You are on page 1of 12

The Leadership Quarterly 23 (2012) 342–353

Contents lists available at SciVerse ScienceDirect

The Leadership Quarterly


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

Embedded leadership: How do a leader's superiors impact


middle-management performance?
Jay J. Caughron a,⁎, Michael D. Mumford b
a
Radford University, United States
b
The University of Oklahoma, United States

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

Article history: This study uses a low-fidelity simulation to test the effect superiors can have on the leadership
Received 25 May 2010 style and cognition of their subordinates who also are leaders. These leaders within the orga-
Accepted 1 August 2011 nization, often referred to as middle managers, occupy an important, albeit overlooked posi-
Available online 16 December 2011
tion within an organization. In order to emphasize the leadership that occurs at the middle
levels of management the term ‘embedded leader’ is used. Using a sample of 224 undergradu-
Keywords: ates, three variables were manipulated to examine their effects on participant sensemaking,
Leadership confidence, and participative leadership. The variables manipulated were the severity of indi-
Embedded leadership
vidual level consequences (high vs. low), the severity of organizational consequences (high vs.
Leader follower relationships
low), and the superior's motivational strategy (coercive vs. supportive vs. passive style). It was
Leader sensemaking
Coercive leadership found that a leader's superior can influence their leadership behaviors in a complex manner.
Supportive leadership Participant sensemaking was higher when their superior emphasized low levels of individual
Middle management (or personal) consequences. Participative leadership was higher when the supervisor framed
organizational level consequences as being high. Findings for participant confidence were
complex, but generally suggest that one's superiors can promote or hinder confidence in a va-
riety of situations.
© 2011 Elsevier Inc. All rights reserved.

As an organization grows in size, multiple layers of management develop in order to keep up with expanding demands.
Of course the most visible leaders are those who occupy the highest hierarchical positions within organizations. However, it is
important to consider leaders who occupy middle or lower level managerial positions as well. Organizational effectiveness may
hinge on the coordination and communication that occurs within the organization between departments and across levels
of management. For example, middle level leaders work to get projects finished through their subordinates, coordinate activities
between groups within the organization, exert influence on others within and outside of the organization, plan for future projects,
and align their groups goals with larger organizational goals (DeChurch, Hiller, Murase, Doty, & Salas, 2010; Jacobs & McGee,
2001).
Research on leadership at multiple levels of management is not new. In a recent review of research on leadership, DeChurch
et al. (2010) found that in the 25 years prior to their analysis upper management was the focus of research in 34% of studies,
lower-levels of management 16% and middle management only 7%. Curiously, leadership behaviors in the middle of the organi-
zation have been researched much less than leadership at the higher and lower ends of the organization, leading DeChurch and
colleagues to conclude “By far the least well empirically-understood aspect of organizational leadership happens in the middle
place” (p. 1078).
It is important to point out that as organizations work to “right-size” their workforce employees in middle management are
often targeted for downsizing. While this means the number of middle managers may be decreasing, it also suggests that those
who remain will have to take on a greater burden. Their position ‘in the middle place’ of the organization dictates that they are

⁎ Corresponding author at: Department of Psychology, Radford University, Radford, VA, United States. Tel.: +1 540 831 8700.
E-mail address: jcaughron@radford.edu (J.J. Caughron).

1048-9843/$ – see front matter © 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.leaqua.2011.08.008
J.J. Caughron, M.D. Mumford / The Leadership Quarterly 23 (2012) 342–353 343

able to exert upward and downward influence effectively as they work to implement upper management's strategy but respond
to the needs of the workforce they lead (Floyd & Wooldridge, 1992, 1994; Yammarino, 1994). They also influence organizational
success by making sense of organizational events for organizational members above and below them in the hierarchy. In this way
they help the organization adapt to emergent events (DeChurch et al., 2010; Weick, 1995).
Superiors can exert a great deal of influence on their subordinates, and there is no reason to suspect that middle managers
are immune to this effect. They have power over the allocation of resources, promotion decisions, performance evaluations,
and a variety of other processes and outcomes that have a real and direct impact on the managers under their supervision and
direction (Bass, 2008; Raven & French, 1958a, 1958b; Smith, 1948). Additionally, superiors often set the stage for sensemaking
in their subordinates by selecting variables to be monitored, emphasizing certain aspects of performance in group and individual
evaluations, and by controlling the flow of information to subordinates (Bass, 2008; Weick, 1995; Yukl, 2006).
While upper level leaders have a notable influence over middle managers, their impact on lower level employees appears to be
negligible. Chun, Yammarino, Dionne, Sosik and Moon (2009) used a survey technique to examine leadership from a multiple
levels of management and a multiple levels of analysis perspective in the same study. Their findings suggest that upper level
leaders can have a wide ranging influence on middle managers, but that their impact on lower level employees is minimal. This
would underscore the importance of the middle manager as an intermediary who is able to exert greater influence on lower
level employees than a ‘distant’ high level leader. Chun et al. (2009) go on to suggest that high level organizational leaders
would be well advised to develop subordinates who can serve as their surrogates. In this way, higher level leaders can exert indi-
rect influence on their front-line employees through their middle managers. This study seeks to follow up on Chun et al. (2009)
examination of the middle levels of management. It should be emphasized that the term ‘levels’ as it is used in this manuscript re-
fers to levels of management, rather than levels of analysis. This manuscript examines the impact a supervisor can have on their
direct subordinate's leadership, but does not use a ‘levels-of-analysis’ approach. Thus, the term ‘level’ is not referring to levels of
analysis as is common in the literature (Chun et al., 2009; Yammarino & Bass, 1991; Yammarino & Dansereau, 2009a, 2009b).
The primary purpose of this research is to examine the ‘second-tier’ leader; the leader who follows, as well as leads. These in-
dividuals are typically referred to as middle managers. Yammarino (1994) refers to them as ‘direct reports’ whereas Chun and
colleagues uses the term ‘close’ followers. These individuals report directly to their own leaders but also work with lower level
employees to accomplish organizational tasks. These lower level employees have been referred to as ‘indirect reports’ or as ‘dis-
tant’ followers (Chun et al., 2009; Yammarino, 1994). Distant followers are separated from a higher level leader by at least one
layer of management. In order to emphasize the leadership aspect of what middle managers do and to avoid confusion about
which level of leadership is being discussed throughout the rest of this manuscript we use the term ‘embedded leader.’ When
we use this term, we refer to a person who reports to another higher level leader within the organization but is also responsible
for leading others towards organizational goals.
Given that studying leadership “in the middle place” is important (Chun et al., 2009; DeChurch et al., 2010; Yammarino &
Bass, 1991), how are researchers to go about this task? The preponderance of research focusing on levels of management issues
in leadership has been conducted in the field. Field research is indeed a valuable approach because it can capture leadership
phenomena at multiple levels of management as they occur in real-world settings. Additionally, it is very difficult to simulate a
variety of aspects of real work settings in the lab, especially those dealing with the nature of interpersonal relationships which
are developed over time.
That being said, it is also important to take a variety of approaches in conducting leadership research. To this point, in
DeChurch et al.'s (2010) review, they found that only 11.5% of studies examining multiple levels of management were conducted
in the lab. Studying people's behavior as they engage in middle management tasks in a lab setting presents several advantages.
First, researchers are able to use levels of control that would be impossible in a real-world setting. Second, researchers can target
and manipulate specific variables of interest based on theory or previous findings. Third, biases that are found in real-world
research can be eliminated in a lab setting. Specifically, the majority of work on multiple levels of management issues employ
surveys; surveys that are typically filled out by the leader of interest as well as that leader's superiors or subordinates. These
surveys collect responses based on co-workers' opinions leader of interest. This is practically the definition of a biased response.
A more objective evaluation of the leader's performance can be obtained in a lab. Fourth, the dependent variables examined in a
real-world setting are much more limited than those that can be examined within a lab. Specifically, Kaiser, Hogan and Craig
(2008) noted that not only are many of the variables examined in leadership research drawn from people's perceptions of a
particular leader, they are also very often some measure of career success for the leaders in question. As they point out, actions
that lead to career success are not the same as those that lead to organizational success, nor are they necessarily the same as
those that would be considered high quality leadership. The outcome variables that are the most theoretically and rationally
interesting can be examined in a lab setting with much less regard for what are convenient sources of outcome data.
Thus, this study seeks to make several unique contributions to the literature. First, it is an experimental examination of lead-
ership phenomenon at multiple levels of management. By simulating middle management tasks and manipulating factors that
influence middle managers we can learn about how these individuals are influenced by their environment and then measure
the outcomes that are relevant to how they lead their subordinates. Second, this study seeks to go beyond correlational findings
that typically result from survey research. By approaching this topic using an experimental technique we can make firm conclu-
sion regarding causal relationships. Third, and finally, this study seeks to expand the outcome variables of interest in multiple
levels of management research. While assessing career attainment and the perceptions of leader performance are important,
they are limited. This study seeks to measure the cognition people engage in when performing middle management tasks
when exposed to select middle management environmental factors.
344 J.J. Caughron, M.D. Mumford / The Leadership Quarterly 23 (2012) 342–353

1. Leadership outcomes

Fortunately, a wide array of leadership variables has been identified and shown to be related to leader and group performance.
One such variable is participative leadership. Participative leadership can be described as a leadership style in which the leader
shares decision-making power, encourages follower input during discussions, and expects followers to be actively involved in
the process of problem-solving (Bass, 2008). Participative leadership has been shown to be related to decision quality (Bass,
1960; Lanzetta & Roby, 1960; Lorge, Fox, Davitz, & Brenner, 1958), decision acceptance (Bass, 1958; Likert, 1961a, 1961b), follow-
er satisfaction (Aspergren, 1963; Harrison, 1985; Preston & Heintz, 1949; Ziller, 1954), task motivation (Aspergren, 1963), and
absenteeism (Mann & Baumgartel, 1952).
In addition to participative leadership, it is also important to consider sensemaking. Leader sensemaking is important for lead-
er problem-solving, decision-making, communication, and planning. Sensemaking is one's ability to develop a coherent, inter-
pretable understanding of an ambiguous situation that facilitates subsequent decision-making and action (Caughron, Shipman,
Beeler, & Mumford, 2009; Weick, 1995). Not only does sensemaking help a leader make decisions and direct action, it also sets
the stage for communication, coordination, and progress monitoring within a given situation (Day, Gronn, & Salas, 2004; Hershey
& Walsh, 2001). As such, it is an important variable to consider when examining how a leader, in this case an embedded leader, is
likely to function in situations calling for problem-solving, communicating with subordinates and superiors, decision-making,
directing action, and monitoring a group's progress toward a goal.
Confidence is another variable that has received a great deal of attention in leadership research. A leader who appears
confident is more likely to instill confidence in followers and engender their commitment to shared goals (Black & Porter,
2000; Northouse, 1997). Leaders who are confident enjoy greater influence over followers than those who appear less confident
(Post, 1986) in addition to higher ratings of leadership from peers and supervisors (Chemers, Watson, & May, 2000). Clearly
confidence is an important variable to consider in evaluating a leader's performance and is likely to be influenced by many situ-
ational factors that can be manipulated in a laboratory setting.

2. Superior's leadership technique

Of course, if an examination of how leaders are influenced by their environment is to be conducted, it is important to consider
the leadership behaviors demonstrated by their superior. Individualized leadership theory suggests that leaders exert a great deal
of influence on followers on a dyadic level. Specifically, it suggests that leaders who support their followers' sense of self-worth
will enable those followers to demonstrate satisfactory levels of performance (Dansereau et al., 1995).
Given the nature of individualized leadership it should not be surprising that two aspects of leadership that are particularly
salient to followers is the extent to which a leader is either supportive or coercive. Bankhart and Lanzetta (1970), Barrow
(1976), and Hinton and Barrow (1975) have all provided evidence for the fact that superiors tend to become more coercive
when their subordinates' performance is inadequate. Similarly, Kipnis (1976) has shown that leaders are more likely to use coer-
cive influence tactics when faced with a potential failure situation. Additionally, it is often the case in real-world organizational
settings that organizational leaders will choose a hard-nosed, coercive leader when they discover a high consequence situation
needing a quick and effective, resolution. However, little research has been conducted to test the wisdom of this strategy.
Although individualized leadership theory suggests supportive leadership will result in higher levels of subordinate perfor-
mance (Dansereau et al., 1995), it is not always entirely clear what an organization gains when their superiors use a supportive
leadership style (Yukl, 2006). Groups with supportive leaders may become more cohesive and demonstrate higher levels of mem-
ber satisfaction, but may also be slower in solving problems (Bass, 2008; Fox, 1954, 1957). Thus the actual benefit of using a sup-
portive leadership style will likely depend on a given set of circumstances.
This is especially true when one considers the relationship between supportive leadership and follower confidence. While it is
easy to imagine a case in which supportive leadership would engender higher levels of confidence in an employee, it is unclear
whether supportive leadership will engender confidence in all situations. In fact, Jones, James, and Bruni (1975) found that con-
fidence is related to perceptions of leadership. It is likely that the relationship between follower confidence and a superior's use of
supportive leadership techniques is a complex one, when one considers that expressions of support may not always be inter-
preted as ‘leadership.’

Hypothesis 1. A supportive leadership style will engender greater levels of follower confidence depending on the situation faced
by the follower.

3. Framing consequences

Given evidence that leaders tend to be more coercive when their followers' performance is lacking and when facing a potential
failure situation, it is important to consider the consequences the organization is facing along with the superior's leadership style.
One of the most important influences leaders have on their subordinates is how they frame consequences. Followers tend to mon-
itor what consequences the leader talks about as being important and use consequence information provided by the leader as
performance feedback (Weick, 1995).
Framing the consequences of failure at the organizational level may bolster embedded leaders' motivation and prompt them
to take a wider view of a situation. Alternatively, the increased attention that organizational level consequences bring may
J.J. Caughron, M.D. Mumford / The Leadership Quarterly 23 (2012) 342–353 345

distract the embedded leader from the task at hand. However, it is likely that organizational level consequences will be seen as
being more distal and less threatening than individual level consequences. Thus a higher level of organizational consequences
are likely to frame the problem as being at the ‘organizational level’ and encourage an embedded leader to rally organizational
members to solve the problem. This would be demonstrated as an embedded leader using higher levels of participative
leadership.

Hypothesis 2. Higher levels of organizational level consequences will cause a leader to become more participative in their
leadership.

While organizational level consequences are no doubt important to a conscientious middle manager, they are undoubtedly
less threatening than consequences directly targeting the individual. Facing personal level consequences suggests a certain
level of personal involvement in the outcome stemming from group performance. If an embedded leader's superior emphasizes
severe negative individual level outcomes, such as losing their job or being demoted, it is arguable that this will result in the em-
bedded leader feeling threatened. Should this occur, it is likely that the embedded leader will become self-focused, thinking less
about resolving the organizational problems at hand and more about dealing with personal outcomes. This would likely serve as a
distraction to the embedded leader.
In fact, in his examination of the Mann Gulch disaster, this is exactly what Weick suggests was a key contributor to the collapse
of leadership resulting in many preventable deaths (Weick, 1993). One's ability to engage in sensemaking seems to be compro-
mised by high levels of individual level consequences. What is unclear is whether or not this effect can been seen for sensemaking
in a middle management situation. In essence, a question remains as to how emphasizing personal level outcomes, especially se-
verely negative ones, impacts one's approach to leadership tasks at the middle levels of an organization. However, given Weick's
suggestions that individual level consequences hinder sensemaking we are left with the following hypothesis (Weick, 1993,
1995).

Hypothesis 3. Higher levels of individual level consequences will cause a decrement in a leader's ability to engage in
sensemaking.

4. Method

4.1. Sample

The sample consisted of 224 undergraduate students (44 males and 180 females) drawn from an introductory psychology
course at a large southwestern university. The study was announced via a website posting describing the study as a leadership
problem-solving study. The mean age of the participants was 19.5 years of age. The mean reported American College Testing
(ACT) exam score was 24.85, suggesting that these students represent a set of typical undergraduate college students, in terms
of demographics and general intelligence.

4.2. General procedures

Upon arriving at the study location, participants read and signed an informed consent form. The study was conducted in a sin-
gle 3-hour session divided into two blocks. The first block was half an hour long and involved a proctor guiding the participants
through a series of timed individual difference measures. The second block was scheduled for two and a half hours. During this
time, the participants were allowed to complete the remainder of the study materials at their own pace.
The primary experimental task was a low-fidelity simulation (Motowidlo, Dunnette, & Carter, 1990) consisting of a scenario in
which the participants assumed the role of a middle-manager involved in directing action of subordinates and giving opinions
about challenges faced by a mid-sized manufacturing organization. In this scenario, the middle-manager oversees the develop-
ment, production, and marketing of a new automobile and writes two open-ended responses to the CEO about how best to
turn around lagging sales. Additionally, they provide two open-ended responses to their subordinates giving them information
about the company's situation, plan of action, and directions on what to do next. The first response is used as a within subject
control because it was solicited from the participants before they encountered any manipulations. The manipulations were ad-
ministered in the form of emails from the CEO and the last three responses were coded for the dependent variables of interest:
participant confidence, use of participative leadership, and sensemaking.

4.3. Individual difference measures

Measures were administered in order to control for the role of individual differences. Participants' gender, personality, intel-
ligence, need for cognition, planning skill, and the number of superiors they have had in their work experience were examined as
covariates. A significant correlation between gender and participative leadership was found (r = 0.140; p b 0.05), suggesting that
females are more likely to engage in participative leadership than males. Prior research has shown that males and females show
differences in leadership style, especially where participative leadership is concerned (Jago & Vroom, 1982; Steers, 1977; Vroom &
Jago, 1995). Given the evidence for, and importance of, gender differences in leadership, the fact that gender showed a significant
346 J.J. Caughron, M.D. Mumford / The Leadership Quarterly 23 (2012) 342–353

correlation with participative leadership, and that there were many more females in this sample than males, it was included as a
control measure in the analysis of the participative leadership variable.
Participants' self-reported ACT score was a significant control measure in the analysis for the sensemaking variable and was
thus retained in that analysis. No other control variables demonstrated significant relationships with sensemaking, confidence,
or participation and were thus removed from the analyses.

4.4. Experimental manipulations

4.4.1. Organizational consequences


Each of the manipulations was written into the scenario, within the CEO's response to their meeting with the board of direc-
tors. Participants read this email after they gave their first response to the problem and before they gave their second response. In
the email, the CEO discusses organizational consequences that could occur if the problem is not resolved adequately. In the high
organizational consequences condition, the CEO chooses to present the board of directors with a plan that is extremely costly in
terms of money (approximately $141 million as stated in the vignette) and time (6–8 months) before the organization will begin
to see if the plan is helping the problematic business unit improve performance. In the low organizational consequences condi-
tion, the CEO presents the board of directors with a plan that is much less costly (approximately $1 million as stated in the vi-
gnette) and while requiring some downtime (4–6 weeks), will not delay production of the vehicle as much as the high
consequence condition.

4.4.2. Individual consequences


As with the organizational consequences manipulation, individual consequences were also manipulated at two levels. In the
high individual consequences condition, the CEO tells the participant that the board of directors was particularly disappointed
with his or her work and that they are considering demoting or laying off the participant. In the low individual consequences con-
dition, the CEO tells the participant that he expects him or her to work through the weekend in order to get a new report finished
by Monday.

4.4.3. Superior's leadership technique


The CEO's motivational strategy was set at three levels. The first level was that of a coercive response. In this condition the CEO
expresses anger with the participant and his or her coworkers in the scenario, questions his or her competence and commitment,
and reprimands him or her for poor performance. The second level of this manipulation was a supportive response from the em-
bedded leader's superior. In this condition the CEO still tells the participant and his or her coworkers that the board meeting did
not go well but emphasizes that they were not to blame and that the primary cause of the problem was his choice of exactly
which material to present to the board of directors. In this condition the CEO is much more supportive and encouraging to the
participant, takes responsibility for what the CEO himself could have done differently, and encourages the employees to move for-
ward in a positive direction. The final condition within this manipulation is that of a passive response. In this condition the CEO
still mentions that the board meeting did not go well, but does not give the employees feedback regarding their performance, nor
does he mention taking responsibility himself for any mistakes that may have occurred.

4.4.4. Content coding


Content coding was used to measure a variety of variables in this study. These variables have been identified as important for
leaders to exhibit in performing their duties and include sensemaking (Weick, 1995), demonstrating participative leadership
(Evans, 1970; Georgopoulos, Mahoney, & Jones, 1957; House, 1971) and confidence (Black & Porter, 2000; Chemers et al.,
2000; Northouse, 1997; Post, 1986).
The four judges involved in this content coding effort were senior-level graduate students working toward obtaining their PhD
in I/O psychology. During their training the judges were introduced to operational definitions for each of the variables to be coded.
Additionally, time was spent during each training session rating materials and comparing ratings on a subset of materials drawn
from the participants' responses to the stimulus materials. Ratings for each construct were made on a 5-point Likert scale. Discus-
sions were held when judges did not agree on how to rate a given response until the judges had a minimum reliability of 0.70 on
ten items drawn from the participant materials for each construct they were rating. After this was achieved, the judges were given
the rest of the participant materials to rate and reliabilities were checked again at the end of the study. The judges were blind to
the participants' conditions. Judges were each given a manual describing the rating strategy, which included definitions of each
construct, markers that highlighted key aspects of the construct, and example materials drawn from participant responses repre-
senting high, medium, and low performance on each construct. Table 1 presents the construct label, definition, reliability, and rel-
evant citations for each of these variables.

5. Results

Table 2 presents the correlations between the leadership dependent variables. In order to test the hypotheses regarding dif-
ferences in behavior by embedded leaders, a series of analysis of covariance (ANCOVA) analyses were conducted. This procedure
allowed the researchers to control for the participant's pre-manipulation response, relevant covariate measures, and test the re-
lationship of the manipulations on the participants' post-manipulation responses. One ANCOVA analysis was conducted for each
J.J. Caughron, M.D. Mumford / The Leadership Quarterly 23 (2012) 342–353
Table 1
Construct definitions and reliability estimates.

Construct Reliability Definition Source

Confidence 0.702 The degree to which the participant seems sure of him or herself as they solve problems, give directions, or interact Hollenbeck & Hall, 2004
with others. A participant who expresses self-doubt, prevaricates about the action steps, or discusses multiple options
without making a decision should be rated as low in confidence. Participants who give directions without prevaricating,
make firm decisions, or express confidence in themselves or team members should be rated as being high in confidence.
Sensemaking 0.734 When solving a problem, it is important that people think about how their position in their group, organization, and Weick, 1995
society relate to the origins of the problem, individuals involved, and relevant principles, goals & values.
Participative decision- 0.818 The extent to which the individual expresses a desire to consult with subordinates, gather information from others, Evans, 1970; Georgopoulos et al., 1957; House, 1971
making asks for opinions or suggestions, or encourages others to make decisions.

347
348 J.J. Caughron, M.D. Mumford / The Leadership Quarterly 23 (2012) 342–353

Table 2
Correlations among leadership outcome variables.

Age Gender ACT Confidence Sensemaking Participative

Age
Gender − 0.029
ACT score − 0.009 0.140⁎
Confidence 0.051 − 0.102 0.045
Sensemaking 0.156 − 0.083 0.059 0.496⁎⁎
Participative leadership 0.128 − 0.172⁎⁎ 0.061 0.316⁎⁎ 0.605⁎⁎
⁎ Denotes p b 0.05.
⁎⁎ Denotes p b 0.01.

leadership outcome variable so that the participant's pre-manipulation response could be used as a covariate for that variable
only. This included the participant's confidence, sensemaking, and participative leadership. Table 3 presents the between-
subject results for all dependent variables.
Table 4 presents the estimated marginal means for the significant interaction between the organizational consequences and
superior's leadership strategy manipulation on the confidence variable (F = 4.47; p = 0.011). These means are charted on graphs
in Fig. 1. Table 4, along with Fig. 1, shows a decline in the participants' confidence when organizational consequences were low
and the participant's superior used a passive or supportive leadership technique. However, participant confidence was highest
when the leader used a coercive strategy in a low organizational consequences situation or a passive or supportive approach
when organizational consequences were high. Overall this supports Hypothesis 1 in that supportive leadership enhanced partic-
ipant confidence but only in certain circumstances. When organizational consequences were high confidence increased with sup-
portive leadership. Alternatively, coercive leadership increased confidence when organizational consequences were low.
Tables 5a and 5b along with Fig. 2 show the estimated marginal means for the three-way interaction between organizational
consequences, individual consequences, and the superior's leadership strategy on the confidence variable. The highest levels of
confidence were found for a supportive superior paired with high organizational and low individual level consequences. This
suggests that a superior can help their followers feel more confident when they use a supportive strategy and deemphasize indi-
vidual level consequences when facing high organizational level consequences.

Table 3
ANCOVA results for outcomes controlling for participants' pre-measure score.

Variable F df p η2

Confidence
Organizational consequences 2.59 1, 212 0.084 0.014
Individual consequences 0.12 1, 212 0.627 0.001
Superior's leadership strategy 1.84 2, 212 0.222 0.014
Org × indiv consequences 0.77 1, 212 0.393 0.003
Org × superior's leadership 4.47 2, 212 0.011⁎ 0.042
Indiv × superior's leadership 0.59 2, 212 0.600 0.005
Org × indiv × superior's leadership 2.89 2, 212 0.046⁎ 0.029

Sensemaking
ACT control measure 7.30 1, 211 0.007⁎⁎ 0.033
Organizational consequences 1.41 1, 211 0.236 0.007
Individual consequences 8.69 1, 211 0.004⁎⁎ 0.040
Superior's leadership strategy 2.63 2, 211 0.075 0.024
Org × indiv consequences 0.12 1, 211 0.727 0.001
Org × superior's leadership 1.20 2, 211 0.303 0.011
Indiv × superior's leadership 1.00 2, 211 0.371 0.009
Org × indiv × superior's leadership 0.89 2, 211 0.414 0.008

Participative leadership
Gender 0.230 1, 211 0.632 0.001
Organizational consequences 4.73 1, 211 0.031⁎ 0.022
Individual consequences 0.01 1, 211 0.907 0.000
Superior's leadership strategy 1.25 2, 211 0.288 0.012
Org × indiv consequences 0.01 1, 211 0.909 0.000
Org × superior's leadership 0.82 2, 211 0.441 0.008
Indiv × superior's leadership 0.12 2, 211 0.885 0.001
Org × indiv × superior's leadership 1.99 2, 211 0.140 0.018
⁎ Denotes p b 0.05.
⁎⁎ Denotes p b 0.01.
J.J. Caughron, M.D. Mumford / The Leadership Quarterly 23 (2012) 342–353 349

Table 4
Confidence (organizational consequences × superior's leadership strategy).

Organizational consequences

Low High

M SD M SD

Superior's leadership strategy Coercive 3.46 0.56 3.28 0.52


Passive 3.24 0.42 3.40 0.43
Supportive 3.29 0.45 3.45 0.52

However, the next highest level of confidence occurs when the participant's supervisor used a coercive leadership style paired
with high individual and low organizational consequences. At face value, this seems to be a counter intuitive finding. That when a
leader emphasizes individual level consequences they should pair that with a coercive leadership style in order to engender
follower confidence, so long as the organizational level consequences are low. However, previous findings have suggested that
follower confidence is related to perceptions of leadership (Jones et al., 1975). It is possible that participants interpreted a coer-
cive leader emphasizing individual level consequences when organizational consequences were low as a display of leadership.
However, it should be noted that the lowest mean level of confidence occurred with coercive leadership with high levels of
both individual and organizational consequences, suggesting that coercion may facilitate follower confidence but only in a nar-
rowly defined set of circumstances.
Interestingly, confidence was the lowest when the situation had low consequences on both individual and organizational
levels and the leader used a passive or supportive approach. While the reasons for this finding need to be researched further, it
is likely that in low consequence situations the participants did not see the situation as a ‘leadership’ task as much as a more rou-
tine management task. If this was the case then it is likely that they would not have felt the need to make overt expressions of
confidence when they interacted with their followers.
One last note about Fig. 2, low levels of confidence occurred when participants faced a very high stakes situation in that the
organization and the individual were both facing severe negative outcomes. The passive leadership approach engendered more
confidence than either a supportive or coercive approach, with coercive being the lowest. It is likely that this results from the
fact that the superior's attempts to manage the participant's motivation and ability to solve the problem at hand by being sup-
portive or coercive were either distracting (in the case of the supportive condition), discouraging (in the case of the coercive con-
dition), or both. It appears that the participants who were allowed to develop and pursue their own solution to a high stakes
problem were the most confident when their superior did not attempt to manipulate their motivation level using an interpersonal
strategy.
Table 6 shows the estimated marginal means for the main effect of individual level consequences on sensemaking. When in-
dividual level consequences were low participant sensemaking was higher (M = 2.90; SD = 0.66) than when individual conse-
quences were high (M = 2.73; SD = 0.73). The interpretation here is straightforward. The less the participants needed to focus
on individual level consequences the more effective they were at engaging in sensemaking. This is entirely in line with Weick's
work on sensemaking. His analysis of the Mann Gulch disaster suggest that high levels of individual level consequences can
cause an individual to draw within themselves and tunnel vision, resulting in lower quality sensemaking (Weick, 1993). This find-
ing supports Hypothesis 3. The importance of this finding hinges on the fact that it is being demonstrated in a simulation of a mid-
dle management task. This effect has been suggested as a causal agent in retroactive reviews of life and death situations and has
been theorized as a causal factor in organizations. However, the fact that it is found experimentally adds validity to these theories
and opens the door for more work on sensemaking as it occurs in the middle layers of organizations.

Note: All scales range from 1-5

Fig. 1. Effect of Organizational consequences × superior leadership technique interaction on confidence.


350 J.J. Caughron, M.D. Mumford / The Leadership Quarterly 23 (2012) 342–353

Table 5a
Confidence in low organization consequence condition (organizational consequences × individual consequences × superior's leadership strategy).

Individual consequences

Low High

M SD M SD

Superior's leadership strategy Coercive 3.41 0.49 3.53 0.64


Passive 3.23 0.46 3.26 0.39
Supportive 3.25 0.53 3.33 0.38

Table 7 shows the estimated marginal means for the main effect of organizational consequences on participative leadership. As
predicted in Hypothesis 2, participants showed higher levels of participative leadership when organizational consequences were
higher (Mlow = 2.18; SDlow = 0.67 vs. Mhigh = 2.38; SDhigh = 0.66). This suggests that when an embedded leader's superiors em-
phasize organizational level consequences they cause the embedded leader to use a more participative leadership style with
their subordinates.

6. Discussion

Before turning to the implications of these findings, a number of limitations should be noted. First, the study uses undergrad-
uates participating in a low-fidelity simulation task. Although evidence has accrued attesting to the validity of using this sample
along with this type of task in order to measure cognitive and behavioral variables (Motowidlo et al., 1990; Wintre, North, &
Sugar, 2001), it is possible that using a different sample will give somewhat different results. It is likely that real middle managers
faced with real organizational or individual level consequences would actually respond in a much stronger fashion, although the
direction of the relationships is likely to remain the same. This would also likely increase the average effect size of the findings
shown here, which are actually quite low. It should be noted, however, that this is a study about cognition, decision-making,
and how these are influenced by one's superiors. Given that these issues do not touch on developmental issues known to change
in adulthood, the findings here should be consistent with findings that would be found using an older sample. The only notable
challenge to this notion would be the experience gained as one ages in their career. This is a noteworthy challenge and should be
addressed in future research.
Additionally, this study used a pre–post manipulation design. This provides for two layers of control; one stemming from the
fact that the study was conducted in a lab and the other stemming from the fact that the participants served as their own statis-
tical controls. This design provides for superior levels of control, especially when compared to field research. Similarly, it allows
researchers to make definitive statements about causality that cannot be made in correlation based research. However, one
should take note of the small effect sizes demonstrated in this study and bear them in mind when interpreting the results. Effect
sizes falling below 0.20 are generally considered small, and the effect sizes here are far below 0.20. This suggests that the relation-
ship between the manipulated variables and the dependent variables is small. It is unclear whether this is due to the technique
used to measure these variables or if the relationship between these variables is actually a weak one. Future research on this
issue may help clarify this issue.
Bearing these limitations in mind, this study makes several important contributions. The findings from this study indicate that
superiors must use caution when attempting to engender confidence in their followers. Supportive leadership was found to en-
gender greater confidence when the embedded leaders were facing high organizational consequences paired with low levels of
individual level consequences. However, coercive leadership increased embedded leader confidence in certain circumstances.
Findings here suggest that superior coercion can cause an embedded leader to approach their tasks with greater confidence,
when they emphasize individual level consequences in a low organizational consequence situation. One might expect that pairing
supportive leadership techniques with high individual level consequences would be helpful. However, when one considers the
practical application of this strategy it becomes clear that pairing expressions of support with an emphasis on personal level con-
sequences is probably not effective for promoting confidence. If one's superior makes a point to mention the personal conse-
quences at stake and then starts using a supportive leadership approach this is a very mixed message. This behavior is likely to
come across as either manipulative or confusing; neither of which are likely to increase follower confidence. It is also important

Table 5b
Confidence in high organization consequence condition (organizational consequences × individual consequences × superior's leadership strategy).

Individual consequences

Low High

M SD M SD

Superior's leadership strategy Coercive 3.33 0.50 3.23 0.54


Passive 3.37 0.36 3.43 0.51
Supportive 3.61 0.58 3.30 0.40
J.J. Caughron, M.D. Mumford / The Leadership Quarterly 23 (2012) 342–353 351

Note: All scales range from 1-5

Fig. 2. Effect of Organizational consequences × individual consequences × superior's leadership strategy interaction on confidence.

to note, however, that the use of a coercive technique did cause the lowest levels of confidence in other situations. Thus, while
being coercive might spur followers on to being confident in a few situations, it appears that those situations are constrained
to a very limited set of circumstances.
It is important to bear these findings in mind when one considers the fact that real-world organizational leaders often put a
coercive leader in charge of high-stakes organizational events. This study suggests that coercive leadership can engender the
most confidence when organizational consequences are lowest, not highest. Further, these findings suggest that when middle
management is facing high organizational consequences, such as the expenditure of large sums of money or large scale changes
in organizational strategy, their confidence is likely to be boosted by bringing in a supportive leader who does not emphasize se-
vere personal level consequences.
However, it also appears that being neither supportive nor coercive can actually be helpful in some circumstances. Findings
here suggest that organizations should bring in a leader who will allow the middle managers to focus on their task without en-
gaging in overt attempts to spur subordinate confidence when facing high levels of both organizational and personal conse-
quences; apparently, trying to spur middle managers' confidence in an overt manner when they are facing a high stakes
situation can cause them to lose confidence. Perhaps embedded leaders interpret these attempts as a lack of trust from their su-
perior. This is an intriguing finding and more research should be conducted to flesh out this issue.
Emphasizing individual level consequences also had an interesting effect on sensemaking. It is not surprising that embedded
leader's sensemaking behaviors were lower when faced with higher levels of individual consequences. These findings agree with
Weick's analysis regarding the Mann–Gulch disaster, it is likely that personal level consequences, especially severe consequences,
act to promote a more individualistic cognitive processing style (Weick, 1993). When facing negative personal outcomes, individ-
uals begin to look for a way to take care of themselves rather than focusing on resolving the larger problem at hand. Finding this
effect for approaching middle management situations is worth noting. This suggests that organizational leaders should be mindful
when selecting the consequences they wish to emphasize with their employees. If those employees are working to create a co-
herent and actionable understanding of an ambiguous situation, it is best to refrain from emphasizing personal level outcomes.
In line with the findings regarding sensemaking, it appears that emphasizing organizational outcomes may be helpful for en-
couraging participative leadership. Framing consequences at the organizational level appears to encourage middle managers to
think in a more collective fashion and recognize the problem as on that should be faced as a group. By emphasizing organizational
level outcomes, the embedded leader's superior is able to focus attention on a collective level problem and encourages a collective
response, rather than encouraging the individual to approach the problem from an individualistic perspective.
One issue that is made very clear by this study is that all consequences are not created equal. The outcomes one's superior em-
phasizes shapes the way those under his or her guidance think about and approach their work. This is an important issue to con-
sider for organizational leaders seeking to motivate their employees. Emphasizing certain outcomes involves more than
influencing which outcomes a subordinate monitors and works to improve, as has been suggested in the past (Weick, 1995).
The findings here suggest that upper level managers can hinder their middle manager's ability to understand organizational

Table 6
Sensemaking.

M SD

Individual consequences Low 2.90 0.66


High 2.73 0.73
352 J.J. Caughron, M.D. Mumford / The Leadership Quarterly 23 (2012) 342–353

Table 7
Participative decision-making.

M SD

Organizational consequences Low 2.18 0.67


High 2.38 0.66

problems by emphasizing individual level consequences. If these managers cannot make sense of organizational problems their
ability to create a cohesive plan of action and promote organizational success will be limited. Additionally, emphasizing organi-
zational consequences helps mid-level leaders use a more participative leadership style. Given the complexity of the problems
many organizations face a participative leadership style is likely to bring together individuals who have a wide range of expertise
and talents and provide higher quality solutions to organizational problems.
A few clear cut implications can be made from these findings. First, organizational leaders should take care when they choose
which consequences to emphasize to their subordinates. Emphasizing the personal consequences for one's subordinate should be
undertaken with great care as it can undermine confidence and derail sensemaking. However, emphasizing organizational level
consequences can boost confidence when paired with supportive comments from the supervisor. Additionally, emphasizing or-
ganizational level consequences can promote participative leadership in one's subordinates; a type of leadership behavior that
is associated with a many positive organizational level outcomes.
Lastly, coercive leadership tactics should be used with caution. While there is evidence to suggest that coercive tactics are like-
ly to be used in certain circumstances (Bankhart & Lanzetta, 1970; Barrow, 1976; Hinton & Barrow, 1975), findings here suggest
that the use of coercion can hinder the development of follower confidence except in a very narrowly defined set of circum-
stances. Specifically, here coercion was shown to promote subordinate confidence when paired with emphasizing the severe con-
sequences for the subordinate, but only when it is clear that organizational consequences are low. While the use of coercion is
more likely when facing failure in high-stakes situations and when the follower's ability to perform is in question, findings
here suggest that coercion is actually only helpful in promoting confidence in low-stakes situations.
Future research can build upon these findings by continuing to examine how superiors influence their subordinate leaders.
This includes how superiors discuss consequences of failure with their subordinates bearing in mind that consequences operate
on multiple levels and are not all of one type. Specifically, examining the effect of emphasizing consequences for a middle man-
ager's subordinates is likely to reveal interesting effects, perhaps causing the middle manager to engage in higher levels of par-
ticipative leadership or changing the degree to which they use coercive or supportive leadership themselves. Additionally,
considering that coercion actually increased follower confidence in some circumstances, further research could also examine
how and when a subordinate is likely to accept coercive leadership.
Lastly, future research on middle management is warranted above and beyond the variables examined here. Middle managers
engage in many tasks that are unique within the organization, such as communicating and influencing others in an upward and
downward direction, identifying and allocating resources, and creating workarounds that help organizational members pursue
appropriate goals and avoid obstacles within the organization. These tasks can and should be examined in both applied settings
but also in the laboratory where superior levels of control can be used to manipulate causal variables and identify cause and effect
relationships beyond correlational associations.
Middle managers serve simultaneously as leaders and followers. Findings here suggest that their superiors can have a substan-
tial impact on their behavior by choosing which outcomes to emphasize and by taking a coercive or supportive leadership style.
This is a noteworthy finding given that a great deal of leadership occurs at the middle levels of an organization; especially when
one considers how poorly ‘leadership in the middle place’ is understood (DeChurch et al., 2010). Middle managers are often re-
sponsible for accomplishing the day-to-day tasks that enable an organization to function smoothly and pursue the goals handed
down from top level leaders. It can also help high level leaders expand their influence over lower level employees (Chun et al.,
2009). Knowing how to facilitate specific types of leadership behavior can enable top level organizational leaders to actively
influence the type of leadership behaviors middle managers demonstrate and ultimately spur organizational performance.

References

Aspergren, R. E. (1963). A study of leadership behavior and its effects on morale and attitudes in selected elementary schools. Dissertation Abstracts, 23, 3708.
Bankhart, C. P., & Lanzetta, J. (1970). Performance and motivation as variables affecting the administration of rewards and punishments. Representative Research in
Social Psychology, 1, 1–10.
Barrow, J. C. (1976). Worker performance and task complexity as causal determinants of leader behavior, style, and flexibility. Journal of Applied Psychology, 61,
433–440.
Bass, B. M. (1958). Leadership opinions as forecasts of supervisory success: A replication. Personnel Psychology, 11, 515–518.
Bass, B. M. (1960). Leadership, psychology, and organizational behavior. New York: Harper.
Bass, B. M. (2008). Handbook of leadership. New York, NY, US: The Free Press.
Black, J. S., & Porter, L. W. (2000). Management: Meeting new challenges. Upper Saddle River, NJ: Prentice Hall.
Caughron, J. J., Shipman, A. S., Beeler, C. K., & Mumford, M. D. (2009). Social innovation: Thinking about changing the system. The International Journal of Creativity
& Problem Solving, 1, 7–32.
Chemers, M. M., Watson, C. B., & May, S. T. (2000). Dispositional affect and leadership effectiveness: A comparison of self-esteem, optimism, and efficacy. Person-
ality & Social Psychology Bulletin, 26, 267–277.
Chun, J. U., Yammarino, F. J., Dionne, S. D., Sosik, J. J., & Moon, H. K. (2009). Leadership across hierarchical levels: Multiple levels of management and multiple
levels of analysis. The Leadership Quarterly, 20, 689–707.
J.J. Caughron, M.D. Mumford / The Leadership Quarterly 23 (2012) 342–353 353

Dansereau, F., Yammarino, F. J., Markham, S. E., Alutto, J. A., Newman, J., Dumas, M., et al. (1995). Individualized leadership: A new multiple-level approach. The
Leadership Quarterly, 6, 413–450.
Day, D. V., Gronn, P., & Salas, E. (2004). Leadership capacity in teams. The Leadership Quarterly, 15, 857–880.
DeChurch, L. A., Hiller, N. J., Murase, T., Doty, D., & Salas, E. (2010). Leadership across levels: Levels of leaders and their levels of impact. The Leadership Quarterly,
21, 1069–1085.
Evans, M. G. (1970). Leadership and motivation: A case concept. Academy of Management Journal, 13, 91–102.
Floyd, S. W., & Wooldridge, B. (1992). Middle management involvement in strategy and its association with strategic type: A research note. Strategic Management
Journal, 13, 153–167.
Floyd, S. W., & Wooldridge, B. (1994). Dinosaurs or dynamos? Recognizing middle management's strategic role. The Academy of Management Executive, 8, 47–57.
Fox, W. M. (1954). An experimental study of group reaction to two types of conference leadership. Doctoral dissertation, Ohio State University, Columbus.
Fox, W. M. (1957). Group reaction to two types of conference leadership. Human Relations, 10, 279–289.
Georgopoulos, B. S., Mahoney, G. M., & Jones, N. W., Jr. (1957). A path-goal approach to productivity. Journal of Applied Psychology, 41, 345–353.
Harrison, T. M. (1985). Communication and participative decision making: An exploratory study. Personnel Psychology, 38, 93–116.
Hershey, D. A., & Walsh, D. A. (2001). Knowledge versus experience in financial problem solving performance. Current Psychology, 19, 261–292.
Hinton, B. L., & Barrow, J. C. (1975). The superior's reinforcing behavior as a function of reinforcements received. Organizational Behavior and Human Performance,
14, 123–149.
Hollenbeck, G. P., & Hall, D. T. (2004). Self-confidence and leader performance. Organizational Dynamics, 33, 254–269.
House, R. J. (1971). A path goal theory of leader effectiveness. Administrative Science Quarterly, 16, 321–338.
Jacobs, T. O., & McGee, M. L. (2001). Competitive advantage: Conceptual imperatives for executives. In S. Zaccaro, & R. J. Klimoski (Eds.), The nature of organiza-
tional leadership: Understanding the performance imperatives confronting today's leaders (pp. 42–78). San Francisco, CA: Jossey-Bass.
Jago, A. G., & Vroom, V. H. (1982). Sex differences in the incidence and evaluation of participative leader behavior. Journal of Applied Psychology, 67, 776–783.
Jones, A. P., James, L. R., & Bruni, J. R. (1975). Perceived leadership behavior and employee confidence in the leader as moderated by job involvement. Journal of
Applied Psychology, 60, 146–149.
Kaiser, R. B., Hogan, R., & Craig, S. B. (2008). Leadership and the fate of organizations. American Psychologist, 63, 96–110.
Kipnis, D. (1976). The powerholders. Chicago: University of Chicago Press.
Lanzetta, J. T., & Roby, T. B. (1960). The relationship between certain group process variables and group problem-solving efficiency. Journal of Social Psychology, 52,
135–148.
Likert, R. (1961). New patterns of management. New York: McGraw-Hill.
Likert, R. (1961). an emerging theory of organizations, leadership and management. In L. Petrullo, & B. M. Bass (Eds.), Leadership and interpersonal behavior. New
York: Holt, Rinehart, & Winston.
Lorge, I., Fox, D., Davitz, J., & Brenner, M. (1958). A survey of studies contrasting the quality of group performance and individual performance, 1920–1957.
Psychological Bulletin, 55, 337–370.
Mann, F. C., & Baumgartel, H. (1952). Absences and employee attitudes in an electric power company. Ann Arbor: University of Michigan, Survey Research Center.
Motowidlo, S. J., Dunnette, M. D., & Carter, G. W. (1990). An alternative selection measure: The low-fidelity simulation. Journal of Applied Psychology, 75, 640–647.
Northouse, P. G. (1997). Leadership theory and practice. Thousand Oaks, CA: Sage Publications.
Post, J. M. (1986). Narcissism and the charismatic leader–follower relationship. Political Psychology, 7, 675–688.
Preston, M. G., & Heintz, R. K. (1949). Effects of participatory vs. supervisory leadership on group judgment. Journal of Abnormal and Social Psychology, 44,
345–355.
Raven, B. H., & French, J. R. P. (1958). Group support, legitimate power, and social influence. Journal of Personality, 26, 400–409.
Raven, B. H., & French, J. R. P. (1958). Legitimate power, coercive power, and observability in social influence. Sociometry, 21, 83–97.
Smith, M. (1948). Control interaction. Journal of Social Psychology, 28, 263–273.
Steers, R. M. (1977). Individual differences in participative decision-making. Human Relations, 30, 837–847.
Vroom, V., & Jago, A. (1995). Situation effects and levels of analysis in the study of leader participation. The Leadership Quarterly, 6, 169–181.
Weick, K. (1993). The collapse of sensemaking in organizations: The Mann gulch Disaster. Administrative Science Quarterly, 38, 628–652.
Weick, K. (1995). Sensemaking in organizations. Beverly Hills, CA: Sage.
Wintre, M., North, C., & Sugar, L. (2001). Psychologists' response to criticism about research based on undergraduate participants: A developmental perspective.
Canadian Psychology/Psychologie Canadienne, 42, 216–225.
Yammarino, F. J. (1994). Indirect leadership: Transformational leadership at a distance. In B. M. Bass, & B. J. Avolio (Eds.), Improving organizational effectiveness
through transformational leadership (pp. 26–47). Thousand Oaks, CA: Sage.
Yammarino, F. J., & Bass, B. M. (1991). Person and situation views of leadership: A multiple levels of analysis approach. The Leadership Quarterly, 2, 121–139.
Yammarino, F. J., & Dansereau, F. (2009a). A new kind of organizational behavior. Multi-level Issues in Organizational Behavior and Leadership Research in Multi-level
Issues, 8, 13–60.
Yammarino, F. J., & Dansereau, F. (2009b). A new kind of organizational behavior. Research in Multi-Level Issues, 8 (Multi-Level Issues in Organizational Behavior and
Leadership) (pp. 13–60).
Yukl, G. (2006). Leadership in organizations (6th ed.). Upper Saddle River, NJ: Pearson-Prentice Hall.
Ziller, R. C. (1954). Four techniques of decision making under uncertainty. American Psychologist, 9, 498.

You might also like