Chapter 22: Transformational School Leadership

KENNETH LEITHWOOD, DIANA TOMLINSON AND MAXINE

GENGE
Ontario Institute for Studies in Education

Out of the varying motives of persons, out of the combat and

competition between groups and between persons, out of the
making of countless choices and the sharpening and steeling
of purpose, arise the elevating forces of leadership and the
achievement of intended change (Bums, 1978, p. 432).

Transformational leadership is a term which has appeared with increas-

ing frequency in writings about education since the late 1980's. Some-
times it has been used to signify an appropriate type of leadership for
schools taking up the challenges of restructuring now well underway in
most developed countries throughout the world (Leithwood, 1992). In
this context, a common-sense, non-technical meaning of the term is
often assumed. For example, the dictionary definition of transform is
'to change completely or essentially in composition or structure'
(Webster, 1971). So any leadership with this effect may be labelled
transformational, no matter the specific practices it entails or even
whether the changes wrought are desirable.
This chapter is not concerned with transformational leadership
defined in this loose, common-sense fashion. It is concerned, rather,
with a form of leadership by the same name that has been the subject of
formal definition and systematic inquiry in non-school organizations
for at least several decades. The small but rapidly growing body of evi-
dence, which has emerged quite recently, inquiring about such leader-
ship in elementary and secondary school settings is reviewed in this
chapter. Much of this research takes the non-schoo I literature on trans-
formational leadership as a point of departure, both conceptually and
methodologically. So it is important, at the outset, to appreciate the
general nature of that literature.
Downton's (1973) study of rebel leadership is often cited as the
beginning of systematic inquiry about transformational leadership in
non-school organizations. However, charisma, often considered an
integral part of transformational leadership, has substantially more dis-

785
K. Leithwood et of. (eds.), International Handbook of Educational Leadership and Administration, 785-840.
i;;. 1996 Kluwer Academic Publishers,
786

tant origins - typically attributed to Max Weber's (e.g., 1947) efforts

almost five decades ago.
James McGregor Burns's (1978) prize-winning book first drew
widespread attention to ideas explicitly associated with transforma-
tional leadership. Based on a sweeping historical analysis, Burns
argued that most understandings of leadership not only overempha-
sized the role of power but held a faulty view of power, as well. There
were, he claimed, two essential aspects of power - motives or purposes
and resources - each possessed not only by those exercising leadership
but also by those experiencing it. The essence of leadership is to be
found in the relationships: between motives, resources, leaders, and
followers: ' ... the most powerful influences consist of deeply human
relationships in which two or more persons engage with one another'
(p. 11). Burns's distinction between transactional and transformational
types of leadership hinges on this appreciation of power-as-relation-
ships.
In contrast to transformational leadership, transactional leadership
occurs when one person takes initiative in making contact with others
for the purpose of exchanging valued things (economic, political, or
psychological 'things', for example). Each person in the exchange
understands that she or he brings related motives to the bargaining
process and that these motives can be advanced by maintaining that
process. But because of the nature of the motives at issue, those
involved are not bound together in any continuing, mutual pursuit of
higher purposes. With this form of leadership, motives or purposes
may well be met using the existing resources of those involved in the
exchange. Neither purposes nor resources are changed, however.
Transformational leadership entails not only a change in the pur-
poses and resources of those involved in the leader-follow relationship,
but an elevation of both - a change 'for the better'. With respect to
motives or purposes: 'transforming leadership ultimately becomes
moral in that it raises the level of human conduct and ethical aspiration
of both leader and led, and thus has a transforming effect on both'
(Burns, 1978, p. 20). This form of leadership, according to Burns's
view, also aims to enhance the resources of both leader and led by rais-
ing their levels of commitment to mutual purposes and by further
developing their capacities for achieving those purposes.
Burns's seminal work provided a solid conceptional footing on
which to build the distinction between transactional and transforma-
tional types of leadership!: it also illustrated the meaning of these
forms of leadership in many different contexts. Not to be found in this
 787

work, however, was a testable model of leadership practices or any

empirical evidence of their effects. The prodigious efforts of Bass and
his associates have been largely in response to these limitations. Bass's
(1985) book Leadership and Performance Beyond Expectations pro-
vided an impressive compendium of survey research evidence about
the effects of one model of transformational leadership. Among the
most important features of this model are the dimensions of leadership
practice it includes and the proposed relationships among these dimen-
sions.
Referred to in more recent publications as the four i's (e.g., Bass and
Avolio, 1993, 1994) Bass and his colleagues define transformational
leadership as including: charisma or idealized influence, inspirational
motivation, intellectual stimulation, and individualized consideration.
In addition to these dimensions of transformational leadership, three
dimensions define the meaning of transactional leadership: contingent
reward, management-by-exception, and a laissez-faire or 'hands off
form of leadership.
Whereas Burns considered transformational and transactional prac-
tices as opposite ends of the leadership continuum (essentially more
and less effective forms of leadership), Bass offers a quite different
conception, a 'two-factor theory' ofleadership: transactional and trans-
formational forms ofleadership, in his view, build on one another (e.g.,
Avolio & Bass, 1988; Waldman, Bass & Yammarino, 1990; Bass &
Avolio, 1993; Howell & Avolio, 1991). Transactional practices foster
ongoing work by attending to the basic needs of organizational mem-
bers. Such practices do little to bring about changes in the organization,
however. For this to occur, members must experience transformational
practices, in addition. Enhanced commitment and the extra effort usu-
ally required for change, it is claimed, are consequences of this experi-
ence.
Transactional practices were the traditional focus of attention for
leadership theorists until the early 1980's. Disillusionment with the
outcomes of that focus, however, gave rise to a number of alternative
approaches, among them transformational leadership. These
approaches have been referred to collectively by Bryman (1992), Sims
and Lorenzi (1992), and others as the 'new leadership paradigm'.
Empirical studies of transformational leadership, reflecting this pessi-
mism with transactional practices, often give them minimum attention.
This is the case with Podsakoff, MacKenzie, Moorman, and Fetter
(1990) for example. While Podsakoff and his associates adopted a
quite limited conception of transactional leadership for their research,
788

they offered arguably the most comprehensive set of transformational

leadership dimensions available to that point, dimensions based on a
synthesis of seven prior perspectives on transformational leadership.
These dimensions, which helped organize parts of the subsequent
review, include: identifying and articulating a vision; fostering the
acceptance of group goals; providing an appropriate model; high per-
formance expectations; providing individual support; providing intel-
lectual stimulation; contingent reward; and management-by-exception.
This brief and selective introduction to the formal study of transfor-
mational leadership in non-school organizations is intended to assist
readers in better appreciating the initial perspectives adopted by many
of those whose research in elementary and secondary school settings is
reviewed. Subsequent sections of this chapter describe: the framework
and methods used for this review; the nature of transformational lead-
ership as it is experienced in schools; what is known about the effects
of such leadership; and the antecedents of (or influences giving rise to)
transformational school leadership.

FRAMEWORK AND METHODS FOR THE REVIEW

Framework

Figure 1 is a comprehensive framework for understanding leadership

and an indication of the particular focus for this review. Relationships
among the constructs in the framework are conceptualized as forming a
causal chain with Leadership Practices in the centre of the chain.
These practices are the more or less overt behaviours engaged in by
leaders and, moving backwards in the chain, are a direct product of
leaders' Internal Processes: the personality traits, demographic charac-
teristics, and capacities, skills and thought processes which figure into
leaders' choices of overt behaviour. While internal processes are, in
part, autonomous (a product of innate traits as well as personal experi-
ences), they are shaped also by many kinds of External Influences, the
far left construct in Figure 1. Formal training, informal socialization
experiences, district policies, staff preferences, the weather, commu-
nity opinion and a host of other factors have the potential for such
influence. Neither set of antecedents to leadership are considered in
this review. Leadership Practices, according to Figure 1, potentially
contribute to both organizational outcomes and outcomes which
schools aspire to for students (the far right construct). There is nothing
789

Figure 1:
A Framework for Guiding The Review of Research
on Transformational School leadership

Antecedents to Leadership

---. Internal ---. ---. ---.

Leadership Effects
Leadership
Practices
External Organizational Student
lllfluences Processes Outcomes Outcomes

especially unique about this framework. Variants on it have been pro-

posed, for example, by Bossert et al. (1982) and YukI (1989).

Methods

Our intention was to conduct an exhaustive review of both published

and unpublished research on transformational leadership in elementary
and secondary school organizations up to approximately August 1993.
In the early stages of the search, we also located about two dozen
empirical studies carried out in non-school settings which were read as
background to this review. Electronic searches were made of ERIC, a
comparable Ontario data base called ELOISE, Sociological Abstracts,
Psychological Abstracts, and Dissertation Abstracts. The reference
lists of all studies located through these sources were read and manual
searches conducted for all promising study titles. Dissertation
Abstracts yielded by far the largest proportion of studies finally
selected for the review, an indication of the recency of attention
devoted to inquiry about transformational leadership in schools.
To the studies identified in this way, we added a half dozen studies
reported by others after the completion of the searches carried out, by
this time a year earlier. A comparable number of studies completed by
the first author and his colleagues were also available by this point.
This three-staged search resulted in a final set of 34 empirical and
formal case studies (see Table 1) conducted in elementary and second-
ary school organizations. Of these 34 studies, 12 were conducted using
qUalitative methods, 17 relied on quantitative methods alone, and 5
studies employed some mixture of qualitative and quantitative tech-
niques. Information in the 34 studies were derived from a single source
790

(e.g., a sample of teachers) in 17 of the studies; sample sizes ranged

from 1 (single case studies) to 770. Surveys, interviews, document
analyses and observations were among the instruments or procedures
used for collecting information. Sixteen studies relied on survey instru-
ments alone, 6 on interviews alone, and 11 studies employed multiple
data collection procedures. In one study (Kirby, King and Paradise,
1992) data were collected through a content analysis of the narrative
writings of those who were able to identify exceptional leaders with
whom they had worked.
Finally, the 34 studies of transformational leadership were largely
concerned with the leadership of school principals (22 studies), but
described transformational leadership offered by those in a number of
other educational leadership roles, as well. These other roles included
superintendents and other central office staff (5), some combination of
school and district roles (4), and multiple roles across schools and dis-
tricts (2). One study examined multiple roles, not only in schools but in
other organizations, as well.
Taken as a whole, the methodological features of these studies avoid
some of the most critical threats to the confidence an aggregated body
of evidence permits. The studies are not distributed across so many
roles as to provide little evidence about any single role: there is clearly
much evidence about principals, in particular. Nor are the methods
used exclusively of one type; a surprising number of studies used qual-
itative or mixed designs. And while many studies use only survey
instruments, a reasonable variety of data collection procedures were
used in the remainder. These qualities of the aggregate body of
research reviewed in this chapter provide initial optimism about the
robustness of conclusions that might be drawn.

THE NATURE OF TRANSFORMATIONAL SCHOOL LEADERSHIP

This section identifies specific dimensions of transformational leader-

ship found to be relevant in school contexts and describes the specific
school leadership practices or behaviours associated with each dimen-
sion.
 InstrumentslData Col-
Authors Design Sample Organizational Role
lection Methods

Bass (1985) Qualitative 23 central administrators - survey (MLQ) central office

Brown (1994) Qualitative - 7 principals - interviews - principal

- 18 VP's - vice principal
- 21 dept. heads - dept. head
- teacher

Bright (1987) Quantitative - 78 board members - surveys - superintendent

- 187 principals -MLQ
- 30 superintendents - other

Buck (1989) Qualitative - 51 superintendents - interview - superintendents

Darling (1990) Qualitative - 174 teachers - Survey (School Work - principals

Culture Profile)

- 53 principals - Survey (The Leader-

ship Report)
- Survey (The Principal
Report)

- 10 principals - interview
I

-...I
Table 1: Methodological Features of Studies of Transformational School Leadership \0
-
 ------

-.l
'D
Instruments/Data Col- IV
Authors Design Sample Organizational Role
lection Methods

Genge (1993) Qualitative 10 secondary principals - interviews - principal

Helm (1989) Mixed - staff in 29 schools - Survey (MLQ) - principal

- 4 schools - observation
- unspecified # - interviews
- teachers - school artifacts
- parents - documents
I
- others
- 4 elementary principals

Hoover (1987) Quantitative - 45 schools - Survey (MLQ) - principals

- 151 staff and faculty
members
I
Kendrick (1988) Qualitative - 155 superintendents - Surveys - superintendent
- 670 'followers' I
King (1989) Quantitative - 208 (variety ofroles) - Survey (MLQ) - principals
- Survey (Organizational - superintendents
Climate Index) - central office adminis-
trators
- university administra-
tors
-- -----

Table 1 (contd.): Methodological Features of Studies of Transformational School Leadership

 InstrumentslData Col-
Authors Design Sample Organizational Role
lection Methods

Kirby, King & Paradise Quantitative - \03 educators (teach- - Survey (MLQ) - principals
(1992) Study A ers, VP's, principals) - superintendents
- other central office
roles

StudyB Qualitative - S8 educators (teach- - Narrative writing - multiple K-12 roles

ers, administration)

Koh (1990) Quantitative - 814 teachers - Survey (MLQ) - principals

- 846 teachers - Survey (Organizational
- 89 principals Citizenship Behaviour)
- Survey (Organizational
Commitment)
- Survey (Index of
Organizational Reac-
tions)
- Survey (Index of Per-
ceived Org. Effect)

Kushner (J 982) Quantitative - 61 elementary princi- - Survey (Leadership - principal

pals Posture Questionnaire)
- 6S6 elementary teach-
ers
-- --- -

Table 1 (contd.): Methodological Features of Studies of Transformational School Leadership -..I

\0
v.>
 -.J
'.D
InstrumentslData Col- ~
Authors Design Sample Organizational Role
lection Methods

Lehr (1987) Qualitative - 31 persons, multiple - interviews - leaders of schools and

roles in and outside other types of organi-
education zations

Leithwood, Cousins & Quantitative - 350 secondary teachers - survey - principal

Gerin-Lajoie (1993) - 3557 secondary - survey
schools

Leithwood, Dart, Jantzi & Mixed - 509 intermediate - surveys - principal

Steinbach (1993a) teachers (168 schools)
- 48 intermediate teach- - interviews
ers
- 5 principals (5 schools)

Leithwood, Dart, Jantzi & Mixed - 74 secondary teachers - interviews - principal

Steinbach (1993 b) (6 schools)
- 534 secondary teachers - survey
(77 schools)
- 2045 secondary stu- - survey
dents

Leithwood & Jantzi Qualitative - 133 teachers (12 - interviews - principal

(1990) schools)
_._-- - -----

Table 1 (contd.): Methodological Features of Studies of Transformational School Leadership

 I
InstrumentslData CoI-
Authors Design Sample Organizational Role
lection Methods

Leithwood, Dart, Jantzi & Mixed - 44 teachers (12 ele- - interviews - principal
Steinbach (1991) mentary schools)
- 247 teachers (35 ele- - survey
mentary schools)
- 31 principals - survey

Leithwood, Jantzi & Fern- Quantitative - 168 secondary teachers - survey - principal
andez (1994) (9 schools)

Leithwood, Jantzi, Silins Quantitative - 770 elementary teach- - survey - principal

& Dart (1993) ers (272 schools)

Leithwood & Steinbach Qualitative (expert vs. - 9 elementary principals - stimulated recall - principal
(1991) typical)

Leithwood & Steinbach Mixed - 9 secondary principals - stimulated recall - principal

(1993)

Orr (1990) Quantitative - 155 supts. - surveys - superintendent

- 670 'followers'
- -- --- - - - - - _... _ -

Table 1 (contd.): Methodological Features of Studies of Transformational School Leadership

--l
\0
VI
 ~
Instruments/Data Col- 0\
Authors Design Sample Organizational Role
lection Methods

Roberts (1985) Qualitative, longitudinal - I superintendent - archival searches - superintendent

- participant observation
- 45 staff (multiple - interviews
roles)

Sashkin & Sashkin (1990) Mixed - 12 principals - Survey (Leadership - school and district
- ?VP's Behaviour Question- leadership roles
- ? lead teachers and oth- naire)
ers - Survey (School Cul-
- 16 central office ture Assessment Ques-
- 12 schools tionnaire)

Silins (1992) Quantitative - 678 elementary teach- - survey - principal

ers (256 schools)

Silins (1994a) Quantitative - 679 elementary teach- - survey - principal

ers

Silins (1994b) Quantitative - 458 primary teachers - survey - principal

(23 schools)

Silins & Leithwood Quantitative - 350 secondary teachers - survey - principal

(1994)

Table 1 (contd.): Methodological Features of Studies of Transformational School Leadership

 InstrumentslData Col-
Authors Design Sample Organizational Role
lection Methods

Skalbeck (1991) Qualitative (interpretive - 1 principal - interview and observa- - principal

case study) tion
- 15 teachers - interviews
- 2 VP's - interviews
- 5 classified staff - interviews
- 2 district administra- - interviews
tors
- 5 others - interviews
- 10 parents, students, - interviews
community members

Smith (1989) Quantitative - 69 principals - Survey (MLQ) - principals

- 88 superintendents - Survey (Educational - superintendents
Work Components
Study)

Vandenburghe & Staes- Qualitative: case study - 1 school - questionnaire - principal

sens (1991) - interviews
i
- observation
- documents
I

Table 1 (contd.): Methodological Features of Studies of Transformational School Leadership

-..J
'-0
-..J
798

Dimensions of Transformational Leadership Relevant to School Settings

Twenty-one studies provided evidence about specific dimensions of

transformational leadership relevant to school contexts: these included
six qualitative and 15 quantitative studies. For each of these studies,
the leadership dimensions explicitly referred to by the author(s) were
identified and the effects reported. Table 2 summarizes the results of
this analysis. The left column of this table indicates that a total of 12
dimensions of transformational leadership were inquired about in the
21 studies. All but two of these dimensions can be found in the
research literature on transformational leadership in non-school set-
tings. Culture building and Structuring are unique to school-based
research: while these dimensions are found primarily in a series of
studies carried out by Leithwood and his colleagues (Leithwood &
Jantzi, 1990; Silins, 1994b; Leithwood & Steinbach, 1991), their rele-
vance to leadership in schools is also reflected in the work of others.
Culture bUilding is prominent in studies by Sashkin and Sashkin
(1990), Skalbeck (1991), Helm (1989). Vandenburghe and Staessens
(1991), Darling (1990), and Hendrick (1988). Structuring, including
the involvement of staff in decision making, for example, is identified
as an important leadership practice by Helm (1989) and Roberts
(1985).
Information in the second and third columns of Table 2 help assess
the extent to which each leadership dimension has been found to be rel-
evant to school settings. Using information in these columns, judge-
ments of relevance depend on: the total number of studies providing
data about the dimension (while there is no non-arbitrary way of estab-
lishing a minimum number, more studies add certainty to the judge-
ment); the average number of significant positive relationships
reported across studies between a leadership dimension and an out-
come or dependent measure (many studies inchlde multiple dependent
measures); and the number of non-significant and significant negative
relationships reported.
Clearly relevant to schools, given the evidence in Table 2, are lead-
ership dimensions that have been the object of a relatively large
number of studies, and in which an overwhelming proportion of signif-
icant positive relationships have been reported. Meeting this standard
are: transformational: composite (C), charisma/inspiration/vision,
intellectual stimulation, and individual consideration. Management-
by-exception, the object of many studies reporting a high proportion of
significant negative relationships is the only leadership dimension
 # of Studies # of Positive # of N.S. Out- # of Neg. Out-
Leadership Dimension
(Quant.lQual.) Outcomes comes comes
Total M
I. Transformational composite: 10 (8/2) 18(1.8) 1 0
2. Transactional composite: 8 (8/0) 7 (.88) 11 0
3. CharismaiInspirationlVision 14(2/12) 29 (2.07) 2 0
4. Goal Consensus 5 (411) 9 (1.5) 0 0
5. Individual Consideration 14 (12/2) 19(1.36) I 0
6. Intellectual Stimulation 14(13/0) 20 (1.42) 2 0
7. Modelling 4 (3/1) 3 (0.6) 4 0
8. High Performance Expectations 4 (4/0) 4 (0.8) 4 0
9. Culture Building 4 (1/3) 5 (1.25) I 0
10. Structuring 2 (0/2) 2 (1.0) 0 0
11. Management-by-Exception (Active/ 9 (9/0) 1 (.1) 11 3
PassivelLaissez-Faire)
12. Contingent Reward 10 (10/0) 9 (.9) 5 2

Table 2: Dimensions of Transformational School Leadership Relevant to School Settings

(Total Number of Studies = 21)

;:cl
\0
800

clearly not relevant to schools. Of the remaining dimensions contingent

reward and transactional: composite (C) have been relatively well-
studied but results are conflicting, possibly suggesting that their effects
are more contingent than are the effects of other dimensions. Structur-
ing and culture building have been too little studied for a judgement of
relevance to be made; the sparse results are promising, however. High
performance expectations and modelling show signs of behaving like
contingent reward, but the amount of evidence is sparse. And finally,
goal consensus shows signs of being clearly relevant to schools, but
more evidence is still needed to make that judgement with reasonable
certainty.

The Theory and Practice of Individual Transformational School Leadership

Dimensions

As Table 2 indicates, nine dimensions of transformational leadership

are either clearly relevant to school settings, show promise of being rel-
evant, or cannot yet be ruled out because of limited or ambiguous
results. Excluded are transformational C and transactional C because
they are a combination of other more specific dimensions; manage-
ment-by-exception is ruled out on the grounds described above. For
each of the nine dimensions, this section offers a definition and
explains its contribution to the motivation of followers. Links with
other relevant leadership theory are also made.
Based on a content analysis of findings of 21 studies, the specific
behaviours or practices associated with each dimension are outlined.
For quantitative studies, each item included in the survey scale measur-
ing a leadership dimension was treated as a potential leadership behav-
iour or practice. Six survey instruments generated behaviours defined
in this way, including: Bass's Multifactor Leadership Questionnaire
(several different forms); Leithwood et al. 's Nature of School Leader-
ship Survey, and a significant adaptation by Silins (l994b); Sashkin's
(1985) Leader Behaviour Questionnaire; The Leadership Report (Dar-
ling, 1990); and the Leadership Posture Questionnaire (Kushner,
1982).
 801

CharismalInspirationNision

Although treated as a single leadership dimension for this review, char-

ismatic or inspirational practices and practices giving rise to shared
school vision are distinct. Or at least they may be distinct, this uncer-
tainty being a function ofthe data available. Evidence concerning cha-
risma largely describes attributions made by 'followers' about the
qualities of those believed to be charismatic. In contrast, evidence
about vision building describes specific behaviours engaged in by lead-
ers with their colleagues.
There is an extensively developed literature on charismatic leader-
ship, in its own right, which substantially overlaps the literature on
transformational leadership. For purposes of this chapter, transforma-
tionalleadership is not considered to be synonymous with charisma but
charisma is an important part of such leadership. Furthermore, efforts
to inquire about charismatic leadership predate research about transfor-
mational concepts by many decades. For example, Weber's (1947)
early work usefully distinguished alternative forms of power on which
authority might be based. These include legal, traditional and personal
forms, the latter being the form undergirding charismatic leadership.
Personal power, Weber claimed, grows out of perceptions that leaders
possess valuable expertise, as well as other unique attributes and char-
acteristics. The term perceived is crucial to the concept of charisma.
Most contemporary views stress the attributions of followers concern-
ing their leaders (e.g., Conger & Kanungo, 1987). If followers do not
'feel' leadership, there isn't any. Such perceptions arise in the context
of particular kinds of social relationships ' ... which, by virtue of both
the extraordinary qualities that followers attribute to the leader and the
latter's mission, the charismatic leader is regarded by his or her follow-
ers with a mixture of reverence, unflinching dedication and awe' (Bry-
man, 1992, p. 41).
Charismatic attributions are a consequence of what leaders do as
well as the circumstances in which followers find themselves. Reflect-
ing these different sources, Boal and Bryson (1988) argue that there are
actually two types of charismatic leaders, 'visionary' and 'crisis-pro-
duced'. The power of visionary, charismatic leaders is to be found in
the attractiveness of the missions which they espouse and the willing-
ness of others to believe in those missions. From the perspective of
motivational theory, then, visionary charismatics influence the nature
of the personal goals motivating the behaviour of followers. Followers
802

will aspire to more ambitious or perhaps even morally more defensible

goals than would be the case in the absence of visionary charisma.
Crisis-produced charismatic leaders, on the other hand, are products
of a set of circumstances which potential followers feel unable to cope
with. Charisma is attributed to persons who are perceived to offer a
way of at least beginning to deal with those circumstances. From a
motivational perspective, crisis-produced charismatic leadership
enhances followers' context beliefs; it increases followers' estimates of
the likelihood of support for their change efforts through the actions of
those awarded leadership status. Such attributions of charisma continue
as long as those perceptions continue. Leadership attributions are not
bi-polar, however. Charisma may vary in intensity and may be dis-
persed beyond single individuals to groups and even to whole organi-
zations.
The experience of charisma, according to the 13 studies 3 of school
leadership in which it was examined, generates increased optimism
among colleagues about the future and generates enthusiasm about
work. Charismatic school leaders are perceived to exercise power in
socially positive ways. They create trust among colleagues in their
ability to overcome any obstacle and are a source of pride to have as
associates. Colleagues consider these leaders to be symbols of success
and accomplishment, and to have unusual insights about what is really
important to attend to; they are highly respected by colleagues.
A leader who assists his or her colleagues in identifying and articu-
lating a vision, whether or not attributed charisma, engages in behav-
iours aimed 'at identifying new opportunities for his or her unit - and
developing, articulating, and inspiring others with his or her vision of
the future' (Podsakoff et aI., 1990, p. 112). Such behaviour is not only
central to theories of charisma, it is a critical part of the explanation for
leaders' effects offered by Nanus (1992). To the extent that vision is
required for planning, its roots can be found.in classical theories of
management developed by Barnard, Krech, and Crutchfield, among
others, at least four decades ago (Bass, 1981).
Those studies which offered evidence about vision-building at the
school level identified 8 associated behaviours:
helping to provide colleagues with an overall sense of purpose;
initiating processes (retreats, etc.) which engage staff in the col-
lective development of a shared vision;
espousing a vision for the school but not in a way that pre-empts
others from expressing their vision;
 8fr3

exciting colleagues with visions of what they may be able to ac-

complish if they work together to change their practices;
helping clarify the meaning of the school's vision in terms of its
practical implications for programs and instruction;
assisting staff in understanding the relationship between external
initiatives for change and the school's vision;
assisting staff in understanding the larger social mission of which
their vision of the school is a part, a social mission which may in-
clude such important end values as eqaality, justice and integrity;
using all available opportunities to communicate the school's vi-
sion to staff, students, parents, and other members of the school
community.

Vision-building behaviours identified primarily in studies of superin-

tendents include:
developing a district mission statement and constantly using it
with staff in communication and decision making;
creating a shared vision for the district in which most district
members share;
using research in decision making and planning;
being sensitive to the views of the community, parents, board and
staff about directions for the district;
willing to take risks in order to bring about change;
incorporating considerations of the district's past and present in
developing plans for the future.

Goal Consensus

Vision building (the dimension of leadership discussed above) and the

development of consensus about goals are closely related sets of lead-
ership practices. The conceptual difference lies in the time frame and
the scope of concern of the direction - setting activities that both sets of
practices entail. Vision building is intended to create a fundamental,
ambitious sense of purpose, one to be pursued over (likely) many
years. Developing a consensus on goals focuses organizational mem-
bers on what will need to be accomplished in the short run (e.g., this
year), in order to move toward the vision.
This dimension of transformational leadership, then, includes
behaviour 'aimed at promoting cooperation among employees and get-
ting them to work together toward a common goal' (Podsakoff et aI.,
804

1990, p. 112). Goal-setting activities fostered by the leader are motiva-

tional to the extent that they increase goal clarity and the perception of
goals as challenging but achievable. The promotion of cooperative
goals may positively influence teachers' context beliefs, as well.
Goal setting also is part of an approach to leadership, pursued vigor-
ously in the 1950's and 60's, that focused on style or behaviour. This
approach awarded substantial importance to the leader's 'initiation of
structure' as a means to help define followers' tasks (e.g., Halpin,
1957; YukI, 1989). Goal setting practices were pivotal to the initiation
of structure, according to this approach. Furthermore the determination
of objectives and the maintenance of goal direction were among the
central functions of leadership identified even earlier by classical man-
agement and behavioural theorists (Bass, 1981). Contemporary social
cognitive theories of leadership continue to ascribe substantial impor-
tance to goals and to goal setting (Sims & Lorenzi, 1992), awarding
special attention to questions about how goals are set and what is the
effect of variation in such processes.
Identified by the 5 studies4 concerned with this dimension of leader-
ship in schools were 10 specific behaviours:
providing staff with a process through which to establish school
goals and to regularly review those goals; this is likely to be a
'problem-solving' process and to include careful diagnosis of the
school's context;
expecting teams of teachers (e.g., departments) and individuals
to regularly engage in goal setting and reviewing progress to-
ward those goals;
assisting staff in developing consistency between school visions
and both group and individual goals;
working toward the development of consensus about school and
group goals and the priority to be awarded such goals;
frequently referring to school goals and making explicit use of
them when decisions are being made about changes in the
school;
encouraging teachers, as part of goal-setting, to establish and re-
view individual professional growth goals;
having ongoing discussion with individual teachers about their
professional growth goals;
clearly acknowledging the compatibility of teachers' and
school's goals when such is the case;
expressing one's own views about school goals and priorities;
 805

acting as an important resource in helping colleagues achieve

their individual and school goals.

Three additional behaviours were identified in studies of superintend-

ents:
influencing others to accept district goals;
focusing on intra-system development;
empowering others through goal setting initiatives.

High Performance Expectations

This dimension of transformational school leadership consists of

behaviour that 'demonstrates the leader's expectations for excellence,
quality, and/or high performance on the part of followers' (Podsakoff
et aI., 1990, p. 112). Expectations of this sort by school leaders will be
motivational as they help teachers see the challenging nature of the
goals being pursued in their school. Such expectations may also
sharpen teachers' perceptions of the gap between what the school
aspires to and what is presently being accomplished. Done well,
expressions of high expectations should also result in perceptions
among teachers that what is being expected is also feasible.
Behaviour which demonstrates high performance expectations is
also consistent with what House and his colleagues termed 'achieve-
ment-oriented leadership' in their path-goal theory. According to this
theory, the effects of such expectations are contingent on both selected
personal characteristics of followers and the environment in which
they are working (e.g., House & Mitchell, 1974).
Four studies 5 in school settings inquiring about this dimension of
leadership identified a total of six specific behaviours:
expecting staff to be innovative, hard working and professional;
these qualities are included among the criteria used in hiring
staff;
demonstrating an unflagging commitment to the welfare of stu-
dents;
often espousing norms of excellence and quality of service;
not accepting second-rate performance from anyone;
establishing flexible boundaries for what people do, thus permit-
ting freedom of judgement and action within the context of over-
all school goals and plans;
being clear about one's own views of what is right and good.
806

Studies of superintendents identified two additional practices:

openly valuing justice, community, democracy, excellence and
equality;
expressing commitment to affective educational goals for stu-
dents.

Individual Consideration

This dimension of transformational leadership encompasses behaviour

indicating that the leader respects followers and is concerned about
their personal feelings and needs (Podsakoff et aI., 1990, p. 112). Such
behaviour may be motivational through its influence on context beliefs,
assuring teachers that the problems they are likely to encounter while
changing their practices will be taken seriously by those in leadership
roles and efforts will be made to help them through those problems.
This dimension of leadership closely parallels the central role of
consideration in the 'style' approach to leadership: 'the extent to which
leaders promote camaraderie, mutual trust, liking, and respect in the
relationship between themselves and their subordinates' (Bryman,
1992, p. 5).
Fourteen studies 6 were sources of information for the twenty spe-
cific behaviours associated with this leadership dimension. These spe-
cific behaviours suggested several distinct facets of individualized
support. One facet, involving the equitable, humane and considerate
treatment of one's colleagues, included:
treating everyone equally; not showing favouritism towards indi-
viduals or groups;
having an 'open door' policy;
being approachable, accessible and welcoming;
protecting teachers from excessive intrusions on their classroom
work;
giving personal attention to colleagues who seem neglected by
others;
being thoughtful about the personal needs of staff.

A second facet of individual consideration is the provision of support

for the personal, professional development of staff. This includes:
encouraging individual staff members to try new practices con-
sistent with their interests;
 807

as often as possible, responding positively to staff members' in-

itiatives for change;
as often as possible, providing money for professional develop-
ment and other needed resources in support of changes agreed on
by staff;
providing coaching for those staff members who need it.

Individual consideration requires leaders to develop close knowledge

of their individual colleagues. This is done by:
getting to know individual teachers well enough to understand
their problems and to be aware of their particular skills and inter-
ests; listening carefully to staffs ideas;
having the 'pulse' ofthe school and building on the individual in-
terests of teachers, often as the starting point for school change.

Consideration is also expressed through recognition of good work and

effort. Transformational leaders:
provide recognition for staff work in the form of individual
praise or 'pats on the back';
are specific about what is being praised as 'good work';
offer personal encouragement to individuals for good perform-
ance;
demonstrate confidence in colleagues' ability to perform at their
best.

Finally, individual consideration is reflected in transformational lead-

ers' approaches to change. These leaders:
follow through on decisions made jointly with teachers;
explicitly share teachers' legitimate cautions about proceeding
quickly toward implementing new practices, thus demonstrating
sensitivity to the real problems of implementation faced by
teachers;
take individual teachers' opinion into consideration when initiat-
ing actions that may effect their work;
instill, in staff, a sense of belonging to the school.

Two studies of superintendents' individual support added nothing

beyond those behaviours already described (Buck, 1989; Smith, 1989).
808

Intellectual Stimulation

Behaviour that 'challenges followers to re-examine some of their work

and to rethink how it can be performed' (Podsakoff et aI., 1990, p. 112)
is the meaning of intellectual stimulation. This may consist of a type of
feedback associated with verbal persuasion. Such stimulation also
seems likely to draw teachers' attention to discrepancies between cur-
rent and desired practices and to understand the truly challenging
nature of school restructuring goals. To the extent that such stimulation
creates perceptions of a dynamic and changing job for teachers, it
should enhance emotional arousal processes, also.
Much of what organizational learning theorists have to say about
leadership could be used to expand on the meaning of this dimension
of leadership. Senge (1990), for example, argues that leaders' new
work will require them 'to bring to the surface and challenge prevailing
mental models, and to foster more systemic patterns of thinking - lead-
ers are responsible for learning' (p. 9). Argyris and Schanted assump-
tions about their work, to engage in 'double loop' learning. To the
extent that leaders' own intellectual resources have a bearing on their
abilities to provide such intellectual stimulation, even cognitive
resource theory (e.g., Fiedler & Garcia, 1987) may be helpful in better
understanding the nature and effects of those practices included as part
of this leadership dimension.
Transformational leaders, in the 14 studies7 concerned with intellec-
tual stimulation, used four basic strategies. One strategy was to change
school norms that might constrain the thinking of staff. This was
accomplished by:
removing penalties for making mistakes as part of efforts toward
professional and school improvement;
embracing and sometimes generating conflict as a way of clari-
fYing alternative courses of action available to the school;
requiring colleagues to support opinions with good reasons;
insisting on careful thought before action.

A second strategy used by transformational leaders was to challenge

the status quo by:
directly challenging staffs' basic assumptions about their work as
well as unsubstantiated or questionable beliefs and practices;
encouraging staff to evaluate their practices and refine them as
needed;
 809

encouraging colleagues to re-examine some of their basic as-

sumptions about their work; making problematic the way things
are;
stimulating colleagues to think more deeply about what they are
doing for their students.

Encouraging new initiatives, a third strategy, entailed such behaviours

as:
encouraging staff to try new practices without using pressure;
encouraging staff to pursue their own goals for professional
learning;
helping staff to make personal meaning of change;
providing the necessary resources to support staff participation in
change initiatives.

A fourth strategy used by transformational leaders for purposes of

intellectual stimulation was to bring their colleagues into contact with
new ideas. This was done by:
stimulating the search for and discussion of new ideas and infor-
mation relevant to school directions;
seeking out new ideas by visiting other schools, attending confer-
ences, and passing on these new ideas to staff;
inviting teachers to share their expertise with their colleagues;
consistently seeking out and communicating productive activi-
ties taking place within the school;
providing information helpful to staff in thinking of ways to im-
plement new practices.

Modelling

This dimension of transformational leadership encompasses behaviour

on the part of the leader 'that sets an example for employees to follow
that is consistent with the values the leader espouses' (Podsakoff et aI.,
1990, p. 112). Such behaviour may enhance teachers' beliefs about
their own capacities, their sense of self-efficacy. Secondarily, such
modelling may contribute to emotional arousal processes by creating
perceptions of a dynamic and changing job.
Earlier trait theories of leadership (Bryman, 1992; Hunt et aI., 1988),
identified a handful of relatively robust personal characteristics of
leaders, the impact of which can be explained, in part, through their
810

modelling effects: for example, energy, honesty, integrity, self-confi-

dence, initiative, and persistence.
Four studies 8 provided information about specific behaviours
included in this dimension, all of which entail the leader acting as a
role model, leading by doing rather than only by telling. Some of these
behaviours model the transformational leader's general commitment to
the school organization:
becoming involved in all aspects of school activity;
working alongside teachers to plan special events;
displaying energy and enthusiasm for own work.

Other behaviours model commitment to professional growth:

responding constructively to unrequested feedback about one's
leadership practices;
requesting feedback from staff about one's work;
demonstrating a willingness to change one's practices in light of
new understandings.

Yet other behaviours seem intended to enhance the quality of both

group and individual problem solving processes:
demonstrating, through school decision-making processes, the
value of examining problems from multiple perspectives;
modelling problem-solving techniques that others can adapt for
their own work.

Finally, transformational school leaders also engage in behaviours

intended to reinforce key values: the basic values of respect for others,
trust in the judgement of one's colleagues, integrity, and even the
instrumental value of punctuality.

Contingent Reward

As defined by Avolio and Bass (1988), contingent reward occurs when

the leader 'is seen as frequently telling subordinates what to do to
achieve a desired reward for their efforts' (p. 35). This leadership
dimension typically is viewed, theoretically, as transactional. But the
possibility of providing informative feedback about performance in
order to enhance teachers' capacity beliefs as well as emotional arousal
processes makes this behaviour potentially transforming, as well.
 811

Contingent reward draws heavily on ideas central to path-goal the-

ory (e.g., House & Mitchell, 1974) which itself draws extensively on
behaviourist approaches to motivation.
Although ten studies9 inquired about this dimension of leadership,
only a small number of specific behaviours were identified. Those
behaviours, largely based on survey items (in particular items from
Bass's (1985) Multifactor Leadership Questionnaire) include:
assuring staff members that they can get what they want person-
ally in exchange for their efforts;
paying personal compliments to staff when they do outstanding
work;
frequently acknowledging good performance;
providing public recognition for good work.

Structuring

This dimension of transformational leadership includes behaviours

aimed at providing opportunities for members of the school organiza-
tion to participate in decision-making about issues which effect them
and to which their knowledge is crucial. Also part of this dimension are
behaviours which create discretion and autonomy for teachers to use
their expertise to greatest effect. Empowering teachers in these ways
contributes to the motivation to change by enhancing teachers' beliefs
about the extent to which their working context will support their best
efforts to implement new practices in their classrooms and schools.
Two studies provided descriptions of structuring behaviours (Helm,
1989; Leithwood & Jantzi, 1990) engaged in by school-level leaders.
These behaviours included:
distributing the responsibility and power for leadership widely
throughout the school;
sharing decision-making power with staff;
allowing staffto manage their own decision-making committees;
taking staff opinion into account when making own decisions;
ensuring effective group problem solving during meetings of
staff;
providing autonomy for teachers (groups, individuals) in their
decisions;
altering working conditions so that staff have collaborative plan-
ning time and time to seek out information needed for planning
and decision making;
812

ensuring adequate involvement in decision-making related to

new initiatives in the school;
creating opportunities for staff development.

A number of structuring behaviours were observed as part of the reper-

toire of the transformational superintendent who was the subject of
Roberts's (1985) case study. This female superintendent:
designed an infrastructure in the district (formal and informal) to
support change initiatives;
hired and replaced staff to reflect changes in the district and the
talent of district employees;
developed joint expectations for the work of district and school
staffs, and allowed considerable discretion in how these expedc-
tations are achieved;
employed participative management.

Culture Building

The culture of a school includes the norms, values, beliefs and assump-
tions shared by members of the school (Shein, 1985). Considerable
evidence suggests that school culture, defined in this way, explains a
large amount of the variation in school effects (e.g., Little, 1982; Nias,
Southworth & Campbell, 1989). This explanation includes not only the
content of the culture (e.g., student-focused norms are associated with
'effective' cultures), but its strength and form as well. The culture
among professionals in schools is typically characterized as weak (lit-
tle consensus) and isolated (Fieman-Nemser & Floden, 1986) whereas
strong, collaborative school cultures contribute more substantially to
school improvement initiatives. Recent research not specifically
framed by concepts of transformational leadership have argued for the
importance of culture building and described relevant leadership prac-
tices (Deal & Peterson, 1990; Cunningham & Gresso, 1993).
Culture-building by transformational leaders includes behaviours
aimed at developing school norms, values, beliefs and assumptions that
are student-centred and support continuing professional growth by
teachers. Such behaviours also encourage collaborative problem solv-
ing when that is likely to be profitable: they have the potential to
enhance teachers' motivation to change through their influence or con-
text beliefs ('we're all in this together'); they may be motivational, as
 813

well, through enhanced self-efficacy resulting from the professional

growth fostered by close working relations with peers.
Four studies provided descriptions of culture building behaviours
(Helm, 1989; Leithwood, Dart, Jantzi & Steinbach, 1993b; Leithwood
& Jantzi, 1990; Skalbeck, 1991) on the part of school leaders. Some of
these behaviours aimed at strengthening the school culture by:
clarifying the school's vision in relation to collaborative work
and the care and respect with which students were to be treated;
reinforcing, with staff, norms of excellence for their own work
and the work of students;
using every opportunity to focus attention on and to publicly
communicate the school's vision and goals;
using symbols and rituals to express cultural values in the context
of social occasions in which most staff participate;
confronting conflict openly and acting to resolve it through the
use of shared values;
using slogans and motivational phrases redundantly;
using bureaucratic mechanisms to support cultural values and a
collaborative form of culture (e.g., hiring staff who share school
vision, norms and values);
assisting staffto clarify shared beliefs and values and to act in ac-
cord with such beliefs and values;
acting in a manner consistent with those beliefs and values
shared within the school.

Other behaviours were aimed at the/orm of the school's culture, in par-

ticular the desire for it to be collaborative. These behaviours included:
sharing power and responsibility with others;
working to eliminate 'boundaries' between administrators and
teachers and between other groups in the school;
providing opportunities and resources for collaborative staff
work (e.g., creating projects in which collaboration clearly is a
useful method of working).

Summary

Evidence provided by twenty-one studies warrant continued attention,

in school settings, to seven dimensions of transformational leadership
originally proposed for non-school settings. This continuing attention
is sometimes justified by substantial evidence of positive effects, as in
814

the case of three dimensions: charisma/inspiration/vision, intellectual

stimulation, and individual consideration. In the case of 4 dimensions,
evidence is meagre but promising or ambiguous: contingent reward,
high performance expectations, goal consensus, and modelling, and
further inquiry is needed. One dimension, management-by-exception
(attending to an aspect of the organization only when something excep-
tional or unusual occurs), clearly can be dismissed as having no pro-
ductive contribution to make to school organizations. Two dimensions
unique to school-based research on transformational leadership have
been proposed: these are culture building and structuring. While there
is only meagre direct evidence of their value in the 21 studies, substan-
tial evidence is available outside the framework of transformational
leadership theory.
Using survey items and qualitative descriptions, a synthesis of spe-
cific behaviours associated with each of nine leadership dimensions
was described in this section. While most of these behaviours seem
plausibly a part of the leadership dimensions in which they are classi-
fied, they are at best an eclectic 'starter set'. Considerably more
research is necessary to approximate a comprehensive set of behav-
iours for each leadership dimension. More research also is needed to
clarify the status of each behaviour's contribution to the effects antici-
pated for the leadership dimension of which it is a part. Is the behav-
iour a critical part of the dimension? Can other behaviours be
substituted with no loss of effect? How contingent are these behaviours
on the situation in which transformational leadership is being exer-
cised? The 21 studies have nothing to say about these important ques-
tions.

EFFECTS OF TRANSFORMATIONAL SCHOOL LEADERSHIP

This section of the chapter takes up two matters concerning the effects
of transformational school leadership. First, Bass's claim that transfor-
mational leadership is 'value-added' (Avolio & Bass, 1988, Bass &
Avolio, 1993) is examined in the context of schools (8 studies). Then,
evidence from 20 studies will be reviewed concerning the effects of
transformational school leadership on the full array of outcomes used
as dependent measures in these studies.
 815

Extent of Support for a 'Two-Factor Theory' of Transformational Leadership in

Schools

Only studies in schools which used either Bass's formulation of trans-

formational leadership or modest adaptations of it were considered as
evidence about the validity of the 'two-factor theory' in schools. This
decision was made because the theory is largely attributable to Bass
and his associates: according to their formulation transactional leader-
ship (contingent reward, management-by-exception, and sometimes
laissez-faire leadership) provides the essential requirements for organi-
zational maintenance. Adding on transformational leadership (includ-
ing charisma/inspiration, intellectual stimulation, and individualized
consideration) produces change. These two forms of leadership are
treated as composite, higher- or second-order factors in the empirical
literature.
As a preliminary qualification to the following analysis, the Bass-
related formulations of transactional leadership do not appear to ade-
quately encompass the range of leadership practices legitimately con-
ceptualized as "transactional". Subsequent discussions of the relative
effects of transactional leadership are clearly limited in their value, as a
consequence.
Eight studies were located which directly tested the two-factor the-
ory in school settings. All but one of these studies relied on survey data
and assumed Bass's five or six dimensional view of leadership prac-
tices. lO These studies used data from Singapore (1 study), Australia
(1), Canada (2), and the United States (4). One study (Buck, 1989)
focused on the superintendent's role, another mixed elementary and
secondary and higher education leadership roles (King, 1989). The
remainder were studies of elementary and secondary school principal
roles.
Two types of evidence are relevant to a test of transformational lead-
ership as a 'two-factor' theory. One type of evidence concerns the
external validity of the theory; this is evidence which confirms that
each of the two higher-order factors in the theory has the consequences
claimed for it. A second type of evidence bears on the internal validity
of the theory. This is the claim that particular dimensions of leadership
practice are related in the manner specified by theory: that is, charisma/
inspiration/vision, intellectual stimulation and individual support form
one related cluster of variables and contingent reward and manage-
ment-by-exception form a second, different cluster.
816

Support for the external validity of the two-factor theory is to be

found in evidence which demonstrates that transactional leadership
practices are necessary to the maintenance of organizational routines,
that they provide essential stability but do not, by themselves, stimulate
much change. Only by 'adding on' transformational practices is change
likely. 11 A quantitative test of this claim might use stepwise regression
analyses, for example; measures of transactional leadership are entered
in the regression equation first, in order to estimate how much variance
in the dependent measure(s) is accounted for. When measures of trans-
formational leadership, entered second into the regression equation,
account for significant additional variation in the dependent meas-
ure(s), the external validity of the two-factor theory receives support.
The external validity of the two-factor theory could also be assessed
using qualitative data. For example, case study data concerning the rel-
ative success of school improvement initiatives undertaken by school
principals varying in the degrees to which they provided both transfor-
mational and transactional leadership practices would be relevant to a
test of external validity.
Claims that the two-factor theory is internally valid require evidence
demonstrating
(a) very low correlations among those leadership dimensions con-
ceptually defined as transactional and those defined as transfor-
mational and;
(b) relatively high correlations among those dimensions within each
leadership type. Factor analysis of responses to all items on a
leadership survey is the most frequently used method of quanti-
tatively testing the internal validity claim. Strongest support for
the claim is provided by results in which all items load on only
two factors. Those items considered part of transformational
leadership, conceptually, load on one factor and all items consist-
ent with the theoretical meaning of transactional leadership load
on a second factor.

It is less obvious how this claim could be tested with qualitative evi-
dence. However, teacher talk explicitly linking together, in terms of
subjective meaning and perceived consequences, an array of practices
theoretically considered to be transactional and yet another array con-
sidered to be transformational would assist with such a test.
Silins's (1992) study of Canadian school leaders provides evidence
from 670 teachers bearing on both the internal and external validity of
the two-factor theory. Confirmatory factor analysis strongly support an
 817

internal validity claim, but as Silins explains, ' ... these two [transac-
tional] factors do not conform to Bass's hypothesized relationships'
(1992, p. 326). Instead, contingent reward loaded on the factor which
included transformational practices, leaving management-by-exception
as the sole dimension defining transactional leadership. These data
support the internal validity of a significantly modified theory of trans-
formational leadership.
Evidence from Silins's (1992) study bearing indirectly on the exter-
nal validity of transformational leadership theory was produced
through a canonical correlation analysis of data using four different
outcome measures, each based on teachers' perceptions. Although
somewhat ambiguous for this purpose, results do not support external
validity claims. Management-by-exception, the single dimension of
practice empirically defining transactional leadership, had negative
effects on school change, whereas at least neutral effects would be pre-
dicted by prior theory. Transformational practices, as a whole, had
expected effects but one of the dimensions, charisma/inspiration/
vision, did not behave as predicted.
A second study by Silins (1994b) of Australian school leaders, using
responses from 458 teachers, assessed the internal validity of the two-
factor theory, as well as a more elaborate conception of trans form a-
tionalleadership theory influenced by work in both Australia and Can-
ada. Confirmatory factor analysis provided weak support for the
internal validity of a reformulated two-factor theory, as in the previous
study. Leadership practices conceptualized as transformational all
loaded on one factor along with practices associated with contingent
reward. Items measuring the remaining transactional dimension, man-
agement-by-exception, loaded either negatively on the transformational
factor or positively on a second factor.
In a third study based on responses from 679 teachers, again focused
on Canadian school leaders, Silins (l994a) assessed both the internal
and external validity of the two-factor theory using both canonical
analysis and partial least squares path analysis with four different out-
come measures. The path model resulting from canonical analysis sup-
ported the internal validity claims of a modified two-factor theory. In
this study, charisma/inspiration/vision helped define, negatively, trans-
actional leadership, whereas contingent reward contributed to the
meaning of both transformational and transactional factors. The results
provide strong direct support for external validity claims. Transforma-
tional leadership explained 55 percent of the variance in transactional
leadership.
818

Results of Koh's (1990) study of Singapore principals do not sup-

port the internal validity of the two-factor theory. Exploratory factor
analysis resulted in a unitary factor defined by three transfonnational
leadership dimensions. However, transactional leadership practices
loaded on three separate factors, factors best interpreted as contingent
reward, passive management-by-exception and active management-by-
exception. Had contingent reward loaded on the transfonnational fac-
tor, these results would not have been substantially different than those
reported in the three Silins studies. This gives rise to the possibility that
contingent reward can take several fonns - one relatively inspirational
in nature and focused on intrinsic rewards, and one that is more
exchange-based and focused on extrinsic rewards.
Koh (1990) also conducted a hierarchical regression analysis with
his data to directly test the external validity of the two-factor theory.
Transactional practices added first to the regression analyses explained
none of the variance in three aspects of a multidimensional outcome
measure (organizational citizenship behaviour). When transfonna-
tional practices were added, significant amounts of the variation in one
aspect of this outcome measure was explained. More detailed analyses
suggested the necessity of transactional leadership in producing trans-
fonnational effects, thereby providing support for external validity
claims.
King's (1989) data from 160 educators in K-12 and higher education
settings was collected using a survey of transformational leadership
practices and a measure of organizational climate. Analysis of these
data provided a test of the external validity of the two-factor theory.
The dependent measure, organizational climate, included perceptions
of leader effectiveness, satisfaction with the leader and leadership
style, developmental press, and control press. Using stepwise regres-
sion, transactional leadership measures were entered at a first stage to
determine how much variance was accounted. for in the dependent
measure. Transfonnational leadership measures were entered next to
estimate the amount of additional variation accounted for.
Results of the analysis support the claim that transformational lead-
ership adds value to transactional leadership, but they also call into
question the necessity of transactional leadership in relation to some
dependent measures. With respect to perceptions of leader effective-
ness, transactional leadership explained 19.6% of the variation, trans-
fonnational leadership explained an additional 44.4%. With respect to
satisfaction with the leader, the comparable figures were 20.8% and
54.6%. Both sets of results are consistent with predictions from the
 819

two-factor theory. Transformational leadership was the only significant

predictor of developmental press, a result still consistent with theory.
But control press (feeling constrained in one's actions by the leader),
an outcome predicted for transactional leadership, behaved contrary to
theory: transformational leadership predicted more variation in this
outcome than did transactional leadership. Simple correlations were
also higher between transformational leadership and control press (r =
.61) than between this outcome and transactional leadership (r = .40).
Kirby, King, and Paradise (1991), using data from 103 educators
mostly about K-12 school leaders, tested the external validity of the
two-factor theory. Based on a stepwise regression procedure, they
found that transactional leadership, entered first in the regression,
explained significant variation in both satisfaction with the leader and
perceptions of leader effectiveness. Transformational leadership,
entered into the regression second, significantly increased the amount
of explained variation on both dependent measures. This same study
called into question the internal validity of the two-factor theory, how-
ever. Only charisma/inspiration/vision and laissez-faire dimensions of
leadership were significant predictors of satisfaction, charisma/inspi-
ration/vision and intellectual stimulation predicting perceptions of
effectiveness. Contingent reward was related to transformational rather
than transactional leadership dimensions.
Hoover's (1987) study provided unambiguous support for the inter-
nal validity of the two-factor theory. As with King's (1989) study,
Hoover's data called into question the necessity of transactional leader-
ship in relation to perceived leader effectiveness, and satisfaction with
the leader. With respect to internal validity, a varimax rotation of first-
order factor loadings resulted in two higher-order factors. One such
factor was transformational (or active-proactive) leadership, including
charisma/inspiration/vision, individualized consideration and intellec-
tual stimulation. The second higher-order factor was transactional (or
passive-reactive) leadership, including contingent reward and manage-
ment-by-exception.
Correlations between the three first-order factors defining transfor-
mationalleadership in this study and the two dependent measures were
positive and significant in all but one case. In contrast, correlations
between transactional leadership factors and the two dependent meas-
ures were non-significant, close to zero and, in half the cases, negative.
These results, as with King's (1989), question the necessity of transac-
tional leadership in relation to some outcomes. Indeed, the two out-
comes for which transactional leadership did predict significant
820

variation in King's study were the same two for which correlations
with transactional leadership in Hoover's study approached zero. Such
contradictory results, however, may be a function of different data
analysis techniques.
Based on qualitative evidence collected from a reputationally effec-
tive sample of 51 Texas superintendents, Buck's (1989) study offers
indirect evidence in support of the external validity of the two-factor
theory. Superintendents varied widely in their use of both transactional
and transformational practices: they believed transactional practices to
be necessary and associated them with routine management, whereas
transformational practices were associated with their change efforts.

Effects of Transformational School Leadership

Evidence about the effects of transformational school leadership was

provided by a total of 20 studies, 5 of which were qualitative. These
studies include estimates of the outcomes or effects of transformational
and transactional higher-order factors considered separately; they also
include estimates of the effects of each of the first-order dimensions of
transformational leadership described in Table 2.
The 20 studies as a whole report evidence about 13 different types of
effects (see Table 3). These are grouped into effects on perceptions of
leaders, effects on the behaviour of followers, effects on followers'
psychological states, organizational-level effects, and student effects.
Table 3 describes: the number of studies in which evidence is reported
about each effect; the number of positive relationships found between
each outcome and the total set of transformational dimensions included
in the studies; and the total number of relationships reported to be neg-
ative or not significant (in the case of the 15 quantitative studies, 'sig-
nificant' means statistically significant). The remainder of this section
examines evidence related to each effect, in turn: how the effect was
defined and measured, the weight of the evidence in support of the
effect, and those leadership dimensions which demonstrably contrib-
uted to the effect.
 821

Number TotaIN.S.
Types of Effects or Outcomes Total Posi-
of Studies or Nega-
Measured tive
(N=20) tive

Effects on Perceptions of Leader

1. perceptions of leader effec- 3 15 4

tiveness

2. satisfaction with leader(s) 5 16 5

and style

Effects on Behavior of Followers

3. extra effort 2 5 2

4. organizational citizenship 1 1 3
behavior

Effects on Followers' Psychological States

5. commitment 5 12 10

6. developmental press 4 12 6

7. control press 1 5 3

8. morale/job satisfaction 1 1 0

Organization-level Effects

9. organizational learning 1 5 1

10. organizational improvement! 7 35 16

effectiveness

II. organizational climate and 7 14 1

culture

Effects on Students

12. teachers' perceptions of stu- 6 5 1

dent effects

13. student participation and 2 0 2

identification

Table 3: Effects of Transformational Leadership (Total Number of Studies = 20)

822

Effects on perceptions of leaders

Three studies, two carried out in the u.s. and one in New Zealand,
examine transformational leadership effects on perceptions of leader
effectiveness and satisfaction with the leader (King, 1989; Kirby, King
& Paradise, 1991; Bass, 1985). Data from leaders at both school and
district levels were collected in these studies, all of which used a ver-
sion of Bass's (1988) Multifactor Leadership Questionnaire (MLQ). In
addition to measuring aspects of transformational leadership, the MLQ
includes a two-item scale measuring respondents' satisfaction with
their leader and a four-item scale measuring perceptions of the leader's
effectiveness.
These studies report largely consistent results in relation to both sets
of effects. Positive relationships are reported between transformational
and transactional higher-order factors. Among first-order leadership
dimensions, these effects are most strongly related to charisma/vision!
inspiration, intellectual stimulation, individual consideration and con-
tingent reward. Non-significant relationships are reported in respect to
management-by-exception. Laissez-faire leadership was negatively
related to both perceptions of leader effectiveness and satisfaction with
the leader.
Two additional studies (Koh, 1990; Orr, 1990) report significant
positive relationships between transformational: C and transactional:
C (these are composite, higher-order factors) and satisfaction with the
leader. Although Koh (1990) included Bass's MLQ among the instru-
ments in his study of Singapore principals, he used a subscale of the
Index of Organizational Reactions (Smith, 1976) to measure satisfac-
tion with leaders. Orr (1990) developed two new instruments for his
study of u.S. superintendents. Included in the instrument used to col-
lect data from those working with superintendents was a two-item
scale asking for a rating of the superintendent's performance.

Effect on behaviours of followers

Three studies (Bass, 1985; Koh, 1990; Orr, 1990) inquired about the
effects of transformational leadership on two types of follower behav-
iours: the extent to which followers are prepared to engage in extra
effort on behalf of their organization; and 'organizational citizenship
behaviour'. Two of these studies were about superintendents (Orr's
 823

u.s. study) or other central administrators (Bass's New Zealand

study); Koh (1990) studied Singapore principals.
Bass (1985) assessed extra effort using responses of followers to a
three-item scale included in the MLQ. These same three items were
used by Orr (1990) to collect data from both superintendents and their
immediate subordinates. Both studies report significant positive rela-
tionships between transformational leadership and followers' extra
effort. In the Bass study, extra effort mostly was accounted for by cha-
risma/inspiration/vision, intellectual stimulation and individual con-
sideration. Contingent reward and management-by-exception had non-
significant and negative correlations, respectively, with extra effort.
Orr (1990) reported positive relations between extra effort and both
transformational: C and transactional: C.
Organizational citizenship behaviour, one of the dependent varia-
bles in Koh's (1990) study, was measured using an adapted version of
a questionnaire originally developed by Smith, Organ and Near (1\983):
sixteen items formed 3 sub-scales measuring altruism, compliance and
non-compliance with leaders' suggestions. Transformational: C was
significantly but negatively related to non-compliance This means, as
Koh explains, 'The more the principals were perceived as transforma-
tional, the lower will be the teacher's tendency to take undeserved
breaks, make unnecessary phone calls, and so forth' (1990, p. 113).
Positively related to non-compliance were active and passive manage-
ment-by-exception. No significant relations were evident with contin-
gent reward.

Effect on followers' psychological states

There is evidence from 5 studies (4 quantitative) that transformational

leadership influences four psychological states of those who experi-
ence such leadership: commitment, developmental press, control press,
and satisfaction. Several forms of commitment served as dependent
variables in these studies including teacher commitment to change
(Leithwood, Jantzi, & Fernandez, 1993; Leithwood et aI., 1993), and
organizational commitment (Koh, 1990; Skalbeck, 1991). Considered
to be part of this category of studies as well is Smith's (1989) study of
teacher work motivation.
Evidence reported in the two studies by Leithwood and his col-
leagues were collected from 534 intermediate and secondary teachers
in a large sample of schools in one Canadian province and from 168
824

teachers in nine secondary schools in a second province. Both of these

studies used path analysis techniques to test a model in which transfor-
mational leadership, as well as sets of other in-school and out-of-
school conditions were treated as independent variables. A multi-scale,
multi-item survey of transformational school leadership, developed
and refined in the earlier work of these researchers, was used to collect
evidence about six dimensions of transformational leadership. Depend-
ent variables, measured with items in the same survey, were four cate-
gories of psychological processes identified in the social-psychological
literature as giving rise to teachers' commitment to change. Transfor-
mational: C had significant direct and indirect effects on teachers'
commitment to change. These effects were accounted for most strongly
by: vision building, high performance expectations, developing con-
sensus about group goals, and intellectual stimulation.
Koh (1990) measured organizational commitment with the widely
used, 15 item Organizational Commitment Questionnaire (OCQ)
developed by Mowday, Steers, and Porter (1979). This instrument
describes respondents' loyalty and attachment to an organization, their
agreement with its purposes and values and their willingness to expend
extra effort. Using hierarchical regression analyses, Koh found that
transactional leadership explained a significant proportion of the varia-
tion in organizational commitment. Adding on transformational leader-
ship explained a significant additional proportion of such variation,
however, more than explained by transactional leadership, or any of
several other variables measured in the study.
Skalbeck's (1991) qualitative study inquired about the transforma-
tional practices of one elementary school principal and the extent to
which these practices influenced teachers' commitment to the school's
mission and vision. Leadership practices accounting for such commit-
ment can be classified as individual support, culture building and con-
tingent reward.
Using Bass's MLQ as a measure of 100 superintendents' transfor-
mational leadership, Smith (1989) also collected evidence about the
work motivation of about 100 principals associated with these superin-
tendents. The questionnaire measuring such motivation was based on
the work of Herzberg and others. Stepwise regression analyses of data
from these samples of administrators identified intellectual stimulation
as the dimension of transformational leadership best predicting princi-
pals' work motivation.
One quantitative (Leithwood, Jantzi, & Dart, 1991) and two qualita-
tive (Skalbeck, 1991; Leithwood & Jantzi, 1990) studies reported
 825

effects of transformational leadership on the 'developmental press'

created among those experiencing such leadership. Leithwood, Jantzi,
and Dart (1991) measured transformational leadership using a version
of the instrument described above: developmental press was defined as
changes in teachers' attitudes and school or classroom behaviour. Data
about both sets of variables were provided by 291 elementary school
teachers and 43 principals in two Canadian provinces. Dimensions of
transformational leadership significantly related to developmental
press included: vision building, developing consensus about group
goals, individual consideration, and intellectual stimulation. A dimen-
sion termed 'leadership pressure' had non-significant relationships
with developmental press.
Skalbeck's (1991) study of a single principal suggested that the
stimulus for growth among teachers (developmental press) was a fimc-
tion of the principal's vision-building initiatives and the collegial cul-
ture she was able to build among teaching staff. Interviews with 133
teachers from nine elementary and three secondary case schools pro-
vided the data for Leithwood and Jantzi's (1990) study. Vision build-
ing, culture building, developing consensus about group goals,
intellectual stimulation, and individual consideration made substantial
contributions to teachers' change initiatives in these schools; neither
modelling nor contingent reward had such effects.
Only King's (1989) study provides evidence about the relationship
between control press and transformational leadership (measured using
Bass's MLQ). Control press, one of two higher-order factors measured
in this study with the 80-item Organizational Climate Index (Stem &
Steinhoff, 1965), is indicative of 'an organizational environment that
emphasized high levels of orderliness and structure. The environment
is work-oriented, rather than people oriented.' (King, 1989, p. 94).
Contrary to expectation, King found that transformational leadership
explained substantially more variation in control press than did trans-
actional leadership. Charisma/inspiration/vision, intellectual stimula-
tion, and individual consideration accounted for most of this variation.
Helm's (1989) qualitative study of principals' leadership in U. S.
Catholic elementary schools was the only source of data about the
effects of transformational leadership on teacher morale. Aspects of
leadership contributing to teacher morale included warm, informal,
positive relationships between principals and teachers (individual con-
sideration) and the creation of opportunities for shared decision mak-
ing and leadership (structuring).
826

Organization-level effects

Included among organization-level effects of transformational leader-

ship that have been studied are organizational learning, organizational
improvement and effectiveness, and organizational climate and cul-
ture. The effects of transformational leadership on organizational
learning were examined by Leithwood and his colleagues (1993) in the
context of secondary school improvement efforts. These efforts were
partly in response to a major reform initiated in one Canadian province.
Conditions giving rise to such learning were extracted from the litera-
ture. Both quantitative and qualitative data were collected from school
staffs about the extent to which these conditions prevailed in their
schools and whether and how school leaders influenced these condi-
tions. Organizational learning was defined as an increase in the collec-
tive capacity of organizational members to better accomplish the
purposes of the school. Transformational leadership practices were
helpful in fostering organizational learning; in particular, vision build-
ing, individual support, intellectual stimulation, modelling, culture
building, and holding high performance expectations.
Seven studies report more evidence about the relationship between
transformational leadership and organizational improvement and effec-
tiveness than any other set of outcomes. The majority of these studies
come from one line of Canadian research (Leithwood et aI., 1991,
1992, 1993; Silins, 1992a, 1992b, 1994a), the context for which was
essentially the same as in the study of organizational learning
described above. Both qualitative and quantitative evidence from these
studies suggests that transformational leadership is a powerful stimu-
lant to improvement. Vision building, developing consensus about
group goals, providing intellectual stimulation and individual support,
culture building, and contingent reward were the leadership dimen-
sions that most accounted for this stimulation.
Two other studies also offer evidence relevant to transformational
leadership and organizational effectiveness and improvement, Kend-
rick's (1988) case study of her own leadership in one school over a five
year period makes a compelling case for the contribution of an evolv-
ing set of transformational leadership practices. Koh (1990), using evi-
dence from the Index of Perceived Organizational Effectiveness (Mott,
1972), did not find support for the contribution of transformational
over transactional leadership.
The contributions of transformational leadership to organizational
climate and culture has been assessed in seven studies. Qualitative
 827

studies by Helm (1989), Kendrick (1988), Leithwood and Jantzi

(1990), and Vandenburghe (1991) report generally positive contribu-
tions. In the case of culture, these studies provide extended descrip-
tions of what it means to offer 'symbolic leadership': such practices
help define the meaning of culture building as a distinct dimension of
transformational leadership.
In three quantitative studies, measures used to describe climate and
culture included the School Work Culture Profile (Darling, 1990), the
Organizational Climate Index (King, 1989), and the Leader Behaviour
Questionnaire (Sashkin & Sashkin, 1990). Each instrument measures
something a bit different. Nevertheless, the studies all provide reasona-
bly strong support for the claim that transformational leadership con-
tributes to more desirable school cultures and climates.

Effects on Students

The effects on students of transformational school leadership are likely

to be mediated by teachers and others. This may account for why so lit-
tle evidence has been reported concerning such effects. The complexity
of analyses required to determine the indirect effects of transforma-
tionalleadership on students may also account for the limited evidence.
Until the link between transformational leadership and student effects
is explored in some depth, however, the utility of transformational
leadership in schools will remain uncertain.
Leithwood and his colleagues offer the only evidence available
about student effects. Six studies by this group inquire about the rela-
tionship between transformational school leadership and a construct
they termed 'teacher-perceived student outcomes' (Leithwood, Cous-
ins & G Steinbach, 1993a; Silins & Leithwood, 1994; Silins, 1992b;
Silins, 1994a; Leithwood, Dart, Jantzi & Steinbach, 1991). This con-
struct was measured through items on a survey which asked teachers to
estimate the effects on students of various innovative practices being
implemented in their classrooms. These practices were usually the
product of school-wide initiatives, such initiatives often promoted (or
at least supported) by those in formal school leader roles. While
teacher perceptions offer indirect evidence of student effects,
Leithwood et al. point to the substantial evidence of high correlations
between such evidence and direct measures of student achievement
such as those provided by standardized tests (Egan & Archer, 1985).
Based on similar path analytic data analysis techniques, five of the six
828

studies reported significant indirect effects of transformational school

leadership on teacher-perceived student outcomes.
This same research group also report two studies inquiring about the
effects of transformational school leadership on a construct they
termed 'student participation in and identification with school'
(Leithwood, Cousins & G3; Leithwood, Dart, Jantzi & Steinbach,
1993b). This construct is based on Finn's theory of factors explaining
variation in student retention rates (Finn, 1989; Finn & Cox, 1992).
Responses to the instrument developed for this research were provided
by two samples of secondary school students (N = 3,557 and 2,045).
Neither studies found significant direct or indirect effects of transfor-
mational school leadership on student participation in and identifica-
tion with school.

Summary

This section of the chapter reviewed evidence concerning the internal

and external validity of a conception of transformational leadership as
a 'two factor' theory. Also reviewed was evidence of the effects of
leadership on 13 types of outcomes.
Eight studies inquiring about this conception of transformational
school leadership provided modest support for its internal validity but
in an adapted form. Such an adaptation appears to be required largely
because contingent reward is more often associated with transforma-
tional than transactional leadership. Modest support is also provided by
these eight studies for the external validity of the two-factor theory.
The sometimes ambiguous findings of the studies about external valid-
ity may be, at least in part, a function of the outcomes chosen as
dependent measures. Transactional leadership may contribute little or
nothing to some types of outcomes, although which outcomes is not yet
clear.
Taken at face value, the twenty studies of transformational school
leadership effects suggest that:
Transformational leadership, as a whole, is strongly related to
satisfaction with the leader and positive perceptions of the lead-
er's effectiveness.
Transformational leadership, as a whole, is strongly related to the
Willingness of organizational members to engage in extra effort
but is weakly or negatively associated with most aspects of or-
ganizational citizenship behaviour.
 829

Psychological states of followers including their organizational

commitment and perceptions of both a development and control
press in the school environment are significantly and positively
associated with transformational leadership.
Organization-level effects are also positively associated with
transformational leadership practices. Such practices explain sig-
nificant variation in organizational learning, teachers' percep-
tions of school improvement and effectiveness, and productive
school cultures and climate.
Transformational leadership has significant indirect effects on
teachers' perceptions of student effects but its effects on direct
measures of student outcomes has yet to be demonstrated.
The specific leadership dimensions most consistently explaining
all transformational effects are charisma/vision/inspiration, in-
tellectual stimulation, and individualized consideration. Less
consistent in their effects, but still important are modelling and
holding high performance expectations.
The relatively little studied dimensions of culture building and
structuring show promising effects but results, as yet, are ex-
tremely sparse.
Contingent reward often has positive effects more consistent
with predictions concerning transformational than transactional
leadership.
Management-by-exception (active or passive) typically has neg-
ative effects, whatever outcome is measured.

Twenty studies seems like a reasonably large number from which to

draw inferences about the effects of transformational leadership How-
ever, the potential of the studies for this purpose is weakened consider-
ably by their dispersion across 11 different outcomes. For six of these
outcomes, there is evidence from only three or fewer studies. Indeed,
the relatively large body of evidence available about organizational
improvement and effectiveness (7 studies) and organizational climate
and culture (7 studies) was created only because studies with loosely
related but not identical outcomes were treated as comparable.
The twenty studies are overweighted in favour of those using quanti-
tative methods (15). This suggests a 'hard edge' on the data with dis-
tinct advantages for the purpose of assessing effects of
transformational leadership. Nevertheless, the hardness of that edge is
partly compromised by the possibility of same-source bias. In most of
the quantitative studies, data about both independent and dependent
830

variables were collected using the same instruments and respondents.

Studies by Koh (1990), Smith (1989), King (1989), Darling (1990),
Leithwood, Cousins and Gerin-Lajoie (1993), and Leithwood, Dart,
Jantzi and Steinbach (1993b) are exceptions to this general trend.
Problems of dispersed evidence and same-source bias preclude
claims about the relative contribution of transformational leadership to
specific types of outcomes. There is warrant for claiming that transfor-
mational school leadership has positive effects on schools but not for
specifying what these effects are. Axiomatically,then, it is even more
important to resist other than the most tentative conclusions about
those specific leadership dimensions best explaining significant varia-
tion in different sets of outcomes.

CONCLUSION

A review of literature usually is carried out as a means of 'summing

up' what is formally known about some phenomenon, issue or problem
and assessing the degree of confidence that can be placed in that
knowledge. The first part of this task (summing up) is much simpler to
accomplish than is the second (assessing the resulting knowledge
claims). When the evidence on which knowledge claims are based lend
themselves to quantitative, meta-analytic techniques (e.g., Light & Pil-
lerner, 1984; Rosenthal, 1984), probability theory can be used as one
way of arriving at judgements of confidence (e.g., whether or not an
'effect size' is statistically significant).
When evidence being reviewed does not lend itself to quantitative
meta-analysis, reviewers typically count the number of studies support-
ing and not supporting a knowledge claim (the so-called 'box score'
method), as we have done to this point. This method, however, does
not fully resolve the desire for an assessment of the robustness of the
'summed up' knowledge, since more than the quantity of evidence is
entailed in such an assessment. Other relevant criteria include, for
example, the methodological sophistication of individual studies
reviewed, the extent to which results converge, and the adequacy of
theory to which the data speak. But since those criteria are not easily
combined, the problem the reviewer confronts in offering an overall
assessment seems comparable to the task facing judges of athletic per-
formances or artistic products. Based on sensitivities and impressions
of such performances or products, these judges make comparisons with
personal models of outstanding performances or products developed
 831

through extensive experience. They then express the outcome of this

clearly private comparison in a publicly clear 'metric'.
What would these publicly clear metrics be in the case of a research
review? As part of a review of research on knowledge about teaching,
Fenstermacher (1994) surveys epistemological positions on the status
of formal knowledge. Of particular value in providing readily under-
stood ways of describing judgements about the robustness of knowl-
edge claims are categories used by some epistemologists for this
purpose and outlined by Fenstermacher. Thirteen categories are
described, ranging from 'certainly false' through 'certain'. For pur-
poses of this review, distinctions signified by three category labels
seem most relevant. These categories are:
Evident: in this review, knowledge claims will be judged as Evi-
dent when (a) there is a large amount of supporting research ev-
idence, (b) there is very little or no disconfirming evidence, (c)
supporting evidence is largely derived from research meeting
conventional standards and (d) there are good theoretical argu-
ments in support of the claims.
Beyond Reasonable Doubt: this classification of knowledge
claims will be used when (a) there is a substantial amount of ev-
idence available, (b) a substantial minority of this evidence is
split between non-confirmation and disconfirmation, (c) evi-
dence is provided from a corpus of research ranging in quality
from poor to excellent, and (d) there is a good theoretical argu-
ment in support of the claims.
Epistemically in the Clear: this category will be used when, in
Fenstermacher's terms, ' ... there are not better grounds for reject-
ing the [knowledge claim] than for accepting it' (1994, p. 24).

We conclude this chapter by revisiting the primary knowledge claims

made or inferred by our review. In each case the status of the claim is
classified using one of the three categories described above. Knowl-
edge claims judged to be epistemically in the clear (and sometimes
beyond reasonable doubt) are used as springboards for identifYing use-
ful foci for further research on transformational school leadership.
832

Status of Knowledge Claims About the Nature of Transformational School

Leadership

That section of the review concerned with the nature of transforma-

tional leadership explained such leadership as a working theory for
motivation. Evidence was examined about the relevance to schools of
10 dimensions of transformational leadership and specific behaviours
associated with each were described. Of those 10 dimensions, we judge
the claim that four are productive in school contexts to be evident:
these dimensions include charisma/inspiration/vision, intellectual stim-
ulation and contingent reward. Also evident is the claim that manage-
ment-by-exception is unproductive in school contexts. Our confidence
in these two claims is bolstered by the considerable body ofunambigu-
ous evidence available in the non-school transformational leadership
literature. 12
The claim that the remaining five dimensions are productive in
school contexts is best judged as epistemically in the clear, a judge-
ment consistent with the weak evidence reported in the non-school
research about goal consensus, modelling and high performance expec-
tations; 13 no evidence has been reported about structuring and culture
building in the non-school literature. This judgement leans toward
being conservative since there are good theoretical arguments in sup-
port of the five dimensions. These are compelling reasons for recom-
mending subsequent research be conducted on the effects of these
dimensions in general, as well as the conditions creating variability of
effects. These dimensions may be highly contingent.
The same section of the review also described specific behaviours
associated with each leadership dimension. The claim that such behav-
iours are a sufficient description of what is entailed in practice by each
dimension is epistemically in the clear. But not even that level of confi-
dence can be awarded the claim that they are necessary. A productive
future inquiry aimed at teasing out the essential or necessary behav-
iours associated with each of the leadership dimensions in order to
realize their theoretical effects would be productive. Such research
ought to remain open to the possibility that alternative sets of behav-
iours within a single dimension are called for depending on the context
in which leadership is exercised. We are not aware of any research in
non-school settings which addresses either of these issues.
 833

Status of Knowledge Claims About the Effects of Transformational School

Leadership

The section of the review concerned with transformational school lead-

ership effects examined the validity of a two-factor theory of transfor-
mational leadership in schools and assessed the contribution of
transformational leadership to 13 different sets of outcomes. Claims for
the internal and external validity of a two-factor conception of transfor-
mational school leadership are judged to be beyond reasonable doubt,
if the theory is adapted; this adaptation entails reconceptualizing con-
tingent reward as a transformational dimension, a view consistent with
evidence from non-school settings. However, that leaves transactional
leadership defined only by management-by-exception, a generally
unhelpful if not actually destructive form of leadership in non-school
as well as school contexts. Subsequent research ought to begin from a
more adequate conception of transactional leadership than typically
has been included in transformational leadership research in either
school or non-school settings. It seems likely that there are important,
as yet unidentified (in this body of research) managerial functions, for
example, to which transformational practices add value. The failure to
more adequately conceptualize and measure these functions is a signif-
icant limitation in the corpus of evidence about transformational lead-
ership.
Thirteen outcomes have been used as dependent variables in studies
of transformational school leadership. The claim that such leadership
contributes to organizational improvement/effectiveness, teachers' per-
ceptions of student outcomes, and organizational climate and culture,
is beyond reasonable doubt. In concert with considerable evidence
from non-school settings,14 this same judgement is warranted about
effects on perceptions of leader effectiveness and satisfaction with the
leader. Claims about the effects of transformational leaders on the
remaining outcomes are judged to be epistemically in the clear. An
important contribution to knowledge about the effects of transforma-
tional school leadership would be a sustained program of research
using the same set of dependent measures or outcomes: these outcomes
ought to include not only selected student effects, but also those medi-
ating variables which are both theoretically influenced by transforma-
tional leadership and demonstrably consequential in the production of
student effects, for example teachers' sense of self-efficacy (Ross,
1994); Leithwood, Jantzi & Fernandez, 1994).
834

Many limitations of a theoretical and methodological nature remain

in research on transformational school leadership carried out to date.
Nonetheless, this evidence does provide a strong argument for expand-
ing school leadership research in the direction of transformational per-
spectives. Such an expanded research base, however, should include
studies which directly compare the power of competing conceptions of
leadership to explain variation in important mediating and dependent
variables. For example, the type of work recently reported by Heck and
his colleagues (e.g., Heck, Marcoulides & Lang, 1991) exploring the
effects of instructional leadership could be expanded with the same
samples of school leaders also using measures of transformational
leadership. Presently, there is evidence of quite positive effects of sev-
eral forms of school leadership. But because of differences in the
choice of dependent variables, research methods and the like in gener-
ating such evidence, the effects cannot be compared meaningfully.
While the claim that 'leadership' is a critical variable in school
improvement is evident, claims about the forms such leadership ought
to take are weak, many based on no empirical evidence at all. An
important project yet to be undertaken is a comparative analysis of the
effects of alternative models of leadership in schools.

FOOTNOTES

1. Sometimes the term 'transformational leadership' is used as a

superordinate term encompassing practices which are both trans-
formational and transactional. On other occasions, the term signi-
fies only dimensions of leadership assumed to have
transformational effects.
2. Portions of this section are based on Leithwood, Jantzi & Fernan-
dez (1994).
3. King (1989), Kirby, King & Paradise (1992), Bass (1985), Silins
(1992), Leithwood, Dart, Jantzi & Steinbach (1993a,b, 1991),
Leithwood, Jantzi, Silins & Dart (1993), Smith (1989), Skalback
(1991), Leithwood, Jantzi & Fernandez (1994), Sashkin & Sashkin
(1990), Vandenburghe & Staessens (1991).
4. Leithwood, Dart, Jantzi & Steinbach (1991, 1993a,b), Leithwood,
Jantzi & Fernandez (1994), Vandenburghe & Staessens (1991).
5. Leithwood, Dart, Jantzi & Steinbach (1991, 1993a,b), Leithwood,
Jantzi, Silins & Dart (1993).
6. King (1989), Kirby, King & Paradise (1992), Leithwood, Dart,
 835

Jantzi & Steinbach (1991, 1993a,b), Leithwood, Jantzi, Silins &

Dart (1993), Silins (1992, 1994b), Smith (1989), Skalbeck (1991),
Leithwood, Jantzi & Fernandez (1994), Helm (1989), Bass (1985),
Sashkin & Sashkin (1990).
7. Same studies as listed in note 6.
8. Leithwood, Dart, Jantzi & Steinbach (1991, 1993a,b), Skalbeck
(1991).
9. King (1989), Kirby, King & Paradise (1992), Silins (1992), Silins
(1994b), Leithwood, Dart, Jantzi & Steinbach (1991, 1993a,b),
Leithwood, Jantzi, Silins & Dart (1993), Bass (1985), Smith
(1989).
10. Sometimes 'management-by-exception' is considered to be two
dimensions: 'active management-by-exception' and 'passive man-
agement-by-exception' .
11. This claim must, of course, acknowledge that factors other than
leadership also give rise to change.
12. Howell & Higgins (1990), Podsakoff, MacKenzie, Moorman &
Fetter (1990), Singer & Singer (1986), Tichy & Devanna (1986),
Waldman, Bass & Einstein (1987), Waldman, Bass & Yammarino
(1990), Bass, Waldman, Avolio & Babb (1987), Bass & Avolio
(1989), Bennis & Nanus (1985), Deluga (1991), Hater & Bass
(1988), Seltzer & Bass (1990), Singer (1985), Spangler & Braiotta
(1990).
13. Podsakoff, MacKenzie, Moorman & Fetter (1990).
14. Seltzer & Bass (1990), Spangler & Braiotta (1990), Waldman,
Bass & Yammarino (1990), Yammarino & Bass (1990), Deluga
(1991), Hater & Bass (1988), Singer (1985), Podsakoff, Todor &
Skov (1982), Podsakoff, MacKenzie, Moorman & Fetter (1990),
Waldman, Bass & Einstein (1987).

REFERENCES

Argyris, C., & Schon, D.A. (1978). Organizational learning: A theory of action per-
spective. Reading, Mass.: Addison-Wesley.
Avolio, B.J., & Bass, B.M. (1988). Transformational leadership, charisma, and
beyond. In J.G. Hunt, B.R. BaJiga, H.P. Dachler, & C.A. Schriesheim (Eds.),
Emerging leadership vistas (11-28). Lexington, Mass.: Lexington Books.
Bass, B.M. (1985). Leadership and performance beyond expectations. New York: The
Free Press.
Bass, B.M. (1988). The Multi/actor Leadership Questionnaire, Form 5 (revised). Bing-
hampton, NY: Centre for Leadership Studies, State University of New York @
Binghampton.
836

Bass, B.M. (Ed.) (1981). Stogdill's handbook of leadership: A survey of theory and
research. London: Collier Macmillan Pub.
Bass, B.M., & Avolio, BJ. (1993). Transformational leadership: A response to cri-
tiques. Leadership theory and research: Perspectives and directions (49-80).
Bass, B.M., & Avolio, B.J. (Eds.) (1994). Improving organizational effectiveness
through transformational leadership. Thousand Oaks, Cal.: Sage.
Bass, B.M., Waldman, D.A., Avolio, BJ., & Babb, M. (1987). Transformational lead-
ership and the falling dominoes effect. Group and Organizational Studies, 12, 73-
87.
Bennis, W.B., & Nanus, B. (1985). Leaders: The strategies of taking charge. New
York: Harper & Row.
Boal, K.B., & Bryson, J.M. (1988). Charismatic leadership: A phenomenological and
structural approach. In J.G. Hunt, B.R. BaIiga, H.P. Dachler, & C.A. Schriesheim
(Eds.), Emerging leadership vistas (11-28). Lexington, Mass.: Lexington Books.
Bossert, S., Dwyer, D., Rowan, B., & Lee, G. (1982). The instructional management
role of the principal. Educational Administration Quarterly, 18(3),34-63.
Bright, K.L. (1987). Leadership behaviors of Ohio school superintendents: An exami-
nation of district situational demands and perceptions. Ann Arbor, MI: UMI Dis-
sertation Services.
Bryman, A. (1992). Charisma and leadership in organizations. Newbury Park, Cal.:
Sage.
Buck, J.T. (1989). Transformational leadership behaviors ofexemplary Texas superin-
tendents. Ann Arbor, MI: UMI Dissertation Services.
Burns, J.M. (1978). Leadership. New York: Harper & Row.
Conger, J.A., & Kanungo, R.N. (1987). Towards a behavioural theory of charismatic
leadership in organizational settings. Academy of Management Review, 12(4), 637-
647.
Cunningham, W.G., & Gresso, D.W. (1993). Cultural leadership. Boston: Allyn &
Bacon.
Darling, S.K. (1990). A study to identifY and analyse the relationship between (1)
transformational leadership and collaboration, and (2) transactional leadership
and collaboration in selected Minnesota elementary schools. Ann Arbor, MI: UMI
Dissertation Services.
Deal, T., & Peterson, K. (1990). The principal's role in shaping school culture. Wash-
ington, D.e.: U.S. Department of Education, Office of Educational Research and
Improvement.
Deluga, R.J. (1991). The relationship of leader and subordinate influencing activity in
naval environments. Military Psychology, 3(J), 25-39.
Downton, J.V. Jr. (1973). Rebel leadership. New York: Free Press.
Egan, 0., & Archer, P. (1985). The accuracy of teachers' ratings of ability: A regres-
sion model. American Educational Research Journal, 22( I), 25-34.
Fenstermacher, G.D. (1994). The knower and the known: The nature of knowledge in
research on teaching. In L. Darling-Hammond (Ed.), Review of research in educa-
tion, volume 20 (pp. 3-56). Washington, D.e.: American Educational Research
Association.
Fiedler, F.E., & Garcia, E. (1987). New approaches to leadership: Cognitive resources
and organizational performance. New York: Wiley.
Fieman-Nemser, S., & Floden, R.E. (1986). The cultures of teaching. In M.e. Wittrock
(Ed.), Handbook of research on teaching (pp. 505-526). New York: Macmillan.
 837

Finn, J.D. (1989). Withdrawing from school. Review of Educational Research, 59(2),
117-143.
Finn, 1.0., & Cox, D. (1992). Participation and withdrawal among fourth-grade pupils.
American Educational Research Journal, 29(1), 141-162.
Genge, M. (1993, June). The development oftransformational leaders: Thejourneys of
male and female secondary school principals, alike or different? Presented at the
annual meeting of the Canadian Association for Studies in Educational Administra-
tion, Ottawa.
Halpin, A.W. (1957). The observed leader behaviour and ideal leader behaviour of air-
craft commanders and school superintendents. In R.M. Stogdill & A.E. Coons
(Eds.), Leader behaviour: Its description and measurement. Columbus: Ohio State
University, Bureau of Business Research.
Hater, J.1., & Bass, B.M. (1988). Superiors' evaluations and subordinates' perceptions
of transformational and transactional leadership. Journal ofApplied Psychology,
73(4), 695-702.
Heck, R., Marcoulides, G., & Lang, P. (1991). Principal instructional leadership and
school achievement: The application of discriminant techniques. School Effective-
ness and School Improvement, 2(2), 115-135.
Helm, C.M. (1989). Cultural and symbolic leadership in Catholic elementary schools:
An ethnographic study. Ann Arbor, MI: UMI Dissertation Services.
Hoover, N.R. (1987). Transformational and transactional leadership: A test of the
model. Ann Arbor, MI: UMI Dissertation Services.
House, R.J., & Mitchell. T.R. (1974). Path-goal theory of leadership. Journal ofCon-
temporary Business, 3(4),81-97.
Howell, J.M., & Avolio, B.J. (1991). Predicting consolidated unit performance: Lead-
ership ratings, locus of control and support for innovation. Paper presented at the
51st annual meeting of the Academy of Management, Miami, Florida.
Howell, J.M., & Higgins, C.A. (1990). Champions of technological innovation.
Administrative Science Quarterly, 35, 317-341.
Hunt, 1.G., Baliga, B.R., Dachler, H.P., & Schriesheim, c.A. (1988). Emerging leader-
ship vistas. Lexington, MA: Lexington Books.
Jantzi, D., & Leithwood, K. (1994). The role ofgender and age in explaining transfor-
mational school leadership. Toronto: Ontario Institute for Studies in Education,
mimeo.
Kendrick, J.A. (1988). The emergence of transformational leadership practice in a
school improvement effort: A reflective study. Ann Arbor, MI: UMI Dissertation
Services.
King, M.1. (1989). Extraordinary leadership in education: Transformational and
transactional leadership as predictors of effectiveness, satisfaction and organiza-
tional climate in K-12 and higher education. Ann Arbor, MI: UMI Dissertation
Services.
Kirby, P.C., King, M.I., & Paradise, L.V. (1992). Extraordinary leaders in education:
Understanding transformational leadership. The Journal of Educational Research,
85(5), 303-311.
Koh, W.L.K. (1990). An empirical validation of the theory of transformational leader-
ship in secondary schools in Singapore. Ann Arbor, MI: UMI Dissertation Serv-
ices.
838

Kushner, R. (1982). Action theory congruence and the exercise of transformational

leadership in Catholic elementary schools. Ann Arbor, MI: UMI Dissertation Serv-
ices.
Lehr, K.A. (1987). A descriptive study of contemporary transformational leadership.
Ann Arbor, MI: UMI Dissertation Services.
Leithwood, K. (1992). The move towards transformational leadership. Educational
Leadership, 49(5),8-12.
Leithwood, K., & Jantzi, D. (1990). Transformational leadership: How principals can
help reform school cultures. School Effectiveness and School Improvement, 1(4),
249-280.
Leithwood, K., & Steinbach, R. (1991). Indicators of transformational leadership in the
everyday problem solving of school administrators. Journal of Personnel Evalua-
tion in Education, 4(3),221-244.
Leithwood, K., & Steinbach, R. (1993). Total quality leadership: Expert thinking plus
transformational practice. Journal ofPersonnel Evaluation in Education, 7(4), 311-
337.
Leithwood, K., & Steinbach, R. (1994). Expert problem solving: Evidence from school
and district leaders. Albany, NY: SUNY Press.
Leithwood, K., Cousins, B., & Gerin-Lajoie, D. (1993). Years of Transition, times for
change: A review and analysis ofpilot projects investigating issues in the Transi-
tion Years (Volume 2: Explaining Variations in Progress). Toronto: Final report of
research to the Ontario Ministry of Education.
Leithwood, K., Dart, B., Jantzi, D., & Steinbach, R. (1991). Building commitment for
change: A focus on school leadership (Final report for year two of the research
project: Implementing the primary program). Prepared for the British Columbia
Ministry of Education.
Leithwood, K., Dart, B., Jantzi, D., & Steinbach, R. (1992). Fostering organizational
learning: A study of British Columbia's intermediate developmental site initiatives
(Final report). Victoria, BC: British Columbia Ministry of Education.
Leithwood, K., Dart, B., Jantzi, D., & Steinbach, R. (1993a). Fostering organizational
learning: A study in British Columbia's intermediate developmental sites, 1990-
1992 (Final report for year three of the research project: Implementing the Year
2000 poliCies). Prepared for the British Columbia Ministry of Education.
Leithwood, K., Dart, B., Jantzi, D., & Steinbach, R. (1993b). Building commitmentfor
change and fostering organizational learning: Final report for phase four of the
research project: Implementing British Columbia's education policy. Prepared for
the British Columbia Ministry of Education.
Leithwood, K., Jantzi, D., & Dart, B. (1991). Toward a multi-level conception ofpolicy
implementation processes based on commitment strategies. Paper presented at the
Fourth International Congress on School Effectiveness, Cardiff, Wales.
Leithwood, K., Jantzi, D., & Fernandez, A. (1994). Transformational leadership and
teachers' commitment to change. In J. Murphy & K. Louis (Eds.) (in press),
Reshaping the principals hip. Newbury Park, Cal.: Corwin Press.
Leithwood, K., Jantzi, D., Silins, H.C., & Dart. B. (1993). Using the appraisal of
school leaders as an instrument for school restructuring. Peabody Journal of Edu-
cation, 68(2),85-109.
Light, R.1., & Pillemer, D.B. (1984). Summing up: The science of reviewing research.
Cambridge, Mass.: Harvard University Press.
 839

Little, J. (1982). Norms of collegiality and experimentation. American Educational

Research Journal, 19(3),325-340.
Mott, P.E. (1972). The characteristics of efJective organizations. New York: Harper &
Row.
Mowday, R., Steers, R., & Porter, L. (1979). The measurement of organizational com-
mitment. Journal of Vocational Behaviour, 14,224-247.
Nanus, B. (1992). Visionary leadership. San Francisco: Jossey-Bass.
Nias. J., Southworth, G., & Campbell, P. (1989). StafJrelationships in the primary
school. London: Cassell.
Orr, D.R. (1990). An expectancy theory investigation ofschool superintendent job per-
formance. Ann Arbor, MI: UMI Dissertation Services.
Podsakoff, P.M., MacKenzie, S.B., Moorman, R.H., & Fetter, R. (1990). Transforma-
tionalleaders' behaviors and their effects on followers' trust in leader, satisfaction,
and organizational citizenship behaviors. Leadership Quarterly, 1(2), 107-142.
Podsakoff, P.M., Todor, W.D., & Skov, R. (1982). Effect ofleader contingent and non-
contingent reward and punishment behaviors on subordinate performance and satis-
faction. Academy ofManagement Journal, 25(4), 810-821.
Roberts, N.C. (1985). Transforming leadership: A process of collective action. Human
Relations, 38(11), 1023-1046.
Rosenthal, R. (1984). Meta-analytic procedures for social research. Beverly Hills:
Sage.
Ross, J. (1994). A review of research on teacher self-efficacy. Toronto: Ontario Insti-
tute for Studies in Education, mimeo.
Sashkin, M. (1985). Trainer guide: Leader behavior questionnaire. Bryn Mawr, PA:
Organization Design and Development.
Sashkin, M., & Sashkin, M. (1990, April). Leadership and culture-building in schools:
Quantitative and qualitative understandings. Paper presented at the annual meeting
of the American Educational Research Association, Boston.
Schein, E.H. (1985). Leadership and organizational culture. San Francisco: Jossey-
Bass.
Seltzer, J., & Bass, B.M. (1990). Transformational leadership: Beyond initiation and
consideration. Journal ofManagement, 16(4), 693-703.
Senge, P. (1990). The fifth discipline: The art and practice of the learning organiza-
tion. London: Doubleday.
Silins, H.C. (1992). Effective leadership for school reform. The Alberta Journal of
Educational Research, 4, 317-334.
Silins, H.C. (1994a). The relationship between transformational and transactional
leadership and school improvement outcomes. Adelaide: The Flinders University
of South Australia.
Silins, H.C. (1994b). Analysing leadership and its components: What makes the differ-
ence? Adelaide: The Flinders University of South Australia.
Silins, H.C., & Leithwood, K. (1994). The relative impact offactors influencing stu-
dent outcomes in the transition years. Adelaide: The Flinders University of South
Australia.
Sims Jr., H.P., & Lorenzi, P. (1992). The new leadership paradigm: Social learning
and cognition in organizations. Newbury Park, Cal.: Sage.
Singer, M.S. (1985). Transformational vs. transactional leadership: A study of New
Zealand company managers. Psychological Reports, 57, 143-146.
840

Singer, M.S., & Singer, A.E. (1986). Relation between transformational vs. transac-
tionalleadership preference and subordinates' personality: An exploratory study.
Perceptual and Motor Skills, 62, 775-780.
Skalbeck, K.L. (1991). Profile of a transformational leader: A sacred mission. Ann
Arbor, MI: UMI Dissertation Services.
Smith, C.A., Organ, D.W., & Near, J.P. (1983). Organizational citizenship behavior:
Its nature and antecedents. Journal ofApplied Psychology, 68, 653-663.
Smith, F.J. (1976). The index of organizational reaction (lOR). JSAS catalogue of
selected documents in psychology: Volume 6, MS. No. 1265.
Smith, J.G. (1989). The effect of superintendent leader behavior on principal work
motivation. Ann Arbor, MI: UMI Dissertation Services.
Spangler, W.D., & Braiotta, L. (1990). Leadership and corporate audit committee
effectiveness. Group and Organizational Studies, 15(2), 134-157.
Tichy, N.M., & Devanna, M.A. (1986). The transformational leader. New York: John
Wiley & Sons.
Vandenburghe, R., & Staessens, K. (1991, April). Vision as a core component in
school culture. Paper presented at the annual meeting of the American Educational
Research Association, Chicago.
Waldman, D.A., Bass, B.M., & Einstein, W.O. (1987). Leadership and outcomes of
performance appraisal processes. Journal of Occupational Psychology, 60(3), 177-
186.
Waldman, D.A., Bass, B.M., & Yammarino, FJ. (1990). Adding to contingent reward
behavior: The augmenting effect of charismatic leadership. Group and Organiza-
tional Studies, 15(4),381-394.
Weber, M. (1947). The theory of social and economic organization (A.M. Henderson
& T. Parsons, translators). T. Parsons (Ed.). New York: Free Press.
Webster's Seventh New Collegiate Dictionary (1971). Toronto: Thomas Allen & Sons.
Yammarino, FJ., & Bass, B.M. (1990). Transformational leadership and multiple lev-
els of analysis. Human Relations, 43( 10), 975-995.
Yuki, G. (1989). Leadership in organizations (2nd edition). Englewood Cliffs, N.J.:
Prentice-Hall.