2

The Measurement of
Transformational Leadership

Although the concept of transformational leadership struck a sym-

pathetic chord with those generally interested in leadership topics,
as evidenced by the popularity of Burnss (1978) book, the surge of
research on transformational leadership is due in large part to the
development of measurement tools to assess the construct. The most
widely accepted instrument to measure transformational leadership
is the Multifactor Leadership Questionnaire (MLQ; Bass & Avolio,
2000), which actually assesses the Full Range of Leadership (FRL)
model, including laissez-faire leadership; the components of transac-
tional leadership, namely, management by exception (both active and
passive forms); and contingent reward, as well as the components of
transformational leadership. Although much of the research on trans-
formational leadership uses the MLQ, it is important to note that other
measures of transformational leadership exist, and these are reviewed
later in this chapter.

DEVELOPMENT OF THE MLQ

Burnss (1978) concept of the transforming leader, along with the

responses of 70 senior executives who reported that they had experi-
ences with transformational leaders, were used to elicit accounts of
leaders who t the description. These were converted to 142 behavioral
statements. Eleven judges were presented with detailed denitions
of transactional and transformational leadership, and these judges

19
20 CHAPTER 2

agreed that 73 of the 142 statements reected either transformational

or transactional leader behaviors.
The 73 items representing transactional and transformational be-
haviors were presented to 176 high-ranking U.S. military ofcials,
who rated how well each item pertained to their immediate superi-
ors. Initial factor analyses of these data, along with subsequent factor
analyses (Bycio et al., 1995; Howell & Avolio, 1993), suggested several
factors. Three factors were extracted from the transactional leadership
items and four factors from the transformational items.
It was originally believed that there were three components to
transformational leadership: charismatic-inspirational, intellectually
stimulating, and individually considerate. However, later factor analy-
ses suggested that the charisma factor, what has been subsequently
termed Idealized Inuence, was separate from the inspiration factor
(Inspirational Motivation). More recent research on the factor structure
of the MLQ is discussed later in this chapter.
The transactional items formed factors of contingent reward, man-
agement-by-exception, and laissez-faire leadership behaviors. Subse-
quent factor analyses supported splitting management-by-exception
into passive and active factors (Hater & Bass, 1988). Combined, these
transactional and transformational leadership factors formed the early
FRL model.
There are two forms of the MLQ. The rst is the Leader Form that
asks the leader to rate the frequency of his or her own leader behavior.
Research has shown that self-ratings of ones own leader behavior are
prone to bias. Therefore, the more important version of the MLQ is the
Rater Form. The MLQ Rater Form requires associates of leaders (usu-
ally supervisees or direct reports) to rate the frequency of their leaders
transactional and transformational leadership behavior using 5-point
ratings scales, with anchors ranging from 0 = Not at all to 4 = Frequently,
if not always. The Rater Form is most commonly used in research to
measure transformational and transactional leadership. Sample items
from the MLQ are presented in Table 2.1.

Revisions to the MLQ

As noted, the original MLQ consisted of 73 items, measuring ve fac-
tors. Responding to criticisms about the incorporation of items that
did not focus directly on leader behaviors (Hunt, 1991; Yukl, 1998)
and concerns about the factor structure and subscales, the MLQ was
MEASUREMENT 21

TABLE 2.1
Sample Items From the MLQ (5X)

Factor Sample Item

Idealized Inuence (Attributed My leader instills pride in me for being

Charisma) associated with him or her.
Idealized Inuence (Behaviors) My leader species the importance of having
a strong sense of purpose.
Inspirational Motivation My leader articulates a compelling vision of
the future.
Intellectual Stimulation My leader seeks differing perspectives when
solving problems.
Individualized Consideration My leader spends time teaching and coaching
Contingent Reward My leader makes clear what one can expect to
receive when performance goals are achieved.
Management-by-Exception My leader focuses attention on irregularities,
(Active) mistakes, exceptions, and deviations from
standards.
Management-by-Exception My leader shows that he or she is a rm
(Passive) believer in If it aint broke, dont x it.
Laissez-Faire My leader delays responding to urgent
requests.

substantially revised. The rst published version of the MLQ (Bass &
Avolio, 1990c) contained 67 items measuring the FRL Model (with 37
of these items assessing transformational leadership). Also included
were nine items measuring outcomes, including ratings of the leaders
effectiveness, satisfaction with the leader, and the extent to which fol-
lowers exert extra effort as a result of the leaders performance.
The current, revised form of the MLQ (5X) (Bass & Avolio, 1997) is
substantially rened and contains 36 standardized items, 4 items as-
sessing each of the nine leadership dimensions associated with the FRL
model (see Table 2.1), and the additional 9 outcome items mentioned
previously. In addition to this version of the MLQ, there is a longer
form (63 items) used for providing feedback in leadership develop-
ment programs (see chap. 10).

Psychometric Properties of the MLQ

Several different approaches have been used to conrm the reliability
and validity of the MLQ. A rst question is the integrity of the MLQ
scales themselves. Do the MLQ scales have good internal consistency?
22 CHAPTER 2

In other words, is there evidence that the items within each of the MLQ
scales hang together and seem to be measuring the same construct? The
MLQ scales have demonstrated good to excellent internal consistency,
with alpha coefcients above the .80 level for all MLQ scales, using
the most recent version of the MLQ across a large sample. Some of the
earlier MLQ scales (particularly the active management-by-exception
scale) had lower levels of internal consistency, but they were still ad-
equate given the small number of items per scale. The MLQ has been
completed by more than 15,000 respondents and translated into many
languages, ranging from German and French to Japanese and Hebrew.

Rate-Rerate Consistency. A second question concerns the consis-

tency of MLQ assessments. If we know a leaders MLQ prole, can we
predict the leaders prole at a later date? The evidence is supportive.
When 6 months intervene between the rst and second assessments of
the MLQ by self and subordinates, Pile (1988) found that the rst set of
results predicted the second set with correlations obtained as shown in
Table 2.2. Subsequent test-retest reliability coefcents have been even
higher (Bass & Avolio, 1997).

SubordinateSuperior Agreement. If supervisors complete the

MLQ on a target leader, does it agree with the ratings of that target

TABLE 2.2
MLQ RateRerate Correlations

Rate-Rerate Rate-Rerate
Leader Self-Ratings Follower Ratings
MLQ Scale (N = 33) (N = 193)
Transformational
Idealized Inuence (Charismatic .60 .79
Behavior)
Inspirational Motivation .45 .66
Intellectual Stimulation .61 .66
Individualized Consideration .70 .77
Transactional
Contingent Reward .44 .52
Management-by-Exception .74 .61
Laissez-Faire .73 .82

Note. From Visionary Leadership: Creating a Generative Internal Map, by S. Pile, 1988, unpub-
lished masters thesis, Pepperdine University, Los Angeles. Copyright 1988 by Pile. Reprinted
with permission.
MEASUREMENT 23

leaders subordinates? There is evidence of general agreement, but

such different perspectives also present unique pictures of the focal,
target leader. For example, Atwater and Yammarino (1993) found that
of the 107 midshipmen at the U.S. Naval Academy, subordinates and
superiors of the focal midshipmen correlated .35 and .34, respectively,
for composite measures of transformational and transactional leader-
ship. Considering that subordinates saw their focal leaders in a quite
different relationship than did the superiors of the focal leaders, the
correlations of .35 and .34 indicated that considerable variance in the
MLQ was attributable to consistent individual differences in the focal
midshipmen.

Peer Ratings Based on Performance in Small Groups. Participants

in the FRL development program, which is described briey in chapter
10, used a list of 21 items to assess the components of the MLQ model
displayed by their peers in the training program. The mean correlation
between the sets, each of three items, dealing with the seven compo-
nents of the model of transactional and transformational leadership,
correlated .35 on the average with the same components of the MLQ
as obtained from the participants subordinates back on the job prior
to beginning the basic program.
There are also some data to suggest that peers and observers of
small group discussions may be able to predict with some validity the
transformational and transactional behavior of examinees placed in a
leaderless group-testing situation because, as Bass (1954) concluded,
performance in a leaderless group discussion (LGD) correlates sig-
nicantly with the examinees subsequent leadership potential, status,
esteem, and merit. The rationale for such positive correlations was
that the LGD requires examinees to be cooperative and competitive,
often at the same time, mirroring an important challenge leaders face
in everyday circumstances (Handy, 1994).

Evidence of Construct Validity of the MLQ. As mentioned, the

MLQ, in its various forms, has been subjected to extensive factor
analyses to examine both the model of transformational leadership,
the larger FRL theory (FRLT), as well as the question of whether the
MLQ adequately measures these constructs. For example, a number
of researchers have been unable to replicate the nine-factor FRLT (e.g.,
Bycio et al., 1995; Den Hartog, Van Muijen, & Koopman, 1997; Tejeda,
Scandura, & Pillai, 2001), although they did nd evidence of many
24 CHAPTER 2

of the hypothesized MLQ dimensions. Typically, these researchers

found fewer than the proposed nine factors, suggesting that factors
could be combined or collapsed. However, recent research (Antonakis
et al., 2003) suggested that the inconsistent ndings of some of these
researchers occurred because they used very heterogeneous samples
of leaders from different cultures, organizational types, and organiza-
tional levels (e.g., combining front-line and upper level managers).
Although there is some evidence of inconsistency in the MLQs
factor structure, we think it is important to distinguish the theorized
subcomponents of transformational leadership and the FRL model for
conceptual clarity and for leadership development purposes. Trans-
formational leadership is clearly multidimensional, and there is sup-
port for the structure represented by the MLQ (e.g., Antonakis et al.,
2003). It is important to also note that a recent alternative model for
the elements of transformational leadership has been proposed, but
it too is multidimensional, consisting of ve components (Rafferty &
Grifn, 2004).

Interdependence of the Components

Although each of the components of the FRL model is conceptually
distinct, there are consistent correlations among them. For instance,
charisma (idealized inuence) and inspirational leadership are highly
correlated, but followers may want to emulate the charismatic leader
and not necessarily the inspirational leader. Nevertheless, the same
persons who are charismatic, in general, are likely to be inspirational.
To a lesser extent, all the components of transformational leadership
are likely to correlate with each other and with contingent reward. In
the same way, passive management-by-exception (MBE-P) is likely to
correlate with laissez-faire leadership, but the leader who frequently
displays MBE-P corrects followers; the laissez-faire leader does not.
A conrmatory factor analysis was completed for 14 samples in-
volving 3,786 MLQ descriptions of leaders from a variety of organiza-
tions and agencies that suggested that the factor model that best t the
data was one combining the charisma-inspirational leadership com-
ponents into one factor and also combining the passive management-
by-exception and laissez-faire components into one factor. Yet all the
remaining components of the FRL model could stand alone as separate
factors (Avolio, Bass, & Jung, 1997, 1999). As mentioned, however, the
more recent investigation by Antonakis et al. (2002) provided evidence
MEASUREMENT 25

to support the full nine-factor structure represented by the FRL model.

Moreover, we maintain that conceptually, and for leadership develop-
ment purposes, it may be helpful to use the full nine factors.
In much of the research on transformational leadership, the compo-
nents are combined. For example, the Idealized Inuence and Inspi-
rational Motivation MLQ subscales are frequently combined to create
a composite index of inspirational charisma or charisma, or all four
components are summed to create a total score on transformational
leadership. This is acceptable given the intercorrelations among the
MLQ transformational leadership scales. Although researchers may
want to make general statements about transformational leadership
using the summed score, it is often important to understand how the
individual elements fare. However, because the transactional factors
tend to be more independent of each other, combining them into a
single composite score is a questionable practice.

Correlations With Independent Criteria

Transformational leadership as measured by subordinates ratings was
shown in early research to correlate more highly than transactional
leadership with various independent criteria. For example, 25 project
champions based on interviews in 28 different organizations were
shown to display more transformational behaviors than 25 matched
nonchampions (Howell & Higgins, 1990). Transformational scores
were also higher among innovative school principals, U.S. Marine
Corps commanders of more highly effective helicopter squadrons
(Salter, 1989), Methodist ministers with greater Sunday church at-
tendance and membership growth (Onnen, 1987), presidents of MBA
teams completing complex simulations with greater nancial success
(Avolio, Waldman, & Einstein, 1988), middle business managers with
future nancial success of their business units (Howell & Avolio, 1993),
middle managers with better management committee evaluations
(Hater & Bass, 1988), and junior naval ofcers with recommendations
for early promotion and better tness reports (Yammarino & Bass,
1990a).
Although numerous individual studies, such as those mentioned
earlier, demonstrated relationships between transformational leader-
ship and various indexes of leadership effectiveness, the most convinc-
ing evidence for the predictive validity of transformational leadership
comes from meta-analyses.
26 CHAPTER 2

Meta-Analyses. The hierarchy of correlations in individual stud-

ies that is found in the correlation of the MLQ components with ef-
fectiveness is usually charisma-inspiration, intellectual stimulation,
and individualized consideration > contingent reward > active man-
aging-by-exception > passive managing-by-exception > laissez-faire
leadership. Three early meta-analyses were completed that support
the model: Lowe, Kroeck, and Sivasubramaniam (1996) completed a
meta-analysis of data for 2,873 to 4,242 respondents in public agen-
cies and the private sector and showed that the correlation between
each component of the MLQ generally was consistent with the model.
Thus, as Table 2.3 shows, the mean corrected correlations with effec-
tiveness for the public (including military) and private sectors were,
respectively, charisma-inspiration, .74, .69; intellectual stimulation, .65,
.56; individualized consideration, .63, .62; contingent reward, .41, .41;
and managing-by-exception, .10, -.02.
Gasper (1992) completed another meta-analysis of demonstrated
transformational and transactional leadership. He pooled the trans-
formational components in one composite and the transactional in a
separate composite. For 20 studies, the mean corrected transforma-
tional leadership correlated respectively, .76, .71, and .88 with effec-
tiveness, satisfaction, and extra effort perceived by the followers. The
corresponding corrected correlations with transactional leadership
were .27, .22, and .32. Patterson, Fuller, Kester, and Stringer (1995) cor-
roborated the ndings of these last two meta-analyses for the effects
of transformational and transactional leadership on selected follower
compliance outcomes.

TABLE 2.3
Correlations With Effectiveness
in Public and Private Organizations

Sector

Leadership Public Private

Transformational
Charisma-inspiration .74 .69
Intellectual stimulation .65 .56
Individual consideration .63 .62
Transactional
Contingent reward .41 .41
Managing-by-exception .10 -.02
MEASUREMENT 27

In summary, there is substantial evidence that transformational

leadership, particularly as measured by the MLQ, correlates signi-
cantly with measures of leadership effectiveness. Moreover, there is
value added from transformational leadership over and above the
positive effects of transactional leadership. Yet the relationship between
transformational leadership and leadership outcomes is complex. We
explore the results of the research summarized by these meta-analyses
more fully in chapters 3 and 4, where we individually examine the
inuence of transformational leadership on performance and on other
follower outcomes (e.g., follower commitment, follower satisfaction).
In subsequent chapters, we further explore the complex nature of the
transformational leadershipoutcomes relationship looking at various
mediators and moderators.

OTHER MEASURES
OF TR ANSFOR MATIONAL
LEADERSHIP

The MLQ has been both a boon and a bane to research on transfor-
mational leadership. The popularity of the MLQ among researchers
has helped lead to a near explosion of research on transformational
leadership, and this work has advanced our knowledge of transfor-
mational leadership greatly. Its computerized data collection, scoring,
feedback, and norms make it readily available not only for research
purposes but also for training and development (www.mindgarden
.com). However, the ready availability of the MLQ, coupled with a bit
of a bandwagon effect, may have somewhat stied the development
of other measures of transformational leadership leading to an over-
reliance on pencil-and-paper report methodology. Yet there are other
methods for assessing transformational leadership, some of which
were used in developing both the MLQ and rening transformational
leadership theory.

Diaries
Virginia Military Academy (VMI) cadets reported in unstructured logs
or diaries the leadership behavior they observed during a given set of
days. It was found that these logs could be reliably scored in terms of
all of the various components of the FRL model. These diary reports
28 CHAPTER 2

could then be linked to independently obtained questionnaire mea-

sures (Atwater, Avolio, & Bass, 1991).

Interviews
Interviews with executives about the leadership they saw produced
numerous other behavioral examples of transformational leadership
and were helpful in both constructing the MLQ and in better under-
standing transformational leadership. Charismatic leadership was
attributed to the interviewees bosses for setting examples, showing
determination, displaying extraordinary talents, taking risks, creating
a sense of empowerment in subordinates, showing dedication to the
cause, creating a sense of a joint mission, dealing with crises using
radical solutions, and engendering faith in the subordinates for their
leadership.
Inspirational leadership included providing meaning and chal-
lenge, painting an optimistic future, molding expectations, creating
self-fullling prophesies, and thinking ahead. Intellectual stimulation
was judged present when superiors questioned assumptions, encour-
aged subordinates to employ intuition, entertained ideas that seemed
unusual, created imaginative visions, asked subordinates to rework
the same problems they thought they had solved, and saw unusual
patterns. Individualized consideration was apparent to interviewees
when their bosses answered them with minimum delay, showed they
were concerned for their subordinates well-being, assigned tasks based
on subordinate needs and abilities, encouraged two-way exchanges
of ideas, were available when needed, encouraged self-development,
practiced walk-around management, and effectively mentored, coun-
seled, and coached.
When peers of VMI military cadet leaders were asked what charac-
terized the important traits of a good leader, they tended to describe
traits of inspiration, intellectual stimulation, and individualized
consideration, such as self-condence, persuasiveness, concern for
the well-being of others, and the ability to articulate ones ideas and
thoughts as well as providing models to be emulated by others, hold-
ing high expectations for themselves and others, keeping others well
informed, and maintaining high motivation in themselves (Atwater et
al., 1994).
In a similar vein, the Full Range of Leadership Development Pro-
gram (Avolio & Bass, 1991) begins with participants describing their
MEASUREMENT 29

implicit theories of leadership as evidenced by ideal leaders they have

known. Invariably, for well over 2,000 trainees, the characteristics of
the ideal leader included the components of transformational leader-
ship and contingent reward.

Observational Methods
To date, there have been very few attempts to assess transformational
leadership via systematic, objective, third-party (i.e., nonfollowers) ob-
servations. One exception is a study by Ployhart, Lim, and Chan (2001)
that rated the transformational leadership behaviors of Singaporean
military recruits during a leadership assessment center at the begin-
ning of their basic training. Some of the transformational leadership
behaviors observed and rated included leading by example, inspiring
others, maintaining trusting and cordial relationships with peers, and
demonstrating initiative and courage. These behaviors were mapped
onto the components of the transformational leadership model.
Clearly, thedevelopmentofsystematicobservationalcodingschemes
for transformational leadership could be an important advancement in
measurement. Researchers would then have a more objective indicator
of a leaders transformational leadership behaviors that does not rely
on followers ratings, which can be affected somewhat by the groups
level of performance.

Alternative Pencil-and-Paper Measures

There are a number of measures that have been developed to assess
transformational leadership besides the MLQ (Goodstein & Lanyon,
1999, provide a nice review). The most widely used of these is the
Transformational Leadership Behavior Inventory (TLI) developed by
Podsakoff, MacKenzie, Moorman, and Fetter (1990). This instrument
measures four key dimensions of transformational leadership. The
rst dimension captures the core transformational leadership behav-
ior, which includes developing and articulating a vision, providing a
positive role model, and motivating employees to look beyond their
self-interest for the good of the group. The remaining three dimen-
sions measure the leaders individualized consideration, intellectual
stimulation, and high expectations for performance. The TLI has been
used in various forms in research by Podsakoff and his colleagues and
by others.
30 CHAPTER 2

Warner Burke (1994) used his Leadership Assessment Inventory

(LAI) to measure transformational and transactional leadership for
some time. Unfortunately, this instrument is now difcult to obtain
and rarely used in research.
The Transformational Leadership Questionnaire (TLQ) is a rela-
tively new instrument that measures nine factors associated with
transformational leaders and is specically designed for use in public
sector organizations in the United Kingdom (Alban-Metcalfe & Alimo-
Metcalfe, 2000; Alimo-Metcalfe & Alban-Metcalfe, 2001). The TLQ
has eight scales, labeled Genuine Concern for Others; Decisiveness,
Determination, Self-Condence; Integrity, Trustworthy, Honest and
Open; Empowers, Develops Potential; Inspirational Networker and
Promoter; Accessible, Approachable; Claries Boundaries, Involves
Others in Decisions; Encourages Critical and Strategic Thinking. Like
the MLQ, the TLQ is completed by the managers direct reports. Initial
validation studies for the TLQ have been completed. As can be seen,
there is conceptual similarity between the TLQ and MLQ.
Carless, Wearing, and Mann (2000) developed a short measure of
transformational leadership, the Global Transformational Leadership
scale (GTL). This seven-item scale assesses a single, global construct of
transformational leadership.
Another new alternative measure of transformational leadership
was developed by Rafferty and Grifn (2004), based on leadership
measures created by House (1998) and Podsakoff et al. (1990). This
15-item rating scale measures the transformational leaders vision,
inspirational communication, intellectual stimulation, supportive
leadership, and personal recognition. The authors claim that these
components provide a better factor structure than the MLQ.
Behling and McFillen (1996) created two measures, the Follower Be-
lief Questionnaire and the Attributes of Leader Behavior Questionnaire
that are follower reports of their leaders transformational/charismatic
leadership. Subscales include inspiration, awe, empowerment, dis-
plays empathy, dramatizes mission, projects self-assurance, enhances
image, assures followers of competence, and provides opportunities
to experience success. It is easy to see how these dimensions map onto
the components of transformational leadership. This measure has not
been widely used in research.
In addition to measures of transformational leadership, some re-
searchers have been interested primarily in the charisma elements of
transformational leadership (and many scholars have used the terms
MEASUREMENT 31

charismatic and transformational interchangeably). In assess-

ing charisma, a variety of instruments have been used. Most popu-
lar among these is the Conger-Kanungo scale (CK-scale; Conger &
Kanungo, 1988), which measures transformational characteristics such
as vision and articulation, environmental sensitivity, unconventional
behavior, sensitivity to member needs, taking personal risks, and not
maintaining the status quo.
A number of other leadership measures assess dimensions related
to transformational leadership, although they have not been labeled
explicitly as such. One such measure that is widely used in leader-
ship development programs is Kouzes and Posners (1988) Leadership
Practices Inventory (LPI). Given the popularity of the Kouzes and
Posner model in leadership development programs, this measure is
widely used in practice but more rarely used in published empirical
research. Sashkin (1996) created a measure, the Leadership Behavior
Questionnaire (LBQ), which measures visionary leadership and is dif-
ferent from, but tangentially related to, transformational leadership.

CONCLUSIONS

The measurement of transformational leadership has been dominated

by the MLQ, although there have been alternative measures devel-
oped to assess both transformational and charismatic leadership. As
research on transformational leadership continues, it will be advanta-
geous to use multiple methods for assessing the construct.