activated identities (Barnes et al., 2011; Dalal,
Lam, Weiss, Welch, & Hulin 2009; Leavitt et al., 2012; Scott et al., 2012; Venus, Stam, & van
Knippenberg, 2013). Leadership is determined in part by dynamic variables such as mood and
identity (Johnson et al., 2012; Venus, Stam, & van Knippenberg, 2013), and should also
naturally vary over time on similar time scales. Indeed, Johnson et al. (2012) found that abusive
supervision varied on a daily basis, and that this variance was greater than between-persons
variance. Thus, abusive behaviors might be linked to within-person variables that vary over
time. The substantial body of work of within-person variability in affect (e.g., Dalal, et al, 2009;
Glomb, Bhave, Miner, & Wall, 2011) suggests that no person is always pleasant or always
unpleasant. Building from this premise, individuals may be abusive on one day but not on others.
Page 6 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Johnson and colleagues (2012) advanced theory by showing the degree to which leaders
were (in)consistent from day-to-day in their level of abusive supervisory behavior; however,
Johnson and colleagues (2012) did not investigate day-level predictors of this daily variance.
Thus, their results pave the way for further research taking a daily view of the factors that lead
to, and result from, daily abusive supervision. In order to enhance the richness of the abusive
supervision literature in this new direction, we examine “daily abusive supervision”, turning our
focus to abusive behaviors, rather than the leader’s style. This construct is defined the same as
the original abusive supervision construct was defined by Tepper (2000), with the exception that
we refer to behaviors—which are variable within-person—rather than a supervisor’s preferred
method of supervision, and we remove the constraint that the behavior is sustained over
prolonged periods of time. This definition enables us to investigate daily fluctuations. In other
words, we argue that abusive behavior engaged in on a single day is still abusive and meaningful.
Thus, in addition to being associated with a leadership style, supervisory abuse is a behavior that
can vary on a daily basis. Our work links back to the larger topic of abusive supervision, but
allows for growth in a useful direction. We hope that it opens further research questions beyond
the model that we test in our paper.
SLEEP, EGO DEPLETION, AND ABUSIVE SUPERVISION
Leaders may often experience situations or events that create tempting impulses or urges
to engage in abusive supervisory behavior in their interactions with subordinates. Frustration
with a lack of progress on a project or with interpersonal conflict may create an urge to yell or
speak uncivilly towards a given subordinate (c.f., Tepper, Moss, & Duffy, 2011). Encountering a
mistake made by an employee might create an impulse to publicly belittle the employee. Having
ideas criticized by an employee might induce the urge to coerce the subordinate into silence. We
Page 7 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

argue that leaders sometimes struggle to control these impulses, and that a primary reason for
their inability to overcome them is failures in self-regulation.
Self-regulation is the psychological process by which counter-normative urges and
impulses are controlled (Muraven & Baumeister, 2000). Ego depletion theory describes how the
ability to exert self-regulation waxes and wanes over time. According to ego depletion theory, all
forms of self-regulation draw from a single finite pool of resources (Baumeister, Bratslavsky,
Muraven, & Tice, 1998; Muraven & Baumeister, 2000). Engaging in acts requiring self-
regulation depletes this pool, leaving them less able to do so until the resources are recovered.
Recent research indicates that ego depletion leads people to be especially likely to fail in
resisting temptations to engage in negative behaviors (c.f., Gino, Schweitzer, Mead, & Ariely,
2011). Examples of such behavior induced by ego depletion include lying (Mead, Baumeister,
Gino, Schweitzer, & Ariely, 2009), cheating (Christian & Ellis, 2011), deception (Welsh, Ellis,
Christian & Mai, 2014), and other unethical behavior (Barnes, Schaubroeck, Huth, & Ghumman,
2011). Moreover, the capacity for self-regulation is dynamic, and can be depleted by a range of
factors (for a meta-analysis, see Hagger, Wood, Stiff, & Chatzisarantis, 2010).
Recent extensions to ego depletion theory indicate an important antecedent to self-
regulation that is relevant to all employees: sleep. Self-regulation may be affected by both sleep
quantity – the amount of time an individual spends in a sleeping state – and by sleep quality –
which refers to difficulty of falling asleep and staying asleep (Barnes, 2012). Barnes further
notes that sleep quantity and quality have parallel additive effects on self-regulation. This is in
line with the proposition of Baumeister, Muraven, and Tice (2000) that sleep is important for the
recovery of physiological resources involved in self-regulation. Moreover, sleep physiologists
Page 8 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

have found that a lack of sleep leads to socially inappropriate behavior (Horne, 1993), suggesting
the possibility of impaired self-control.
Emerging physiological evidence supports this view, suggesting that sleep deficiencies
impair the functioning of structures in the brain that are critical to self-regulation. A growing
literature in neurophysiology indicates that self-regulation relies disproportionately on the
prefrontal cortex and amygdala regions of the brain (Banks, Eddy, Angstadt, Nathan, & Phan,
2007; Beauregard, Levesque, & Bourgouin, 2001; Chuah et al., 2010; Nilsson et al., 2005;
Ochsner et al., 2004). These regions are fueled by glucose (Fairclough & Houston, 2004), which
is utilized throughout the day and replenished during sleep. Brain-imaging studies indicate a
decrease in cerebral metabolism under conditions of sleep deprivation and insomnia, most
notably in the prefrontal cortex (Altena et al., 2008; Thomas et al., 2000). Thus,
neurophysiological research indicates that sleep is an important determinant of self-regulation.
Given the importance of self-regulated behavior in organizations, its connection with
sleep has recently been targeted by management researchers. Christian and Ellis (2011) found
that compared to sleeping 6 hours or more, nurses sleeping fewer than 6 hours in a night had
reduced resources and increased organizational deviance the next day. Barnes et al. (2011)
similarly found that a lack of sleep led to resource depletion, producing unethical behavior.
Ghumman and Barnes (2013) found that a lack of sleep led to impairments in the suppression of
prejudice. Barber, Barnes, and Carlson (2013) found that sleep difficulties led to decrements in
self-regulation, in turn undermining attempts at social desirability. Wagner, Barnes, Lim, and
Ferris (2012) found that a lack of sleep led to an increase in cyberloafing at work.
Every day and night, employees make choices between allocating time toward sleep
versus other competing activities such as time spent working, with family, or recreating (Barnes,
Page 9 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Wagner, & Ghumman, 2012). Consistent with this, Knutson, Rathouz, Yan, Liu, and Lauderdale
(2007) conducted a large scale study of sleep, and found that the within-person standard
deviation exceeded the between-person standard deviation. Recent management studies have
found day-level relationships between sleep (quantity and quality) and several workplace
phenomena, including affect, job satisfaction, unethical behavior, surface acting, and time spent
working (Barnes et al., 2011; Barnes et al., 2012; Christian & Ellis, 2011; Scott & Judge, 2006;
Sonnentag, Binnewies, & Mojza, 2008; Wagner et al., 2014; Welsh, et al., 2014).
Specific to the topic of daily self-regulation, Barnes et al. (2011) and Christian and Ellis
(2011) extended theory on ego depletion to indicate that sleep varies along with self-regulatory
capacity on a daily basis. Self-regulatory resources are depleted daily, and replenished during
sleep. Thus, a lack of sleep in a given night leaves an individual with depleted self-regulation the
next day. Consistent with this reasoning, Barnes et al. (2011) provided evidence from a diary
study showing daily relationships between sleep quantity/quality and self-regulation, as did
Christian and Ellis (2011), who manipulated one night of sleep deprivation. Thus, we expect
daily leader sleep quantity and quality to influence the leader’s ego depletion on the next day.
Hypothesis 1a: Daily leader sleep quantity is negatively related to daily leader ego depletion.
Hypothesis 1b: Daily leader sleep quality is negatively related to daily leader ego depletion.
As we note above, leaders face many temptations to engage in abusive behavior toward
subordinates, especially when they experience stress, frustration, and difficulties at work.
Suppressing those temptations and behaving in a civil manner requires self-regulation. As noted
in Hypotheses 1a and 1b, we expect sleep to influence self-regulation. Thus, we contend that
Page 10 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

sleep on a given night (both quantity and quality) will influence abusive supervisory behavior the
next day, and that ego depletion will mediate this effect.
Although previous research has not examined this relationship directly, it does lend
indirect support. Horne (1993) found that sleep deprivation led to an increase in interpersonally
inappropriate behavior. Kahn-Greene, Lipizzi, Conrad, Kamimori, & Killgore (2006) found that
sleep deprivation leads to frustration, lack of willingness to accept blame, increased tendency to
blame others, and weakened inhibition of aggression. As reviewed by Tepper (2007), several
studies show that displaced feelings of aggression are a likely antecedent of abusive supervision
(Aryee, Chen, Sun, & Debrah, 2007; Hoobler & Brass, 2006; Tepper, Duffy, Henle, & Lambert,
2006). Thus, leaders who have weakened inhibition from a nightly sleep deficiency and are
frustrated or blame others are likely to engage in abusive supervision. Indeed, Barnes (2012)
argued that low sleep quantity and poor sleep quality would lead to workplace incivility.
Accordingly, drawing from an ego depletion approach, we hypothesize that daily sleep quantity
and quality will negatively influence daily abusive supervision through the mediator of ego
depletion.
Hypothesis 2: Daily leader ego depletion is positively related to daily abusive supervisor
behavior.
Hypothesis 3: Daily leader ego depletion mediates the effects of (a) daily leader sleep quantity
and (b) daily leader sleep quality on daily abusive supervisor behavior.
EFFECTS ON UNIT WORK ENGAGEMENT
Page 11 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Abusive supervision involves subordinate perceptions of mistreatment by their supervisor
and should thus affect subordinate outcomes. We focus, in particular, on a motivational
outcome—daily unit level work engagement—for three reasons. First, our ego depletion
framework specifies the critical role of self-regulatory resources underlying supervisor behavior,
and we extend this line of reasoning to suggest that follower behavior is similarly underpinned
by motivation and willingness to allocate self-regulatory energy to tasks. Second, subordinate
motivation is proximal psychologically to abusive behavior and thus likely to be a strong
psychological outcome for subordinates experiencing abusive supervision (c.f., Schyns &
Schilling, 2013). Third, a primary function of leadership is to instill motivation and meaning in
their group of followers (e.g., Avolio, 1990), thus work engagement is conceptually linked to
leader behavior (Christian et al., 2011; Macey & Schneider, 2008).
Work engagement is a state of cognitive, emotional, and physical investment in one’s
personal experience or performance of work (Christian et al., 2011; Kahn, 1990; 1992; Rich,
LePine, & Crawford, 2010). In a review of the work engagement literature, Bakker (2014) notes
that work engagement fluctuates on a daily basis, and that this daily fluctuation is driven in part
by negative employee experiences at work. Indeed, several recent articles empirically support the
idea that daily fluctuation in engagement is meaningful and predictable (Bakker & Despoina,
2009; Breevaart, Bakker, & Demerouti, 2014; Christian, et al., 2011; Lanaj et al., 2014;
Sonnentag, 2003; Sonnentag et al., 2009).
Moreover, work engagement occurs collectively. Workers who are treated poorly and
experience disengagement as a result will be likely to engage in collective sensemaking
processes whereby they affect each other’s daily work engagement (Costa, Passos, & Bakker,
2014). We conceptualize work engagement at the unit level, following the lead of others (e.g.,
Page 12 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Barrick, Thurgood, Smith, & Courtright, in press, Harter, Schmidt, & Hayes, 2002; Salanova,
Agut, & Piero, 2005). Unit-level engagement is a conceptually appropriate level of analysis for
examining outcomes of leadership, because a leader’s influence tends to produce shared
responses among subordinates (e.g., Christian, Christian, Garza, & Ellis, 2012; George, 2000).
Subordinates within the same work unit are likely to have similar levels of exposure to abusive
supervision on a given day, such that multiple subordinates will be exposed to that same
behavior. Moreover, group members tend to converge in their affect, attitudes, and behavior
(Bhave, Kramer, & Glomb, 2010; Duffy, Shaw, & Stark, 2000; Felps, Mitchell, Hekman, Lee,
Holtom, & Harman, 2009; Sy & Choi, 2013), as they interact and make sense of social and
environmental information as a group (Salancik & Pfeffer, 1977).
Indeed, unit level engagement has been shown to converge among unit-members
(Salanova et al., 2005). Work unit engagement has beneficial effects on important outcomes such
as firm performance (Barrick, et al., in press), service climate, unit performance, and customer
loyalty (Salanova, Agut, & Piero, 2005), as well as having cross level effects on individual
burnout (Bakker, van Emmerik, & Euwema, 2006). This is further consistent with our topic of
abusive supervision; recent research illustrates aggregate subordinate responses to abusive
supervision (Priesemuth, Schminke, Ambrose, & Folger, in press).
Work engagement is related to leadership to the extent that effective leaders help
subordinates to view their work as meaningful and valuable, and to attach their identities to the
work itself (Bono & Judge, 2003; Grant, 2012). Thus, leader behavior may influence the extent
to which subordinates feel personally invested in the work they perform, especially to the degree
to which a leader is fair and trustworthy and engenders feelings of psychological safety (Kahn,
1990; Macey & Schnieder, 2008). Abusive supervision is inconsistent with signals of
Page 13 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

competence, value, and respect (Mayer, Thau, Workman, Dijke, & De Cremer, 2012). By
providing an experience that employees will find aversive, abusive supervision should leave
employees more likely to withdraw than to engage themselves heavily in their work. Related
research indicates that CEO leadership influences collective organizational engagement (Barrick
et al., in press). Although we examine work units rather than firms as a whole, the logic is similar
in that leaders influence the engagement of subordinates within their collective.
We integrate unit work engagement into our model of sleep and abusive supervision.
Specifically, as indicated in Hypothesis 3, sleep quantity and quality will negatively influence
daily abusive supervision through the mediating mechanism of ego depletion. As indicated in our
logic above, daily abusive supervision will negatively influence daily unit engagement.
Therefore, we expect a relationship with two levels of serial mediation, such that the effects of
leader sleep are transmitted to subordinate unit engagement first through leader ego depletion
and then through daily leader abusive supervision. Figure 1 depicts the full conceptual model.
Hypothesis 4: Daily abusive supervision will be negatively related to daily unit work
engagement.
Hypothesis 5a: Daily leader ego depletion and daily abusive supervisor behavior will serially
mediate the daily leader sleep quantity to daily unit work engagement relationship.
Hypothesis 5b: Daily leader ego depletion and daily abusive supervisor behavior will serially
mediate the daily leader sleep quality to unit work engagement relationship.
METHOD
Sample
Page 14 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

We drew our participants from Amadeus - Bureau Van Dijk (https://aida.bvdinfo.com/), a
database of public and private firms which include domestic and multinational firms in Italy. We
contacted managers from these organizations and informed them about our study. After
managers expressed their respective organizations’ willingness to participate in the study, we
informed employees of these organizations via e-mail about the project and invited them to sign
up individually for the study. We offered feedback about the study results after completion of
data collection as an incentive for participation. Participants were from a variety of industries
and occupations, including accounting, supply chain, operation management, human resources,
and marketing in the industries of banking, information technology, and health care. 99
supervisors agreed to allow the administration of surveys, and completed the surveys themselves.
Their workgroups ranged from 3-8 members, with a mean of 4.6 per group. We received
completed questionnaires from 261 subordinates, representing a response rate of 57%. Across
groups, response rates ranged from 25% to 100%.
For supervisors, 28% of respondents were female; 6% were between 18 and 30 years old,
25% between 31 and 40, 33% between 41 and 50 and 35% were older than 51 (mean 46 years,
SD 9.8). Twenty two percent of supervisors worked within the current position for less than 2
years, 15% between 2 and 4 years, 24% between 4 and 6 years, and 38% worked more than 6
years in the current positions. Mean supervisor tenure with their organizations was 7 years (SD
5.8).
Forty percent of subordinates were female; 26% were between 18 and 30 years old, 36%
between 31 and 40, 25% between 41 and 50 and 13% were older than 51 (mean 38 years, SD
9.8). Twenty eight percent of subordinates worked with the current supervisor for less than 1
year, 47% between 2 and 4 years, 12% between 4 and 6 years, and 13% worked more than 6
Page 15 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

years with the current supervisor (mean 3.6 years, SD 3.3). In the sample, on average
subordinates interacted with their current supervisors “rarely” 2%, “once every few weeks” 9%,
“once per day” 13%, “a few times per day” 23% and “several times per day” 53%.
Overall, we received 826 out of 990 possible supervisor surveys (83.43% response rate)
and 2148 out of 2610 possible subordinate surveys (82.30%) out of those participating. Ten
supervisors either provided less than 2 surveys or did not meet the minimum requirement of
having at least one subordinate who completed at least 2 surveys, and were thus removed from
the data. As noted below, subordinate-days in which subordinates had either “none” or “little”
contact with their supervisors were left out of the data. After the available daily subordinate
surveys were matched with the available daily supervisor surveys, and were aggregated to the
supervisor-level, it yielded a final sample of 606 unit-days nested within 88 supervisors.
Procedures
Participants were recruited through contacts with their organizations. Individuals who
indicated an interest in participating were presented with the informed consent document. This
provided instructions for the study, as well as assurances of confidentiality. Surveys were
provided in Italian, the native language of the participants. To develop the Italian version of the
surveys, we followed the translation-back translation procedures outlined by Brislin (1986).
Supervisors first completed a baseline survey. In order to capture daily variance in the
constructs of our model, the rest of the study used interval-contingent experience-sampling
methodology (see Alliger & Williams, 1993; Wheeler & Reis, 1991). Similar to the majority of
experience sampling research in the management literature, we chose a 2 week period. This is
consistent with Reis and Wheeler’s (1991) suggestion that two weeks represents a generalizable
sample of individuals’ lives. During the two weeks of the study (including only workdays),
Page 16 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

participants were asked to complete 1 survey per day. Supervisors were asked to complete their
surveys at the beginning of their workday (measuring supervisor sleep and ego depletion), and
subordinates were asked to complete their surveys at the end of their workday (all other
measures). This allowed for temporal precedence, with the independent variables and first stage
mediator (supervisor sleep quantity and quality, and supervisor ego depletion) measured earlier
in the workday than the second stage mediator and outcome (abusive supervisor behavior and
subordinate work engagement).
Measures
Supervisor sleep. Supervisor sleep was measured with the Pittsburgh Sleep Diary (Monk
et al., 1994). Participants were asked the time at which they went to bed, how long it took to fall
asleep, what time they woke in the morning, and how long they were awake after initially falling
asleep. Time awake after initially falling asleep is referred to in the sleep physiology literature as
“wakefulness after sleep onset” (WASO). In the instructions to participants, in the WASO
question, participants were provided with an example to help them understand the meaning (“For
example, if you were asleep until 1am, woke at 1am and fell back asleep at 1:20am for the rest of
the night, your answer would be 20 minutes”). These times were used to calculate the number of
minutes spent asleep, which was how we operationalized sleep quantity. Previous research
indicates that this measure of sleep quantity correlates with objective measures of sleep quantity
(Barnes et al., 2011). We reverse-coded WASO, which captures interruptions to sleep, as our
operationalization of sleep quality. This follows the same approach as previous research in
management measuring sleep quality with interruptions to sleep (Wagner et al., 2012).
Daily Leader Ego Depletion. To measure daily ego depletion, we used the 5 item scale
that Lanaj et al. (2014) selected to measure ego depletion in a diary study format. These items
Page 17 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

originally came from Twenge et al. (2004). Participants were instructed to indicate the degree to
which they agreed with each item on a 5 point Likert scale, in which 1=very slightly or not at all
and 5=very much. A sample item is “My mental energy is running low.” Average coefficient
alpha for this scale was .92.
Daily Abusive Supervisor Behavior. To measure daily abusive supervisor behavior, we
used the 5 item scale that Johnson et al. (2012) specifically developed to measure daily abusive
supervision. Participants were instructed to indicate the “frequency with which your supervisor
engaged in each of the 5 behaviors today at work”, using a 6 point Likert scale provided in
increments of 1 occasion, in which 1=never and 6=5 or more. A sample item is “Behaved in a
nasty or rude manner toward a group member”. Average coefficient alpha for this scale was .78.
Daily Unit Work Engagement. Work engagement was measured with 3 items drawn
from Schaufeli, Bakker, and Salanova (2006) and validated for use in a daily survey context by
Lanaj et al. (2014), one item for each conceptual dimension of work engagement, physical,
emotional, and cognitive (cf. Rich et al., 2011). In their pilot work, Lanaj et al. (2014) found that
the shortened version of the Schaufeli et al. (2006) work engagement scale correlated with the
full version at r=.83 (p<.01).The items were reworded to focus on daily engagement, and
participants were asked to indicate the degree to which they agreed with the items on a 5 point
Likert scale in which 1=strongly disagree and 5=strongly agree.
A recent meta-analysis indicates that when assessing and aggregating affectively-laden
variables (i.e., work engagement) at the group-level, a direct consensus model may be more
appropriate than a referent shift model (Wallace, Edwards, Paul, Burke, Christian, & Eissa,
2013). Wallace et al. argue that an employee’s assessment of the work environment relative to
their personal affective experience is more accurate than an employee’s assessment of the
Page 18 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

affective experiences of others inside and outside of one’s workgroup. As Wallace and
colleagues (2013) recommend, direct consensus models are appropriate for constructs that have
affective components, withstanding sufficient within-group agreement statistics. Thus, we used
items that measured individual engagement and then aggregated to the work unit level
(aggregation information is provided later in this manuscript). An example item for cognitive
engagement is “Today, I was immersed in my work.” Average coefficient alpha for this measure
was .86.
Control and Cutoff Measures. Participants who have either no contact or only a little
contact with their supervisor in a given day possess insufficient information to rate the abusive
behavior of their supervisor on that day. Accordingly, we asked participants “How much contact
did you have today with your supervisor?”, with responses on a 5 point Likert scale in which
1=none, 2=little, 3=a moderate amount, 4=quite a bit, and 5=a high amount of contact. We
included subordinate responses only on days in which their contact with their supervisor was 3 or
greater on this scale, analogous to the approach previously used by Pugh, Groth, and Hennig-
Thurau (2011) and Bono et al. (2007) to target the questionnaire toward those with sufficient
information to answer it.
Trait anxiety has been linked to both sleep problems (LeBlanc, Mérette, Savard, Ivers,
Bailargeon, & Morin, 2009) and negative behaviors from leaders (Kant, Skogstad, Torsheim, &
Einarsen, 2013). Thus, in order to eliminate anxiety as a confound in the relationship between
leader sleep and abusive supervision, we included leader trait anxiety as a control variable.
Following Wagner, Barnes, and Scott (2014), we used a 4 item measure of anxiety drawn from
MacKinnon, Jorm, Christensen, Korten, Jacomb, and Rodgers (1999). Participants rated on a 5
Page 19 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

item Likert scale the degree to which each of 4 adjectives described them “on average”. Sample
adjectives include “nervous” and “distressed”. Coefficient alpha for this scale was .74.
ANALYSIS
Given the multilevel nature of our model, the data collected to test our model included
nesting that violates assumptions of independence of observations required for ordinary least
squares regression analyses. Accordingly, we conducted our analyses in multilevel format using
Mplus (Muthén & Muthén, 1998-2010). Furthermore, in order to test the proposed serial
mediation, we performed multilevel path analysis (MacKinnon, 2008; Preacher, Zyphur, &
Zhang, 2010). Specifically, in order to test the serial mediation, in our model daily leader sleep
quantity and quality were the independent variables, daily leader ego depletion was the first-
stage mediator, daily abusive supervision was the second-stage mediator, and unit work
engagement was the dependent variable. We test and report mediation through a test of the
statistical significance of the indirect effect and its associated confidence interval (MacKinnon,
2008). The data consisted of 2 levels. The lowest level (Level 1) comprised daily unit ratings,
leader sleep and ego depletion, which were nested within leader our leader control variable-
(Level 2).
In order to empirically justify aggregation of subordinate ratings of a given leader on a
given day and aggregation of subordinate daily ratings to unit daily ratings, we conducted ICC
analyses. This analysis indicates what proportion of the variance is accounted for by the group
level, and whether or there is significant nesting. For leader daily abusive supervision ICC(1)=
.43 (p<.01) and ICC(2) = .91, F=1.87. For daily unit work engagement, ICC(1) = .48 (p<.01) and
ICC(2) = .65, F=2.81. These values all support the aggregation we indicated in our conceptual
development.
Page 20 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

RESULTS
Table 1 reports the descriptive statistics and the correlations at the within-person level.
As a preliminary step in the analysis, partitioning of the variance indicated that 57% of the
variance in abusive supervision was within-supervisors, and 54% of the variance in work
engagement was within-work units (i.e., the work unit-day level of analysis). Table 2 reports the
results of our hypotheses and the multilevel path analysis. We proposed that daily leader sleep
quantity (H1a) and sleep quality (H1b) will be negatively related to leader ego depletion. Results
provided support for H1b: daily sleep quality (γ = -.13, p < .05) was negatively related to leader
ego depletion; there was no statistically significant relationship between daily sleep quantity and
leader ego depletion (γ = .02, p > .05).
Hypothesis 2, which posited that daily leader ego depletion positively relates to daily
abusive supervision, received empirical support (γ = .35, p < .01). Hypothesis 3 posited that daily
leader ego depletion will mediate the relationship between daily sleep quantity (H3a) and daily
sleep quality (H3b) and daily abusive supervisor behavior. Hypothesis 3a was not supported.
Hypothesis 3b was supported: the indirect effect of daily sleep quality on daily abusive
supervisor behavior via daily leader ego depletion was significant (ab = -.04, p<.05; 95% CI [-
.084, -.003]).
Hypothesis 4, which posited that daily abusive supervision will be negatively related to
daily unit work engagement, was supported (γ = -.45, p < .01). Hypothesis 5 posited that daily
leader ego depletion and daily abusive supervision will serially mediate the relationship between
daily sleep quantity (H5a) and daily sleep quality (H5b) and daily unit work engagement.
Hypothesis 5a was not supported but Hypothesis 5b received support: the indirect effect of daily
sleep quality on daily abusive supervision via daily leader ego depletion (the first-stage
Page 21 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

mediator) was statistically significant (ab = -.15, p<.01; 95% CI [-.229, -.080]); notably, the
indirect effect of daily sleep quality on daily unit work engagement via both daily leader ego
depletion (the first-stage mediator) and daily abusive supervision (the second-stage mediator)
was also significant (ab = .02, p<.05; 95% CI [.000, .039]; 90% CI [.003,.036]. The estimate (ab)
of the indirect effect for the serial mediation is statistically significant at the 5% level of
significance (i.e., p<.05), and the lower bound of the 95% confidence interval is a non-zero
positive value beyond the three decimal places, which, as such does not include zero.
Nevertheless, as an additional check, we also report the 90% confidence interval for this indirect
effect.
Overall, the results indicated that daily leader ego depletion mediated the relationship
between daily leader sleep quality and daily abusive supervision. Furthermore, there was
evidence of serial mediation such that the relationship between daily leader sleep quality and unit
work engagement was mediated by leader ego depletion and daily abusive supervision. We did
not observe these indirect effects for the daily sleep quantity and unit work engagement
relationship.
Supplementary Analyses
To provide additional insight on different functional forms of abusive supervisory
behaviors, we performed a number of supplementary analyses. Specifically, we considered the
variability in abusive supervision (as indicated by the standard deviation of abusive supervision)
and the trend in abusive supervision over the study period (as indicated by the linear trend of
abusive supervision). Although we do not have any a priori hypotheses for these analyses, we
explore them to allow the possibility of finding useful information. We focus solely on the
Page 22 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

variables of interest in these analyses and exclude the various potential combinations of control
variables.
As a first step, we examined whether abusive supervision variability is related to unit
work engagement above and beyond the mean level of abusive supervision. Results indicated
that abusive supervision variability was related to unit work engagement (γ = .09, p< .10) only at
less stringent cutoff level of an alpha value of .10. Next, to understand whether units may react to
the same daily abusive supervision differently depending on the predictability of the behavior,
we examined abusive supervision variability as a moderator of the abusive supervision – unit
work engagement relationship (see Table 3). Results indicated that abusive supervision
variability may play such a moderating role (γ =.13, p < .10) such that unit work engagement
was highest when both mean and variability in abusive supervision were low, albeit again only at
a less stringent alpha cutoff level (see Figure 2).
In a third exploratory analysis we examined the trend in abusive supervision as a
predictor of unit work engagement to assess whether subordinates respond differently if abusive
supervision is getting worse (or lessening) over time (i.e., within the two-week time period of our
study). There were no statistically significant effects for the trend of abusive supervision as a
predictor of unit work engagement (γ = -.04, p > .05). In a related analysis, to assess whether
units may respond to the same level of daily abusive supervision differently depending on the
trend in abusive supervision, we examined the trend as a moderator of the daily abusive
supervision – unit work engagement relationship. Results indicated that the moderator effects
were not significant (γ = -.09, p > .05).
Finally, given the null findings associated with sleep quantity, we examined the
interactive effects of sleep quantity and sleep quality on daily abusive supervision. Results
Page 23 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

indicated that the interaction term was not statistically significant (γ = .08, p > .05). We also
examined the interactive effects of sleep quantity and sleep quality on daily ego depletion, a
more proximal outcome. Results indicated that the interaction term was statistically significant,
although at a less stringent cutoff of an alpha of .10 (γ = -.14, p < .10), such that daily ego
depletion was lowest when both sleep quantity and sleep quality were high.
Overall, the exploratory analyses revealed few statistically significant findings.
Moreover, there was no clear and meaningful pattern that emerged from those analyses. It is
plausible that a larger sample, or more specifically, data collected over a longer timeframe, may
be necessary to detect trends and variability in abusive supervision, and to more clearly isolate
interactive effects of sleep quantity and quality.
DISCUSSION
We used an experience sampling design to examine the daily relationships among
supervisor sleep, subsequent supervisory abusive behaviors towards subordinates, and
subordinate outcomes. The results generally supported our hypotheses with regards to sleep
quality, but not sleep quantity. Supervisor sleep quality was associated with daily abusive
behaviors through the mediator of daily ego depletion. Supervisor sleep quality was also linked
indirectly—via daily leader ego depletion and daily abusive supervisor behavior—to subordinate
unit work engagement. Our results have several theoretical and practical implications.
To begin, we challenge the prevailing static viewpoint that assumes that leaders are either
abusive to some degree or not abusive at all. Whereas the majority of research has considered
abusive supervision to be a chronic factor—much like a trait or a consistent style—our study
suggests that supervisors vary in their level of abusive behavior on a daily basis. Our results
stand with those of Johnson and colleagues (2012) as the only two studies that have tested this
Page 24 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

approach. Our findings add to the literature by suggesting that in order to accurately describe
abusive supervision, theory and research should focus as much on “momentary” abuse as it has
on “sustained” abuse. Thus our research moves the literature on abusive supervision, and on
leadership more generally, forward by demonstrating the importance of a daily perspective in
understanding leader behavior.
Further, our results suggest at least two reasons why fluctuations in abusive behavior are
theoretically important. First, daily abusive behaviors are associated with fluctuations in
supervisor sleep quality. This is unlikely to be found in a between-subjects research design,
because the effects of nightly sleep quality are proximal and most likely to have effects on
behavior the following day. More generally, by linking sleep to leader behavior, we contribute to
the very small body of research on antecedents to abusive supervision, arguing that the
exogenous causes of abusive supervision may vary on a daily basis. Our study explains one
reason why leaders exhibit inconsistency in their abusive behaviors, filling a critical gap in our
understanding of the reasons why managers may be abusive (cf. Tepper, 2007; Tepper, et al.,
2011).
In addition to examining antecedents, our results suggest a connection between the non-
work domain of leaders and the work domain of subordinates, supporting our hypothesis that, at
the day-level, a supervisor’s sleep quality impacts subordinate outcomes indirectly by increasing
supervisors’ daily abusive behaviors. The finding that daily abusive supervisor behavior leads to
detrimental subordinate outcomes on a daily basis is in contrast to other studies of the outcomes
of abusive supervision, which have focused exclusively on differences between individuals’
emotions, attitudes, and behaviors (see Schyns & Schilling, 2013). As such, our study
demonstrates how a daily perspective enables tests of subordinate reactions to abuse from a
Page 25 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

within-person vantage. This has clear implications for negative spillover processes (Eby, Maher,
& Butts, 2010), in which difficulties outside of the work domain can negatively impact work
experiences. In this case, poor sleep quality outside of work negatively influences leader
behavior toward subordinates.
Moreover, our findings that leader sleep has an indirect effect on daily work unit
engagement highlight an important “crossover” process that might have long-term downstream
outcomes for organizations and employees. For organizations, workers who are disengaged on
any given day will have lower job performance (Rich, LePine, & Crawford, 2010) which can
affect the quality of work output by the organization. A single day of work can represent a
sizeable amount of value for an organization. Imagine if a supervisor’s employees disengaged
entirely from work for a single day. Subtracting weekend (or alternative days off), holidays,
vacation time, and sick days, full-time employees will typically work somewhere between 220
and 240 days per year. Thus, even a single day of work represents somewhere around half of a
percent of the full value that employee brings for a year. Given that there is considerable daily
variance in each of the outcomes we study, it is reasonable to expect that there can be many days
in a given year in which a supervisor suffers a poor night of sleep, is high in abusive supervision
the next day, and elicits disengagement from his/her subordinates. Obviously, exact amounts are
difficult to estimate, given that these constructs vary continuously and that there are also
between-individual differences that also influence these frequencies. Nevertheless, we posit that
lost value from disengagement on a given day can represent considerable amounts of lost value
to organizations. This becomes even more apparent when one begins to aggregate across many
such days of low work unit engagement, many work units, many organizations, and many years,
Page 26 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

or when one examined high reliability contexts in which a moment of disengagement can
produce disastrous consequences.
Our study also contributes to research on sleep in organizations in several ways. Using an
ego depletion framework, we theorized that sleep would affect leadership behaviors, a
relationship that has not previously been proposed. Moreover, we respond to calls to “focus more
directly on sleep quality in addition to sleep quantity” (Barnes, et al., 2011, p. 178), given results
indicating that sleep quality may play an important role in determining behavior (Barnes, et al.,
2011) and attitudes (Scott & Judge, 2006) in the workplace. In our study, sleep quality—the
difficulty of falling asleep and staying asleep (Barnes, 2012)—emerged as an explanatory
variable. This finding is important because it adds to the range of factors relating sleep to
workplace dynamics. Barnes further notes that sleep quantity and quality have parallel additive
effects on self-regulation.
Although we predicted such parallel effects, the effects for sleep quantity were generally
not supported. It is possible that this is simply the result of sampling error, but this is difficult to
assess. The p values for sleep quantity were not close to conventional cutoffs, indicating that
there would have had to be considerable levels of such sampling error to create a Type II error.
An alternative possibility is that supervisors are more aware of their sleep quantity than quality,
and are more carefully monitoring their behavior after low sleep quantity but not poor sleep
quality. Another possibility is that chronic sleep deprivation may be more powerful than acute
sleep deprivation in predicting abusive supervision. Although we do not have any measures of
chronic sleep deprivation, future research may do well do examine this question. Moreover, there
is some evidence in the extant literature of a possible threshold effect with quantity. Christian
and Ellis (2011) found a difference between those above and below 6 hours of sleep, and in our
Page 27 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

sample we found that only 13 percent of the observations of supervisor sleep quantity achieved
this level of deprivation. This may have limited our ability to explain variance in abusive
supervision by restricting the range of the independent variable. Further, sleep quality is a
variable that might be subject to more variation than quantity.
Limitations and future research
Despite the strengths of our methodology—we collected data over time from the separate
sources of subordinates and supervisors—helping us to avoid inflated correlations commonly
found in same-source data (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003), our study has
several limitations. We did not manipulate variables or use random assigment techniques, which
would enable more clear causal inferences. Instead, we rely on our theory, the time-separated
nature of our daily assessments, and a within-participant design that partials out any between-
participant differences in order to test our hypotheses specifying directional relationships.
Although we control specifically for supervisor trait anxiety, and use a design that parses out any
potential between-supervisor confounds, we did not include an exhaustive list of day-level
controls. We encourage future research to include such day-level control variables, perhaps
starting with the most conceptually relevant, such as supervisor day-level workload, supervisor
day-level stress, and supervisor day-level health and well-being related variables.
Because the focus of our research was on supervisor sleep, we did not examine the effects
of subordinate sleep (either as predictors or consequences), despite some compelling theoretical
possibilities that can be addressed in future studies. Niedhammer, David, Degioanni, Drummond,
and Philip (2009) found that the experience of workplace bullying was associated with sleep
disturbance. Given the conceptual overlap between workplace bullying and abusive supervision,
it may well be that daily abusive supervisor behavior would lead to sleep difficulties that night
Page 28 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

for subordinates. Indeed, Rafferty, Restubog, and Jimmieson (2010) found more directly that
abusive supervision is associated with subordinate insomnia. Thus, researchers could examine
the relationship between subordinate and supervisor sleep quantity and quality. These
relationships could be modeled as having lagged crossover effects (Eby et al., 2010), with causal
effects emerging over time, as one role’s lack of sleep crosses over to the other. Supervisor sleep
might have effects on subordinate sleep or vice versa, because it is possible that tired and
fatigued employees lead coworkers to experience stress that is disruptive to their subsequent
sleep patterns. Alternatively, the relationship between supervisor and subordinate sleep might be
modeled as an interaction effect. Tired leaders working with tired subordinates could result in a
particularly toxic combination. Subordinate deviance and unethical behavior are the results of
both lack of subordinate sleep (e.g., Barnes et al., 2011; Christian & Ellis, 2011) and abusive
supervision (e.g., Tepper, Henle, Lambert, Giacalone, & Duffy, 2008). Thus, it is quite possible
that the result of lack of sleep in the supervisor-subordinate dyad could exacerbate the tendency
towards deviance for the subordinate.
Despite a growing body of work that supports an ego-depletion view of sleep and
resulting effects on antisocial or unethical behaviors (e.g., Barnes et al., 2011; Christian & Ellis,
2011), a recent study has shown, by manipulating sleep and ego depletion, that these effects may
not consistently hold (Vohs, Glass, Maddox, & Markman, 2011). Vohs and colleagues’ work
suggests that the effects of ego depletion on aggressive behavior hold across sleep-deprivation
conditions, and thus ego-depletion might be differentiated from fatigue induced by suboptimal
sleep. Future research might be conducted to try to untangle the differences and similarities
between sleep’s effects on fatigue and ego depletion. Also, as Vohs and colleagues point out, it is
possible that self-report measures of depletion are more likely to capture variance associated
Page 29 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

with people’s lay beliefs about the effects of sleep on self-control. Although this is a question
perhaps best addressed by controlled laboratory research, future field studies that expand on our
results might benefit from measuring self-control depletion by developing and using behavioral
measures.
Another potential area for future research would be individual differences that would
moderate the effects of leader sleep on daily abusive supervisor behavior. Our model focuses on
the causal mechanism of ego depletion, from a perspective of self-control. However, future
research may find that individual differences in trait self-control play an important moderating
role; for leaders who are dispositionally high in trait self-control, they may manifest a weaker
relationship between sleep and abusive supervision, whereas the relationship may be stronger for
those low in trait self-control. Similarly, leader agreeableness and emotional stability may play
moderating roles. Moreover, future research should also include other potential predictors of the
daily subordinate outcomes of unit work engagement. Not only are such antecedents relevant in
predicting these outcomes, but some may also play a moderating role of the effects of daily
abusive supervisor behavior on subordinate outcomes.
In our exploratory analyses, we found a preliminary indication that one potential
individual difference moderator to consider for further study is the consistency of the behavior of
the leader. We found a marginally significant interaction, such that subordinates are hesitant to
heavily engage in their work if their leaders are highly variable in abusive supervision, even if
the supervisor is not being abusive on a given day. Perhaps such subordinates are waiting for the
other shoe to drop, in that they might expect the leader’s behavior to become abusive at any
moment. Although we are reluctant to infer conclusively from this exploratory analysis, we hope
that future research follows up on this idea both conceptually and empirically.
Page 30 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Future research should extend our model to other workplace outcomes and over other
periods of time. Our initial work focuses on daily work unit engagement. However, there are
potentially many other outcomes that could be impacted by leader daily sleep and daily abusive
supervisor behavior, to include attitudes, psychological safety, turnover intentions, perceived
supervisor support, and subordinate stress and strain. Moreover, examining the effects over
different durations of time could help to identify effects yet to be uncovered. All of the
constructs in our model, including abusive supervision (to more clearly isolate differences
between momentary versus sustained abuse), can be examined on a longer timescale, considering
average differences between people, which would be relatively stable. For example, we focused
on daily variation in sleep, which influences daily variation in the other constructs in our model.
However, future research could also examine chronic sleep deprivation or chronic insomnia, and
how this might influence between-leader differences in abusive supervision. Whether to focus on
the momentary or sustained aspects of abusive supervision should be driven by the research
question and the nature of the variation of other constructs in a given model.
Finally, future research should delve deeper into the causal steps suggested by our model.
Although we already present a model with a mediational chain, future research should examine
the processes underlying each of these links and fill in greater detail in order to further enhance
our understanding of how these relationships play out. For example, it may be that subordinate
psychological safety or emotions mediate the relationship between abusive supervision and work
unit engagement.
Practical implications
Our study also makes several contributions to practice. By focusing on the antecedents to
daily abusive supervisor behavior, we offer important guidelines to organizations interested in
Page 31 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

limiting abusive behaviors among supervisors at work. Our daily approach towards supervisory
behavior has implications for management practice, and possesses distinct advantages over more
traditional static approaches. For example, static approaches assuming that leaders are
consistently abusive imply selection or termination as the only effective methods by which
abusive supervision can be curbed: abusive managers are abusive, through and through.
However, our study suggests that the rate of abusive behavior is related to exogenous daily
factors such as sleep. Our sleep framework suggests that within-persons interventions to aid
sleep will lead to lower levels of abusive supervision behavior the next day. Indeed, treating
abusive supervision on a given day may be much less intimidating and much more manageable
than preventing all occurrences of abusive supervision in a between-persons approach.
The finding that abusive behavior varies daily suggests that certain factors, including and
in addition to sleep, might lead to rises and falls in abusive supervision. Leaders should thus be
aware of their own abusive “triggers.” For example, they can attempt to delay important
interactions or decisions on days when they have a poor night of sleep the night before. Through
leadership training, organizations can increase awareness of the connections that we observed in
our research, by helping leaders to connect the dots between their sleep, their abusive behavior
towards subordinates, and resulting subordinate hostility and attitudes. In customer service
organizations, cultivating a clear understanding of the relationship between a leader’s behavior
and resulting subordinate hostility could have positive effects on customer perceptions of service
quality and emotional delivery. Moreover, subordinates can learn from our results as well—it is
advisable that a subordinate refrain from behaviors that could instigate an abusive episode, if
they are aware that their manager has slept poorly.
Page 32 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Research has begun to uncover strategies for managing sleepy employees that could
potentially apply to leaders in a manner that would decrease levels of abusive supervisor
behavior. Barnes (2011), Caldwell (2012), and Caldwell, Caldwell and Schmidt (2008)
summarize some of these strategies. Research conducted by Welsh et al. (2014) suggests that
consuming caffeine might help mitigate the effects of poor sleep on ego depletion and
susceptibility to following unethical instructions from an authority figure. It is possible that
caffeine could play a beneficial role for tired supervisors and help to reduce their propensity for
abusive behaviors. Finally, recent research by Lanaj et al. (2014) indicates that eliminating
smartphone use late at night can additionally help employees sleep better. Thus, boundary work
that establishes off-hours for smartphones and work email should help with the effects of sleep
on daily abusive supervisor behavior.
Conclusion
In conclusion, our study connects leader sleep quality to daily abusive supervisor
behavior, which ultimately results in deleterious outcomes for subordinates. Organizations
wishing to create positive work environments for their workforce should take note of the
importance of considering the effect of daily events, both non-work (e.g., sleep) and during work
(e.g., abusive supervision behavior) as precursors to important motivational factors such as unit
work engagement. Our study shows that abusive supervision varies within-person, not just
between person, creating a complicated—but increasingly complete—picture for organizational
scholars, managers, and workers.
Page 33 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

References
Alliger, G. M., & Williams, K. J. 1993. Using signal-contingent experience sampling
methodology to study work in the field: A discussion and illustration examining task
perceptions and mood. Personnel Psychology, 46: 525–549.
Altena, E., van Der Werf, Y., Sans-Arigita, E. J., Voorn, T. A., Rombouts, S. A. R. B., Kuijer, J.
P. A., & van Someren, E. J. W. 2008. Prefrontal hypoactivation and recovery in
insomnia. Sleep, 31, 1271–1276.
Aquino, K., Grover, S. L., Bradfield, M., & Allen, D. G. 1999. The effects of negative
affectivity, hierarchical status, and self-determination on workplace victimization.
Academy of Management Journal, 42: 260-272.
Aryee, S., Chen, Z. X., Sun, L. Y., & Debrah, Y. A. 2007. Antecedents and outcomes of abusive
supervision: Test of a trickle-down model. Journal of Applied Psychology, 92: 191-201.
Bakker, A. B. 2014. Daily fluctuations in work engagement. European Psychologist, 1: 1-10.
Bakker, A. B., & Despoina, X. 2009. The crossover of daily work engagement: Test of an actor-
partner interdependence model. Journal of Applied Psychology, 94: 1562-1571.
Bakker, A. B., van Emmerik, H., & Euwema, M. C. 2006. Crossover of burnout and engagement
in work teams. Work and Occupations, 33: 464-489.
Banks, S. J., Eddy, K. T., Angstadt, M., Nathan, P. J., & Phan, K. L. 2007. Amygdala-frontal
connectivity during emotion regulation. Social Cognitive & Affective Neuroscience, 2:
303–312.
Barber, L., Barnes, C. M., & Carlson, K. 2013. Sleepy respondents: Random and systematic
error effects of insomnia on survey behavior. Organizational Research Methods, 16:
616-649. doi: 10.1177/1094428113493120
Page 34 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Barnes, C. M. 2011. “I’ll sleep when I’m dead”: Managing those too busy to sleep.
Organizational Dynamics, 40(1): 18-26. DOI: 10.1016/j.orgdyn.2010.10.001
Barnes, C. M. 2012. Working in our sleep: Sleep and self-regulation in organizations.
Organizational Psychology Review, 2: 234-257. doi: 10.1177/2041386612450181
Barnes, C. M., Schaubroeck, J. M., Huth, M., & Ghumman, S. 2011. Lack of sleep and unethical
behavior. Organizational Behavior and Human Decision Processes, 115: 169-180.
Barnes, C. M. & Wagner, D. T. 2009. Changing to daylight saving time cuts into sleep and
increases workplace injuries. Journal of Applied Psychology, 94: 1305-1317. doi
10.1037/a0015320
Barnes, C. M., Wagner, D. T., & Ghumman, S. 2012. Borrowing from sleep to pay work and
family: Expanding time-based conflict to the broader non-work domain. Personnel
Psychology, 65: 789-819. doi: 10.1111/peps.12002
Baron, R. M., & Kenny, D. A. 1986. The moderator–mediator variable distinction in social
psychological research: Conceptual, strategic, and statistical considerations. Journal of
Personality and Social Psychology, 51: 1173.
Barrick, M., Thurgood, G., Smith, T., & Courtright, S. (in press). Collective Organizational
Engagement: Linking Motivational Antecedents, Strategic Implementation, and Firm
Level Performance. Academy of Management Journal.
Baumeister, R. F., & Boden, J. M. (1998). Aggression and the self: High self-esteem, low self-
control, and ego threat. In R. G. Geen & E. D. Donnerstein (Eds.), Human Aggression:
Theories, Research, and Implications for Social Policy (pp. 11–137). San Diego, CA:
Academic Press.
Page 35 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Baumeister, R. F., Bratslavsky, E., Muraven, M., & Tice, D. M. (1998). Ego depletion: Is the
active self a limited resource? Journal of Personality and Social Psychology, 74, 1252-
1265.
Baumeister, R. F., Muraven, M., & Tice, D. M. 2000. Ego depletion: A resource model of
volition, self-regulation, and controlled processing. Social Cognition, 18: 130-150.
Beauregard, M., Levesque, J., & Bourgouin, P. 2001. Neural correlates of conscious self-
regulation of emotion. Journal of Neuroscience, 21(RC165): 1–6.
Bhave, D.P., Kramer, A., & Glomb, T.M. 2010. Work family conflict in workgroups: Social
information processing, support, and demographic similarity. Journal of Applied
Psychology, 95: 145-158.
Blanc, M., Mérette, C., Savard, J., Ivers, H., Baillargeon, L., & Morin, C. M. 2009. Incidence
and risk factors of insomnia in a population-based sample. Sleep, 32: 1027- 1037.
Bono, J. E., Foldes, H. J., Vinson, G., & Muros, J. P. 2007. Workplace emotions: The role of
supervision and leadership. Journal of Applied Psychology, 92: 1357-1367.
Bono, J. E., & Judge, T. A. 2003. Self-concordance at work: Toward understanding the
motivational effects of transformational leaders. Academy of Management Journal, 46:
554-571.
Breevart, K., Bakker, A. B., & Demerouti, E. 2014. Daily self-management and employee work
engagement. Journal of Vocational Beahvior, 84: 31-38.
Brief, A. P., Butcher, A. H., & Roberson, L. 1995. Cookies, disposition, and job-attitudes: The
effects of positive mood-inducing events and negative affectivity on job-satisfaction in a
field experiment. Organizational Behavior and Human Decision Processes, 62: 55-62.
Page 36 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Brief, A. P., & Weiss, H. M. 2002. Organizational behavior: Affect in the workplace. Annual
Review of Psychology, 53: 279-307.
Brislin, R. W. (1986). The wording and translation of research instruments. In W. J. Loner & J.
W. Berry (Eds.) Field Methods in Cross-Cultural Research (pp. 137-164). Newbury
Park, CA: Sage.
Caldwell, J. A. 2012. Crew schedules, sleep deprivation, and aviation performance. Current
Directions in Psychological Science, 21: 85-89.
Caldwell, J. A., Caldwell, J. L., & Schmidt, R. M. 2008. Alertness management strategies for
operational contexts. Sleep Medicine Reviews, 12: 257-273
Christian, M. S., & Ellis, A. P. J. 2011. Examining the effects of sleep deprivation on workplace
deviance: A self-regulatory perspective. Academy of Management Journal, 54: 913–
934.
Christian, M. S. Garza, A. S., & Slaughter, J. E. 2011. Work engagement: A quantitative review
and test of its relations with task and contextual performance. Personnel Psychology, 64:
89-136.
Chuah, L. Y. M., Dolcos, F., Chen, A. K., Zheng, H., Parimal, S., & Chee, M. W. L. 2010. Sleep
deprivation and interference by emotional distractors. Sleep, 33: 1305–1313.
Costa, P., Passos, A. M., & Bakker, A. B. 2014. Team work engagement: A model of emergence.
Journal of Occupational and Organizational Psychology, 87: 414-436.
Dalal, R. S., Bhave, D. P., & Fiset, J. (2014). Within-Person Variability in Job Performance A
Theoretical Review and Research Agenda. Journal of Management, 40:1396-1436.
Dalal, R. S., Lam, H., Weiss, H. M., Welch, E. R., & Hulin, C. L. 2009. A within-person
approach to work behavior and performance: Concurrent and lagged citizenship-
Page 37 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

counterproductivity associations, and dynamic relationships with affect and overall job
performance. Academy of Management Journal, 52: 1051-1066.
Danner, D. D., Snowdon, D. A., & Friesen, W. V. 2001. Positive emotions in early life and
longevity: Findings from the nun study. Journal of Personality and Social Psychology,
80: 804-813.
DeRue, D. S. 2011. Adaptive leadership theory: Leading and following as a complex adaptive
process. Research in Organizational Behavior, 31: 125-150.
Diener, E., Nickerson, C., Lucas, R. E., & Sandvik, E. 2002. Social Indicators Research, 59:
229-259.
Doyle, W. J., Gentile, D. A., & Cohen, S. 2006. Emotional style, nasal cytokines, and illness
expression after experimental rhinovirus exposure. Brain Behavior and Immunity, 20:
175-181.
Duffy, M. K., Shaw, J. D., & Stark, E. M. 2000. Performance and satisfaction in conflicted
interdependent groups: When and how does self-esteem make a difference? Academy of
Management Journal, 43: 772-782.
Eagly, A., & Chaiken, S. 1993. Psychology of Attitudes. New York: Harcourt, Brace
Jovanovich.
Eby, L. T., Maher, C. P., & Butts, M. M. 2010. The intersection of work and family life: The role
of affect. Annual Review of Psychology, 61: 599-622.
Fairclough, S. H., & Houston, K. 2004. A metabolic measure of mental effort. Biological
Psychology, 66: 177–190.
Page 38 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Felps, W., Mitchell, T. R., Hekman, D. R., Lee, T. W., Holtom, B. C., & Harman, W. S. 2009.
Turnover contagion: How coworkers’ job embeddedness and job search behaviors
influence quitting. Academy of Management Journal, 52, 545-561.
Fredrickson, B. L. 2000. The role of positive emotions in positive psychology: The Broaden-and-
Build theory of positive emotions. American Psychologist, 56: 218-226.
Ghumman, S. & Barnes, C. M. 2013. Sleep and prejudice: A resource recovery approach.
Journal of Applied Social Psychology, 43: E166-E178. doi: 10.1111/jasp.12045
Glomb, T. M., Bhave, D. P., Miner, A. G., & Wall, M. 2011. Doing good, feeling good:
Examining the role of organizational citizenship behaviors in changing mood. Personnel
Psychology, 64: 191-223.
Gino, F., Schweitzer, M. E., Mead, L., & Ariely, D. 2011. Unable to resist the temptation: How
self-control depletion promotes unethical behavior. Organizational Behavior and
Human Decision Processes, 115: 191–203
Grant, A. M. 2012. Leading with meaning: Beneficiary contact, prosocial impact, and the
performance effects of transformational leadership. Academy of Management
Journal, 55: 458-476.
Hagger, M. S., Wood, C., Stiff, C., & Chatzisarantis, N. L. D. 2010. Ego depletion and the
strength model of self-control: A meta-analysis. Psychological Bulletin, 136: 495–525.
Harter, J. K., Schmidt, F. L., & Hayes, T. L. 2002. Business-unit-level relationship between
employee satisfaction, employee engagement, and business outcomes: A meta-
analysis. Journal of Applied Psychology, 87: 268.
Hoffmann, S. A., Griffin, M. A., & Gavin, M. B. 2000. The application of hierarchical linear
modeling to organizational research. In K. Klein & S. Kozlowski (Eds.), Multilevel
Page 39 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

theory, research, and methods in organizations: Foundations, extensions, and new
directions: 75–170. San Francisco: Jossey-Bass.
Hoobler, J. M., & Brass, D. J. 2006. Abusive supervision and family undermining as displaced
aggression. Journal of Applied Psychology, 91: 1125-1133.
Horne, J. A. 1993. Human sleep, sleep loss and behavior: Implications for the prefrontal cortex
and psychiatric disorder. British Journal of Psychiatry, 162: 413–419.
Hox, J.J. 2010. Multilevel Analysis. New York, NY: Routledge.
Ilies, R., & Judge, T. A. 2002. Understanding the dynamic relationships among personality,
mood, and job satisfaction: A field experience sampling methodology. Organizational
Behavior and Human Decision Processes, 89: 1119-1139.
Johnson, R. E., Venus, M., Lanaj, K., Mao, C., & Chang, C. H. 2012. Leader identity as an
antecedent of the frequency and consistency of transformational, consideration, and
abusive leadership behaviors. Journal of Applied Psychology, 97, 1262-1272.
Judge, T. A., & Piccolo, R. 2004. Transformational and transactional leadership: A meta-analytic
test of their relative validity. Journal of Applied Psychology, 89: 755–768.
Kahn-Greene, E. T., Lipizzi, E. L., Conrad, A. K., Kamimori, G. H., & Killgore, W. D. S. 2006.
Sleep deprivation adversely affects interpersonal responses to frustration. Personality
and Individual Differences, 41: 1433–1443.
Kant, L., Skogstad, A., Torsheim, T., & Einarsen, S. 2013. Beware the angry leader: Trait anger
and trait anxiety as predictors of petty tyranny. Leadership Quarterly, 24: 106-1024.
Klein, K. J., & Kozlowski, S. W. J. 2000. From micro to meso: Critical steps in conceptualizing
and conducting multilevel research. Organizational Research Methods, 3: 211-236.
Page 40 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Knutson, K. L., Rathouz, P. J., Yan, L. 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: 11-23
Lazarus, R. S. 1991. Emotion and Adaptation. New York: Oxford University Press.
Leavitt, K., Reynolds, S., Barnes, C. M., Schilpzand, P., & Hannah, S. T. 2012. Different hats,
different obligations: Plural occupational identities and situated moral judgments.
Academy of Management Journal, 55: 1316-1333.
LeBlanc, M., Merette, C., Savard, J., Ivers, H., Baillargeon, L., & Morin, C. M. 2009. Incidence
and risk factors of insomnia in a population-based sample. Sleep, 32: 1027-1037.
Lorinkova, N. M., Pearsall, M. J., & Sims, H. P. 2013. Examining the differential longitudinal
performance of directive versus empowering leadership in teams. Academy of
Management Journal, 56: 573-596.
Lyubomirsky, S., King, L., & Diener, E. 2005. The benefits of positive affect: Does happiness
lead to success? Psychological Bulletin, 131: 803-855.
MacKinnon, A., Jorm, A. F., Christensen, H., Korten, A. E., Jacomb, P. A., & Rodgers, B. 1999.
A short form of the Positive and Negative Affect Schedule: Evaluation of factorial
validity and invariance across demographic variables in a community sample.
Personality and Individual Differences, 27: 405-416.
MacKinnon, D. P. 2008. Introduction to Statistical Mediation Analysis. New York: Lawrence
Erlbaum Associates.
Page 41 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Mayer, D. M., Thau, S., Workman, K. M., Van Dijke, M., & De Cremer, D. 2012. Leader
mistreatment, employee hostility, and deviant behaviors: Integrating self-uncertainty and
thwarted needs perspectives on deviance. Organizational Behavior and Human
Decision Processes, 117: 24-40.
Monk, T. H., Reynolds, C. F., Kupfer, D. J., Buysse, D. J., Coble, P. A., Hayes, A. J. . . .
Ritenour, A. M. 1994. The Pittsburgh Sleep Diary. Journal of Sleep Research, 3: 111–
120.
Moskowitz, J. T. 2003. Positive affect predicts lower risk of AIDS mortality. Psychosomatic
Medicine, 6:, 620–626.
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. and Muthén, B.O. (1998-2010). Mplus User’s Guide. Sixth Edition.
Los Angeles, CA: Muthén & Muthén
Niedhammer, I., David, S., Degioanni, S., Drummond, A., & Philip, P. 2009. Workplace
bullying and sleep disturbances: Findings from a large scale cross-sectional survey in the
French working population. Sleep, 32: 1211-1219.
Nilsson, J. P., Soderstrom, M., Karlsson, A. U., Lekander, M., Akerstedt, T., Lindroth, N. E., &
Axelsson, J. 2005. Less effective executive functioning after one night’s sleep
deprivation. Journal of Sleep Research, 14: 1–6.
Ochsner, K. N., Ray, R. D., Cooper, J. C., Robertson, E. R., Chopra, S., Gabrieli, J. D. E., &
Gross, J. J. 2004. For better or for worse: Neural systems supporting the cognitive down-
and upregulation of negative emotion. Neuroimage, 23: 483–499.
Page 42 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Ostir, G. V., Markides, K. S., Black, S. A., & Goodwin. J. S. 2000. Emotional well-being
predicts subsequent functional independence and survival. Journal of the American
Geriatrics Society, 48: 473-478.
Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. 2003. Common method
biases in behavioral research: a critical review of the literature and recommended
remedies. Journal of Applied Psychology, 88: 879 - 903.
Preacher, K. J., & Hayes, A. F. 2008. Asymptotic and resampling strategies for assessing and
comparing indirect effects in multiple mediator models. Behavior Research Methods, 40:
879-891.
Priesemuth, M., Schminke, M., Ambrose, M. L., & Folger, R. in press. Abusive supervision
climate: A multiple-mediation model of its impact on group outcomes. Academy of
Management Journal.
Rafferty, A. E., Restubog, S. L. D., & Jimmieson, N. L. (2010). Losing sleep: Examining the
cascading effects of supervisors’ experience of injustice on subordinates’ psychological
health. Work & Stress, 24: 36-55.
Raudenbush, S. W., & Bryk, A. S. 2002. Hierarchical Linear Models: Applications and Data
Analysis Methods. Thousands Oaks, CA: Sage.
Reis, H. T., & Wheeler, L. 1991. Studying social interaction with the Rochester Interaction
Record. In M. P. Zanna (Ed.), Advances in experimental social psychology, vol. 24:
270–318. San Diego: Academic Press.
Rich, B. L., Lepine, J. A., & Crawford, E. R. 2010. Job engagement: Antecedents and effects on
job performance. Academy of Management Journal, 53: 617-635.
Page 43 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Richman, L. S., Kubzansky, L., Maselko, J., Kawachi, I., Choo, P., & Bauer, M. 2005. Positive
emotion and health: Going beyond the negative. Health Psychology, 24: 422–429.
Salanova, M., Agut, S., & Peiró, J. M. 2005. Linking organizational resources and work
engagement to employee performance and customer loyalty: the mediation of service
climate. Journal of Applied Psychology, 90: 1217-1227.
Schaufeli, W. B., Bakker, A. B., & Salanova, M. 2006. The measurement of work engagement
with a short questionnaire – A cross-national study. Educational and Psychological
Measurement, 66, 701-716.
Schyns, B., & Schilling, J. 2013. How bad are the effects of bad leaders? A meta-analysis of
destructive leadership and its outcomes. Leadership Quarterly, 24: 138-158.
Scott, B. A., & Barnes, C. M. 2011. A multilevel investigation of emotional labor, affect,
withdrawal, and gender. Academy of Management Journal, 54, 116-136.
Scott, B. A., Barnes, C. M., & Wagner, D. 2012. Chameleonic or consistent? A multilevel model
of self monitoring and variability in emotional labor. Academy of Management Journal.
55, 905-926. doi: 10.5465/amj.2010.1050
Scott, B. A., & Judge, T. A. 2006. Insomnia, emotions, and job satisfaction: A multilevel study.
Journal of Management, 32: 622–645.
Shoss, M. K., Eisenberger, R., Restubog, S. L. D., & Zagenczyk, T. J. 2013. Blaming the
organization for abusive supervision: The roles of perceived organizational support and
supervisor’s organizational embodiment. Journal of Applied Psychology, 98: 158-168.
Smith, C. A., Haynes, K. N., Lazarus, R. S., & Pope, L. K. 1993. In search of the ‘‘hot’’
cognitions: Attributions, appraisals, and their relation to emotion. Journal of Personality
and Social Psychology, 65: 916–929.
Page 44 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

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.
Sy, T., & Choi, J. N. 2013. Contagious leaders and followers: Exploring multi-stage mood
contagion in a leader activation and member propagation (LAMP) model. Organizational
Behavior and Human Decision Processes, 122: 127-140.
Tepper, B. J. 2000. Consequences of abusive supervision. Academy of Management Journal,
43: 178-190.
Tepper, B. J. 2007. Abusive supervision in work organizations: Review, synthesis, and research
agenda. Journal of Management, 33: 261-289.
Tepper, B. J., Duffy, M. K., Henle, C. A., & Lambert, L. S. 2006. Procedural injustice, victim
precipitations, and abusive supervision. Personnel Psychology, 59: 101-123.
Tepper, B. J., Henle, C. A., Lambert, L. S., Giacalone, R. A., & Duffy, M. K. 2008. Abusive
supervision and subordinates' organization deviance. Journal of Applied Psychology, 93:
721-732.
Tepper, B. J., Moss, S. E., & Duffy, M. K. 2011. Predictors of abusive supervision: Supervisor
perceptions of deep-level dissimilarity, relationship conflict, and subordinate
performance. Academy of Management Journal, 54, 279-294.
Twenge, J., Muraven, M., & Tice, D. 2004. Measuring state self-control: Reliability, validity,
and correlations with physical and psychological stress. Unpublished manuscript, San
Diego State University.
Thomas, M., Sing, H., Belenky, G., Holcomb, H., Mayberg, H., Dannals, R., . . . Redmond, D.
2000. Neural basis of alertness and cognitive performance impairment during sleepiness.
Page 45 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

I. Effects of 24 h of sleep deprivation on waking human regional brain activity. Journal
of Sleep Research, 9: 335–352.
Venus, M., Stam, D., & van Knippenberg, D. 2013. Leader emotion as a catalyst of effective
leader communication of visions, value-laden messages, and goals. Organizational
Behavior and Human Decision Processes, 122: 53-68.
Vohs, K. D., Glass, B. D., Maddox, W. T., & Markman, A. B. 2011. Ego Depletion Is Not Just
Fatigue Evidence From a Total Sleep Deprivation Experiment. Social Psychological and
Personality Science, 2: 166-173.
Wagner, D. T., Barnes, C. M., Lim, V., & Ferris, D. L. 2012. Lost sleep and cyberloafing:
Evidence from the laboratory and a Daylight Saving Time quasi-experiment. Journal of
Applied Psychology, 97: 1068-1076. doi: 10.1037/a0027557
Wagner, D. T., Barnes, C. M., & Scott, B. A. 2014. Driving it home: How workplace emotional
labor harms employee home life. Personnel Psychology, 67, 487-516
Wallace, J. C., Edwards, B. D., Paul, J., Burke, M., Christian, M., & Eissa, G. (2013). Change
the Referent? A Meta-Analytic Investigation of Direct and Referent-Shift Consensus
Models for Organizational Climate. Journal of Management.
Weiss, H. M. 2002. Antecedents of emotional experiences at work. Motivation and Emotion, 26:
1-2.
Weiss, H. M., & Cropanzano, R. 1996. Affective events theory: A theoretical discussion of the
structure, causes, and consequences of affective experiences at work. Research in
Organizational Behavior, 18: 1-74.
Page 46 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Weiss, H. H., Nicholas, J. P., & Daus, C. S. 1999. An examination of the joint effects of affective
experiences and job beliefs on job satisfaction and variations in affective experiences
over time. Organizational Behavior and Human Decision Processes, 78: 1-24.
Welsh, D. T., Ellis, A. P. J., Christian, M. S., & Mai, K.M. 2014. Expanding the self-regulatory
model of sleep deprivation and unethical behavior: The role of caffeine and social
influence. Journal of Applied Psychology.
Wheeler L., & Reis, H. T. 1991. Self-recording of everyday life events: Origins, types, and uses.
Journal of Personality, 59: 339–354
Page 47 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Table 1: Descriptives and Correlations
Mean SD 1 2 3 4
1. Daily Sleep Quality -8.29 14.23 -
2. Daily Sleep Quantity 435.59 70.73 .18 -
3. Daily Leader Ego Depletion 1.82 .83 -.13 -.03 -
4. Daily Abusive Supervision 1.58 .55 .13 .01 .33 -
5. Daily Unit Work Engagement 3.06 .87 -.04 .03 -.35 -.52
Notes. N = 606.
Correlations greater than |.12| are statistically significant at p < .05, two-tailed
Page 48 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Table 2: Multilevel Path Analysis Results
Ego Depletion Abusive Supervision Unit Work Engagement
Main Effects
Supervisor Trait Anxiety -.08 -.05 .03
Sleep Quality -.13* .19** -.02
Sleep Quantity .02 -.01 .02
Ego Depletion .35** -.19**
Abusive Supervision -.45**
R-sq .02 .14 .30
Indirect Effects Estimate LLCI UCLI
Sleep QualityAbusive Supervision
(via Ego Depletion)
-.04* -.084 -.003
Sleep Quantity Abusive Supervision
(via Ego Depletion)
.01 -.036 .050
Ego Depletion Unit Work Engagement
(via Abusive Supervision)
-.15** -.229 -.080
Sleep Quality Unit Work Engagement
(via Ego Depletion & Abusive Supervision)
.02* .000 .039
Sleep Quantity  Unit Work Engagement
(via Ego Depletion & Abusive Supervision)
-.00 -.022 .016
Notes. N = 606. LLCI = lower level of the 95% confidence interval. UCLI = upper level of the
95% confidence interval. The model was estimated simultaneously. Standardized estimates are
reported.
* p < .05, ** p < .01, two-tailed.
Page 49 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Table 3: Exploratory Analysis of Abusive Supervision Variability as a Moderator
Unit Work Engagement
Model 1 Model 2 Model 3
Supervisor Trait Anxiety .10 .04 .04
Abusive Supervision Variability -.25** .09† .01
Abusive Supervision -.66** -.64**
Abusive Supervision X Abusive Supervision
Variability
.13†
R-sq .07 .39 .40
∆ R-sq .32 .01
Notes. N = 606. † p<.10, * p < .05, ** p < .01, two-tailed. Standardized estimates are reported.
Page 50 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Figure 1: Conceptual Model
Page 51 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Figure 2: Exploratory Analysis of Abusive Supervision Variability as a Moderator
Low Abusive
Supervision
High Abusive
Supervision
Low Abusive
Supervision
Variability
High Abusive
Supervision
Variability
Page 52 of 53Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

Bios
Christopher M. Barnes (chris24b@uw.edu) is an assistant professor of Organizational
Behavior at the Michael G. Foster School of Business, University of Washington. He received
his PhD in Organizational Behavior from Michigan State University. His research interests focus
primarily on sleep and fatigue in organizations, with other interests in behavioral ethics,
emotional labor, and team performance.
Lorenzo Lucianetti (llucianetti@unich.it) is an assistant professor in business administration in
the Department of Management and Business Administration at the University of Chieti and
Pescara (Italy). He is also a visiting research fellow at University of Cranfield (UK). His PhD is
from the University of Chieti and Pescara. His research has an inter-disciplinary focus that spans
multiple disciplines including Accountancy (Financial Reporting), Operations Management
(Performance Measures Systems) and Psychology (Abusive Supervision and Mistreatment on the
workplace).
Devasheesh P. Bhave (dbhave@smu.edu.sg) is an assistant professor of organizational behavior
and human resources in the Lee Kong Chian School of Business at Singapore Management
University. He received his Ph.D. from the Carlson School of Management at the University of
Minnesota. His research interests include dynamic processes of affect and performance,
interpersonal relationships at work, and customer service.
Michael S. Christian (mike_christian@unc.edu) is an assistant professor of organizational
behavior at the Kenan-Flagler Business School, University of North Carolina at Chapel Hill. His
research focuses on understanding how energy, engagement, self-control and other self-
regulatory processes affect behavior at work. He studies how these factors dynamically relate to
unethical behavior as well as desirable work performance. He received his PhD in management
from the University of Arizona and his master’s degree in industrial and organizational
psychology from Tulane University.
Page 53 of 53 Academy of Management Journal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60