Content uploaded by Paul Englert
Author content
All content in this area was uploaded by Paul Englert on Mar 24, 2016
Content may be subject to copyright.
0
Word Count: 5270
The Development and Application of a three-factor
Leadership Model
Paul Englert, Sue Seymour & Sarina Johnstone
Paul Englert, Phd
OPRA Consulting Group
Level 4 James Cook Arcade
Wellington
New Zealand
Ph: +64 4 499 2884
Fax: +64 472 3500
Sue Seymour
Previously with: Ministry of Social Development
Level 9, Bowen State Building
Wellington
Wellington
New Zealand
Ph: +64 49168-375
Fax: +64 49163-441
Sarina Johnstone
Previously with: OPRA Consulting Group
Christchurch Business Centre
240 Armagh Street
Christchurch
Ph: +64 3 379 7377
Fax: +64 3 379 7379
1
The Development and Application of a three-factor
Leadership Model
Paul Englert, Sue Seymour & Sarina Johhstone
There is a range of theories that have been developed to describe leadership.
The application of these theories is often hindered by both their complexity and
the knowledge required by Human Resource practioners to effectively implement
a leadership model to solve organisational issues. Pragmatic leadership models
must, therefore, be practical and theoretically robust. This paper outlines the
development of a leadership model developed for use with managers in a New
Zealand Government Department. Repertory grid analysis was used to
reconceptualisation the previous leadership models used by the Department .
The leadership questionnaire development is discussed with specific reference to
a three-factor model of leadership that accounted for 52% of the variance related
to leadership behaviour. The three factors were labelled vision, operational focus
and positive energy. The practical usefulness of the leadership model is
discussed with reference to current literature in the field of leadership and
performance appraisal as well as specific interventions now employed by the
organisation as part of its leadership programme.
2
Since at least the 1800’s value has been placed on qualifying the components of
leadership (Landy, 1989). Bass (1990) stated that:
“There are almost as many different definitions of leadership as there are people who
have tried to describe the concept” (Bass, 1990, pg. 11).
With this noted, leadership from an organisational perspective could be considered as a
social influence in an organisational setting in order to achieve organisational goals
(Saal & Knight, 1995). Social influence as a definition fails however to describe the
qualities that underpin leadership, the identification of which is important to all
organisations seeking development. There is still debate today as to whether a leader is
born, as specified by trait theories, or can be developed or learned as suggested by
theories such as Vertical-Dyad Linkage model (Dansereau, Graen and Hager, 1975) and
Path-Goal theories of leadership (House, 1971). There is nevertheless agreement
between leadership theorists that management is very different to leadership (Hart and
Waisman, 2005).
The initial theories of leadership were focussed around traits that could be used to
predict a person’s leadership potential. Trait theories of leadership suggested:
“...there are certain inborn qualities such as initiative, courage, intelligence and humour,
which altogether pre-destine a man (sic) to be a leader ...the essential pattern is given at
birth” (Adler, 1991, p. 4)
The trait theory of leadership has its origin in the ‘great-man’ theory (Saal & Knight,
1995). Indeed, much of the initial work on leadership focussed on one key trait, that of
being male! (Saal & Knight, 1995). The theory assumes that the course of history has
been shaped by the acts of key individuals, for example Mahatma Gandhi in India.
Thus, in turn, the identification of key figureheads in society gave rise to the theories of
leadership that centred on the identification of traits that may underpin leaders. While
many theories of leadership suggest that, to a degree, traits are responsible for
leadership behaviour, trait theories often neglect the situational influences or
environment factors of behaviour.
If trait theory were to be fully utilised in organisations within New Zealand emphasis
would be heavily placed on the identification of leaders at the time of selection as no
scope would be available for the development of employees to become leaders unless
they had the appropriate traits. The implications of this theory would be large in New
Zealand’s current tight labour market (http://www.stats.govt.nz). Those born with the
desired traits would need to be enticed from their current positions to fill needed roles.
This would not be easy for a Government Department with limited resources and
3
funding. Measuring a trait would also be difficult. Psychometric assessments available
today infer traits from preferred behaviour (Barrick, Mount, and Judge, 2001).
Theories of leadership have evolved over time from enforcing one best form of
leadership to multiple best approaches to leadership as prescribed by the situation
(House, Jarvidan, Hanges and Dorfman, 2002). Stogdill (1948) noted that no single trait
was necessary for leadership and that the traits that were associated with successful
leadership varied from situation to situation. Stogdill (1948) reviewed the early literature
on leadership traits and found evidence suggesting leaders were more intelligent,
scholarly, dependable, active, participative, and were of higher socio-economic status
than non-leaders. While later work has identified traits associated with leadership such
as intelligence, masculinity, and dominance (Lord, Devade, & Alliger, 1986); self-
monitoring behaviour (Zacarro, Foti, & Kenny, 1991); and motivation to manage (Miner,
1978); the role of the situation in promoting leadership has been found to complement
the role of traits in the prediction of leadership. This is supported by the research of
behaviourists who have argued that the key to identifying the characteristics of
leadership is not the identification of certain traits, but rather how those who are led
reinforce these traits (Liden & Graen, 1980). However the reliance on an individual to
possess the relevant traits to lead successfully is still core to this approach which again
does not allow for scope to develop a team of managers already onboard within an
organisation.
The interaction between the environment and traits gave rise to contingency models of
leadership (Saal & Knight, 1995). The most prominent of the contingency theories is
that of Fiedler (1964, 1967). According to Fiedler, leader effectiveness depends on the
interaction of two factors: the degree to which the leader’s situation is favourable for the
exercise of influence and leadership style or personality. Fiedler (1967) defined the
favourability of the situation for leadership in terms of how well the leaders and
subordinates get along; how clearly defined people’s tasks were; and the perceived
status of the leader. Furthermore, a person’s relationship with their LPC can also be
used to differentiate leadership styles as to whether one is more likely to be task –
oriented (reflective in a negative description of the LPC) or relationship-oriented
(reflective in a positive description of their LPC). The value that a leader places on a
person with whom they find it particularly difficult to work with is used by Fiedler (1967)
as a measure of a person’s interpersonal skills. The Contingency Model (Miller, Butler
and Cosentino, 2004) proposes that leaders identified as task-oriented by the LPC were
more likely to succeed when in the very unfavourable or very favourable situations.
Those identified as relationship oriented are proposed to be more successful in
moderately favourable situations. This model does not lend itself easily for practical
4
application in regard to organisational leadership implementation exercises for
development. Management participants would need to be coached to adapt their style
(task or relationship oriented) to fit with the favourability of the situation. Participants
would also need to be trained to constantly monitor and categorise the situations
favourability.
The emphasis on relationship versus task orientation is also reflected in the model of
transformational leadership (Bass, 1985). Transformational leadership involves the
engagement of followers in such a way that their leader transforms them. In this sense
followers undergo personal-growth by adopting the values and goals of the leader.
Transactional leadership involves influencing subordinates by exchange of rewards for
compliance with the leaders instructions. The definition of purely Transactional
leadership as opposed to more Transformational leadership echoes the distinction made
between management and leadership (Hart and Waisman, 2005). Bass (1989) and
Bass and Avolio (1985) have found that transformational leaders are more effective,
have better leader-subordinate relations, and greater productivity and growth. To
measure Transformation leadership Bass and Avolio (1995) developed the Multifactor
Leadership Questionnaire (MLQ). Avolio, Bass and Jung (1999) adapted the MLQ to
correct item wording and to increase the differentiation of leadership factors. They
identified six lower order factors and 3 higher order factors: transformational
(charismatic/inspirational and intellectual stimulation), transactional/developmental
(individualised consideration and contingent reward) and corrective avoidant
(management by exception and laissez-faire) leadership. The MLQ (Form 5X) is
completed by subordinates, which raises issues of rater bias due to pressure to respond
in a manner for which the manager will look favourable.
Leader – Member Exchange (LMX) is based on the Vertical Dyad model of leadership
(Schriesheim, Castroa, Zhoua and. Yammarino, 2001). This model focuses specifically
on the leader – subordinate relationship to determine leadership effectiveness. The
situation is proposed to either restrict a manager to a supervisory role or allow for
influence of subordinate behaviour without authority. The exchanges between the
leader and subordinate over time are proposed to shape the relationship and therefore
leadership effectiveness. Schriesheim, Castroa, Zhoua and. Yammarino (2001)
conducted an analysis of the LMX model at a dyadic level as previous research
conducted on this approach had only investigated either party (leader or subordinate)
and not the relationship between the two. They found that generally it was supported
however further research was required to qualify the theory at the appropriate level-of-
analysis that it hypothesized.
5
Path-goal theory adapted to include situational consideration by House (1971)
hypotheses that the leader influences subordinates by making them aware of the value
of the specified goals. Global Leadership and Organisational Behaviour Effectiveness
(GLOBE) is an international study conducted by House and colleagues, (House,
Jarvidan, Hanges and Dorfman, 2002) on leadership and the effect that culture has on
the leadership effectiveness. In order to measure leadership attributes a questionnaire
was developed specifically for the purpose of global distribution and measurement of
organisation and societal culture, and leadership (House, Jarvidan, Hanges and
Dorfman, 2002). Six attributes of leaders were assessed using the questionnaire.
Respondents marked on a quantitative scale what they believed constituted leadership.
Data has been collected across 62 societies and a book published on the collection:
Culture, Leadership, and Organizations The GLOBE Study of 62 Societies (House,
Jarvidan, Hanges, Dorfman and Gupta, 2004) The third phase, analysis of leadership
behaviour on subordinate’s attitudes, performance and leadership effectiveness (House,
Jarvidan, Hanges and Dorfman, 2002).
Posner and Kouzes (1988) constructed the Leadership Practices Inventory (LPI) in order
to measure leader behaviour. They identified five core factors of leadership: challenging
the process, inspiring a shared vision, enabling others to act, modelling the way and
encouraging the heart. These scales are assessed by leaders and their subordinates on
a 5 point likert scale of frequency scale. Analysis has shown this measure of leadership
to be a good predictor of leader performance (Posner and Kouzes, 1988).
Measuring leadership
While theories of leadership provide a framework for understanding leadership, it is the
measurement of leadership that is crucial for identification of future leaders within, and
for, organisations. The benefit of theoretical models to the practioner is directly
dependant on the ability to quantify the components that a particular model says
characterises leadership. Consequently, measurement of leadership is a necessary
precursor to demonstrating the utility of leadership models. Practical utility in this
context can be defined by the application of leadership models to core Human Resource
(HR) functions such as training (training future leaders), selection (selecting staff with
leadership qualities), succession planning (identification of potential leaders from
existing staff), and performance appraisal (monitoring and reinforcing leadership
qualities among staff). Thus, a theoretical model of leadership relies on appropriate
measurement of its constituent
components in order to demonstrate tangible benefits to
the organisation.
6
In order to measure the components of leadership one must assume that leadership
exists as a quantifiable amount. The greater the number of factors included in a
questionnaire does not necessarily increase the information obtained of the leader in
question. In fact the greater the number of factors could increase the difficulty of
implementation for those using the questionnaire. In addition, there are some aspects of
leadership that are less easily assessed by a questionnaire. Avolio and Gardner (2005)
differentiate ‘authentic leadership’ an altruistic core of leadership as opposed to
leadership with personal agendas that they guide others to achieve on their behalf. The
ethically/authenticity or integrity of leadership has often been a point of difficulty in
regard to the comprehensive measurement of leadership behaviour (Kouzes and
Posner, 1992).
Another difficulty of leadership measurement is whether leadership questionnaires need
to be contextualised to accurately measure leadership behaviours so as to relate more
specifically to a particular organisation. If the tool is more closely related to the
organisation the face validity of the tool is likely to increase, in turn increasing meaning
for respondents however this may also increase respondent distortion (Cattell and
Warburton, 1967). Issues of parsimony must also be considered. Questionnaires must
be succinct and easy to interpret, while maintaining validity. This increases usefulness
and applicability, as many organisations do not have access to trained test interpreters
and are hindered by competing demands on time. Implicit leadership theories of
respondents must also be considered when constructing a leadership questionnaire.
Attention should be placed on behaviours rather than subject judgements in order to
avoid bias (Gioia and Sims, 1986).
The current study
This study set out to create a behavioural level leadership questionnaire to measure
leadership within a large New Zealand public sector organisation. The measurement of
leadership was required to identify those with leadership potential in order to take part in
leadership development programmes to grow their ability. Strong leadership was
identified by the department as key to achieving its core and operational objectives, as
well as adherence to the organisational values.
The questionnaire was designed in line with common psychometric principles and
formed part of training needs analysis by:
a) Identifying the leadership competencies in the organisation;
b) Identifying staff who were deficient in the leadership competencies and therefore
had training needs.
7
The leadership questionnaire was built over a series of iterative stages aimed at
ensuring the tool was robust, both statistically and theoretically. The following research
study outlines each of the various development stages. Application of the resulting
leadership questionnaire is then discussed.
Stage 1: Developing a new model of leadership
A workshop was conducted with the HR team to clearly identify the competencies that
conceptually defined leadership within the Department. The aim of the workshop was to
identify the smallest set of competencies that were independent and captured their
vision of leadership. A revised repertory grid procedure (Kelly, 1955) was used that
encouraged members of the project team to identify how the current competencies used
to define leadership in the Department were both similar and different from each other.
This procedure involved the development of competencies from opposite ends of the
scale i.e. expedient versus conscientious, analytical versus creative. These opposite
pairs of competencies were compared to a single competency to identify similarities and
differences. For example:
“ How are modelling and influencing similar, but different from visionary?”
Each of the original leadership competencies used by the Department was clustered in
this way, alternating the way that competencies were paired, until definitions of
competencies emerged. In this way a new core set of competencies was defined that
were theoretically independent of each other and empirically differentiable from one
another. This revised model in turn provided the framework for the development of the
leadership questions.
The repertory grid workshop provided two key outcomes for the project. Firstly, the
competency model developed to define leadership within the department was altered to
create greater distinction between each of the competency areas. This created a degree
of theoretical independence between the various factors, which in turn is essential to
ensure that the measurement is not confounded by reducing the clarity of what is being
measured inside each factor.
A secondary outcome of the workshop was that the behaviours that underpinned each
competency were identified. In having to explain how each competency area was
different or the same as others the HR team was able to identify the types of behaviours
that they believed defined each competency area. From this process behaviours, which
8
in turn could be developed into questions that best captured the essence of each of the
leadership competencies, were identified.
Stage 2: Developing a pilot version of the leadership questionnaire
The revised leadership assessment tool was developed as an on-line or e-mail
questionnaire that assessed leadership behaviour and potential, as measured through
self and managerial ratings. The tool was developed by the HR team to measure the
revised competency model identified as important for leadership within the Department.
The revised competency model consisted of six core leadership competencies.
Innovates and Inspires were kept as one competency as they were unable to be
separated in the original rep grid workshop. The six competencies were:
• Visionary;
• Influencing;
• Modelling;
• Motivating;
• Achieving Results;
• Innovates and Inspires
Based upon the revised leadership competencies discussed above a series of questions
were developed. 16 questions per competency were developed. The rationale behind
this was that it allowed redundant questions to be identified and excluded from a final
solution without potentially hindering the reliability of the questionnaire. Given that the
final questionnaire would have approximately seven to eight questions per factor, and
that 2/3 to 1/2 of the questions would be discarded, 16 questions achieved the balance
of also ensuring the minimisation of respondent burden.
Both positive and negative valence items were included in the questionnaire. This
aimed to increase the range of responses and ensure that respondents would have to
engage in each question and therefore not just default their responses to the agree box
only.
A behavioural observation rating scale was used to measure responses. This meant
judgement rating was replaced with ratings based on observable behaviour. For
example, the upper rating of behaviour was changed from excellent (a value judgement)
to always (frequency of behaviour).
9
Participants
An initial study using the revised questionnaire was completed. This involved sending an
invitation to all managers within the Department and asking them, and their managers, to
complete the questionnaire. It was anticipated that over 200 self-raters would complete
the questionnaires and for each of those raters a manager would also complete a
questionnaire. The resulting sample size however was much smaller than anticipated
with only 90 self and 90 managerial ratings being completed. Reasons for the low
response rate were discussed with the HR team who believed that a lack of support for
the process by some managers and heavy workloads were the key reason for the low
response rate.
180 participants therefore completed the revised leadership questionnaire.
Demographics details were not recorded, as the HR team viewed the custom
development the key factor in determining the applicability of the instrument in their
organisation. As the questionnaire was designed as a managerial training needs
analysis tool to be completed by managers and their managers, a breakdown was
available on whether a participant was a manager or a staff member (i.e. a person who
was responsible to the manager being rated). 90 participants were classified as staff
members and 90 participants were classified as managers.
Materials
The pilot questionnaire consisted of 96 items that were evenly spread across the 6
leadership competencies. These competencies aimed to measure behaviour in the
areas of: vision, influence, modelling, motivating, achieving results, and innovating and
inspiring.
Design and procedure
The questionnaire was completed via the World Wide Web. A web site was set up that
hosted the leadership questionnaire and participants who were asked to complete the
questionnaire were all sent an email with the URL. Participants visited the website,
completed the questionnaire, and the responses were then submitted automatically to a
database that was used to collate the results. Access to this website was controlled by a
third party. Information was not captured on who and why people failed to complete the
questionnaire.
10
Once all questionnaires had been completed, a principal component factor analysis was
conducted on the data. The reason for using principle component analysis was tied to
its advantage for the analysis of psychological data. These are as follows (Nunnally,
1978):
1. The principle components are ordered by the magnitude of variance accounted
for. Each factor accounts for more variance than any other method would allow.
2. In most cases the first component explains far more variance than do the other
components and the Eigen values drop off sharply as a result. The result for
psychological tests is that a small set of key factors will result. This adds in
interpretation and test use, as the final solution is high in parsimony.
The purpose of this factor analysis was to identify the underlying factor structure of the
questionnaire. This in turn would potentially shorten a revised version of the
questionnaire by identifying questions that did not account for sufficient variance and did
not fit within the defined factor structure of the questionnaire. Managerial and staff
results were combined into one data set and factor analysed using principal
components.
11
Results
The responses from the completed questionnaires were analysed to establish a
structure for the final questionnaire. This analysis included a principal component
analysis to identify the underlying structure of the questions in terms of how respondents
clustered the various aspects of leadership. Cronbach-Alpha coefficients were also
calculated to measure the internal consistency of each component. Inter-item
correlations were assessed to measure the breadth of measurement within a factor.
Correlations between components were assessed to measure independence and
redundancy in scales. Practical usefulness was also assessed to identify the theoretical
meaning of the question clusters.
12
Table 1
Factor solution for the Leadership questionnaire
Factor Loadings (Varimax normalized)
(st4.sta)
Extraction: Principal components
Factor Factor Factor
1 2 3
Q10 -0.04278 -0.65734*
-0.04756
Q11 -0.08991 -0.73954*
-0.07243
Q20 0.246659
0.255474
0.68044*
Q23 0.196757
0.19634 0.551009*
Q25 0.730105*
0.181127
0.224046
Q26 0.266232
0.604163*
0.160849
Q30 0.625422*
0.070831
0.128903
Q35 0.545844*
0.012584
0.257845
Q36 0.294232
0.199727
0.807598*
Q38 0.739147*
0.158493
0.308705
Q40 0.307251
0.631766*
0.112531
Q43 -0.20379 -0.67829*
-0.27175
Q44 -0.10549 -0.51331*
-0.26057
Q46 0.766773*
0.11996 0.263451
Q53 0.309157
0.146993
0.775494*
Q54 0.36791 0.033586
0.134152
Q56 0.290731
0.222916
0.752798*
Q58 0.205013
0.295474
0.712709*
Q60 0.250509
0.177138
0.850662*
Q61 0.327032
0.235759
0.754654*
Q64 -0.12255 -0.568* -0.23667
Q65 0.628268*
0.205229
0.156677
Q71 0.672079*
0.280742
0.153809
Q72 0.69726*
0.245759
0.232469
Q74 0.366976
0.498072*
0.24235
Q75 0.477327*
0.145901
0.210477
Q84 0.017759
-0.57027*
-0.20422
Q85 0.698125*
0.085477
0.110132
Expl.Var
5.495414
4.040455
5.206914
Prp.Totl
0.196265
0.144302
0.185961
Mean 45.22 54.44 48.75
SD dev
5.92 6.82 5.73
Sample
180 180 180
* Primary loading (>0.4)
13
Plot of Eigenvalues
Number of Eigenvalues
Value
0
1
2
3
4
5
6
7
8
9
10
11
12
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Figure 1: Scree-plot of the factors derived from the Leadership Questionnaire
Table 2
Eigenvalue of each factor, total variance explained and cumulative Eigen value
and variance explained
% total Cumul. Cumul.
Eigenval Variance Eigenval %
1 10.46971
37.39184
10.46971
37.39184
2 2.311242
8.254437
12.78096
45.64628
3 1.961825
7.006517
14.74278
52.65279
14
Table 1 outlines the results from the factor analysis. The data matrix was rotated using
a varimax rotation. Varimax rotation was chosen by the researcher to minimise
correlation between factors (Cattell, 1978). Table 1 indicates that there are clearly
loadings within each factor evident by the presence of eigenvectors over 0.3. Moreover,
many of these loadings appear to be prime loaders within that factor evident by the
prevalence of Eigenvectors over 0.7.
The scree plot in figure 1 indicates that there are only three factors that are above or
close to an Eigen value of 2. Table 2 indicates that these three factors account for over
50% of the variance in the way that people answered the questionnaire.
It is clear from table 1 that most items appear to be loading on only one factor.
Moreover, for the items that load on two factors, given that one of these loadings is a
prime loader, this negates the low but moderate loading on another question. This is the
case with three out of the five questions that have a double loading in the leadership
questionnaire.
Table 3
Canonical correlation and redundancy in each factor
Factors being
correlated
Canonical
correlation (2 dp.)
Redundancy in second
factor given first factor
1-2 0.63 17%
1-3 0.66 30%
2-3 0.68 27%
Table 3 indicates that the three-factor leadership questionnaire is highly internally
consistent, i.e. the questions in each factor are measuring aspects of the same
construct. Coefficients of over 0.7 are considered as indicating good reliability, however
the validity of a test must also be taken into consideration for an overall judgement of a
test (Guilford, 1956). The items are, however, measuring a unique part of the overall
factor as given by the lower Inter-Item Correlation (IIC). Factor 3 however has a fairly
high IIC indicating that the questions that make up this factor are likely to overlap to a
greater degree and contain greater item redundancy than items within the other factors.
15
Table 4
Alpha-Cronbach coefficients and average inter-item correlation (IIC) for the
factors in the Leadership questionnaire
Reliability
IIC
Factor 1
0.84 0.44
Factor 2
0.81 0.36
Factor 3
0.92 0.63
Table 4 notes the correlations between each of the factors. These correlations are
relatively high (over 0.6). This indicates that the three factors are likely to have a degree
of overlap. This is supported by an examination of the redundancy between factors.
Redundancy between factors varied between 17-30%, with factor 3 having a high
(>25%) level of redundancy with both factors 1 and 2.
Table 5 indicates the theoretical consistence of the final PCA solution. The titles given
to each component have been developed to capture the cluster of items. As each
cluster of items could meaningfully be related internally the theoretical consistence of the
scale can be seen to be strong.
16
Table 5
Items that make up each factor
Component 1: - Future oriented
• Has a vision of where they would like the Department to move
• Enjoys the challenge of new and unusual tasks
• Has credibility as a spokesperson for the Department
• Anticipates the future needs of the Department
• Sees ahead and anticipates future consequences and trends accurately
• Is not afraid to take people out of their comfort zones
• Sets stretching objectives
• Continually seeks new and better ways of doing things
Component 2:-Operationally focussed
• Rarely influences staff members to behave in line with the values of the
Department
• Does not provide regular feedback to staff
• Is fully aware of how individual staff members are performing against their key
performance indicators
• Provides feedback to staff on how they are performing against their key
performance indicators
• Rarely makes each individual feel that their work is important and worthwhile
• Seldom looks for clues to improve in either successes or failures
• Consistently informs people when their performance is not up to scratch
• Is not in touch with the day-to-day operations in the office.
Component 3: -Positive energy
• Facilitates a strong sense of involvement with staff
• Demonstrates empathy with staff
• Has a positive impact on the morale within the office in which he or she is
working
• Is someone people want to work with
• Takes action to contribute to high morale
• Creates strong morale and spirit in the team
• Has a positive influence on the people that work for them.
17
Summary discussion of findings: The identification of a three factor model
of leadership
The study indicated that there were three factors that could be used to explain
leadership within this Government Department. The first of these factors was labelled
future oriented. This factor covered aspects of leadership such as the ability to have
vision, set stretching objectives, identify trends, and seek better ways of doing things.
The second factor was labelled operational focus. This factor captured behaviours
around the day-to-day management of an office and employees, such as providing
feedback and staying in touch with others’ performance. The final factor was labelled
positive energy. This factor captured aspects of a leaders charisma and their effect on
others in terms of morale and empathy. The identification of these factors and
questionnaire to measure leadership attributes enabled the Government Department to
identify leadership competencies within those already employed. A leadership
development programme was designed using these factors to assist those in leadership
roles.
This three-factor model has many similarities to other older models of leadership,
although presented in a clearly delineated fashion. For example, positive energy
captures the essence of transformational leadership. The traits of empathy and a
person’s ability to affect change in others are traits that are seen as responsible for
facilitating change in others through a transformational approach (Bass & Avolio, 1989).
Furthermore, the ability to envisage a future is a core trait in many models of leadership
from those that focus on goal attainment (Bellows, 1959) to descriptive models that
attempt to identify core leadership traits (Stogdill, 1974).
Hart (2001) noted a very similar model of leadership to that identified in the current
study. In the culmination of work across both New Zealand and Australia on
organisational stress and wellbeing, Hart found the role of leaders as crucial to
moderating stress levels in organisations. Hart in turn described leaders as having three
key competencies. These competencies are a mirror image of those described in the
current research and have been titled: providing direction (vision); attending to core
business (operational focus) and providing support to staff (positive energy). This
provides further support to the validity of the leadership model identified in the current
study.
18
The theoretical application and development of a three-factor model of leadership
The development of this statistically robust three-factor model of leadership has
numerous advantages within the Government Department. If this model of leadership
was applied and tested in other organisations the advantages gained by this
Government Department would easily translate to the respective organisation.
A three-factor model of leadership can easily be integrated into a selection procedure.
Selection requires the integration of information from a variety of sources, such as
personality tools, structured interviews, and reference checking. A common problem
with selection methodologies however is that there is an overabundance of information
and the complexity of the decision-making exceeds cognitive limitations. This is likely to
be especially the case with complex models of personality-based leadership, which can
be comprised of over 30 factors (Gough, 1984). Furthermore, development only three
structured interview questions to capture the three factors is unlikely to cause excessive
respondent burden and is well inside best practice guidelines for structured interviewing
(Taylor & O’Driscoll, 1995).
In addition to selection, the three-factor model of leadership has application in other
Human Resource initiatives such as training and performance management. A training
course structure is simplified by the necessity to only focus on three factors. This is also
the case with respect to performance appraisal where raters will often have difficulty
differentiating among performance dimensions. Raters are more likely to give valid
ratings when performance dimensions are kept to a minimum (Saskin, 1981).
The factor structure that was derived from the leadership questionnaire accounted for
over fifty percentage of the variance related to how leadership was viewed inside the
Department. This is equivalent to the LPI which accounts for 55% of the variance around
assessment of leader performance by subordinates (Posner and Kouzes, 1988).
19
Limitations of the current study
The factor structure model of the leadership should however be confirmed before it is
accepted as a robust model for describing leadership, and validated empirically against
performance data before it is used for selection. This is especially important given that
the current factor structure was derived with a smaller sample size than would be
desirable. A confirmatory factor analysis should therefore be completed on a second
sample, using the revised 24-item leadership questionnaire. The initial PCA reported
was exploratory and only via confirmatory factor analysis can it be refuted or supported.
While a three factor solution was agreed upon, the model should be seen as a macro
rather than complete model of leadership. It is likely that the global factors of leadership
herwithin reported are in turn compromised of facet constructs. Until the facets of future
oriented, operational focus, and positive energy are identified the model remains too
macro and over simplified to be considered a complete model of leadership.
The current study provided an initial examination of the psychometric properties of a
potential leadership questionnaire. In particular the PCA identified three constructs that
underpin leadership within this Department, and measures of internal reliability and inter-
item correlations were evaluated. Various psychometric properties are still, however, to
be investiagted including; external reliability; construct validity; and criterion validity.
External reliability or test-retest reliability is required to assess the stability of the results
over a time. Criterion validity which is the ability of the three-factor model to predict some
real world criterion and construct validity would need to be evaluated by correlating the
factors in the WLQ with other leadership tools such as those measuring transformational
leadership (Bass and Avolio, 1985).
The current application of the three factor model of leadership
The three-factor leadership model and questionnaire has recently been successfully
adopted by one of New Zealand’s largest Government Departments. The resulting factor
structure of the leadership questionnaire provided a clear indication for the Department
as to the key elements for leadership in their organisation. These elements (visionary
behaviour, positive focus or charisma, and results orientation) are now measured by the
organisation through the use of a 360-feeedback questionnaire that is completed on
managers to identify areas of strength and areas for development. This feedback
process is in turn the precursor to a modular training programme for enhancing
leadership. The core leadership training has therefore been built to enhance three main
competencies; visionary behaviour, positive focus, and results orientation.
20
A key advantage of the three-factor leadership model identified by the organisation is the
ability to develop a training programme that is seen to cover the major components of
leadership in an efficient manner. This is achieved through three two day training
modules each dedicated to a separate area of leadership as identified in the three-factor
model. Furthermore, while, at the commencement of the training roll out, individual
modules could be undertaken, it has subsequently been found that the modules have a
synergistic effect such that the greatest value for increased outcomes is gained from
attending all three training modules. After the opportunity for on the job application of
learning, impact is assessed through presentations on what has changed as a result of
attending the workshops. This is complemented with coaching and mentoring
opportunities made available to participants to further embed learning. The key outcome
achieved has been increased leadership among managers as identified by subsequent
feedback processes and adoption of specialised training for managers at the conclusion
of the core leadership course.
These leadership development initiatives have been linked to a new leadership
qualification that is being registered on the New Zealand Qualifications framework. This
creates another element to the development of leaders. The success of the development
programme therefore starts with a strong model and is complemented by sound
programme design, best fit of provider to organisation, and incorporating strong quality
assurance processes.
Conclusion
The development of an empirically derived three-factor model of leadership has been
described, and the model has been discussed in relation to leadership theory. The
factors identified were operational focus, vision, and positive energy. The resulting
model has potential for application in many areas of HR from selection through to
performance management, and is currently the foundation for a leadership development
programme run by a large New Zealand Government Department. It is also being
developed further for recruitment and performance appraisal. The usefulness of the
model comes from both the psychometric soundness of the three factors, as well as the
ease with which the model can be applied. While, additional work should be completed
on the model the current work indicates that it is a useful and parsimonious model of
leadership.
21
REFERENCES
Avolio, B.J., Bass, B. M. and Jung, P. I. (1999). Re-examining the components of
transformational and transactional leadership using the multifactor leadership
questionnaire. Journal of Occupational and Organisational Psychology, 72, 441-
462.
Avolio, B. J. & Gardner, W. L. (2005) Authentic leadership development: Getting to the
root of positive forms of leadership. The Leadership Quarterly, 16(3) 315-338.
Barrick, M. R., Mount, M. K. and Judge, T. A. (2001). Personality and performance at the
beginning of the new millennium: What do we know and where do we go next?
Personality and Performance. 9, 9-29.
Bass, B.M. (1985). Leadership and performance beyond expectations. New York: Free
Press.
Bass, B.M., & Avolio, B.J. (1989). Potential biases in leadership measures: How
prototypes, leniency, and general satisfaction relate to ratings of transformational
and transactional leadership constructs. Educational and Psychological
Measurement, 49, 509-547.
Bass, B.M. (1990). Bass and Stogdill’s handbook of leadership: Theory, research and
managerial applications (3
rd
ed.). New York: Free Press.
Bellows, R.M. (1959). Creative Leadership. New Jersey: Prentice-Hall.
Cattell, R. B. and Warburton, F. W. (1967). Objective personality and motivation test.
Urbana:University of Illinois Press.
Dansereau, F. Jr, Graen, G. and Hager, W. J. (1975). A vertical dyad linkage approach
to leadership within formal organisations. A longitudinal investigation of the role
making process. Organisational Behavior and Human Performance, 13, 46-78.
Fiedler, F.E. (1964). A contingency model of leader effectiveness. In L. Berowitz (Ed.),
Advances in experimental social psychology (Vol. 1). New York: McGraw-Hill.
Fiedler, F.E. (1967). A contingency model of leader effectiveness. In L. Berowitz (Ed.),
Advances in experimental social psychology (Vol. 1). New York: McGraw-Hill.
Gioia, D. A. and Sims, H. P.(1986). Cognitive behavior connections: Attribution and
verbal behavior in leader-suboridinate interactions. Organizational Behavior and
Human Decision Processes, 37(2); 197- 230.
Gough, H.G. (1984). A managerial potential scale for the California Psychological
Inventory. Journal of Applied Psychology, 69, 233-240.
22
Hart, L. B. & Waisman, C. S. (2005) The sorts of leadership. American Society for
Training and development, 59 (3), 20-22.
Hart, P.M. (2001). Occupational stress, employee well-being, organisational health and
effectiveness. Key note address of the New Zealand Psychological Society
Conference, Auckland University: Auckland.
House, R. J. (1971). A path–goal theory of leader effectiveness. Administrative Science
Quarterly, 16, 321–328.
House, R., Jarvidan, M. Hanges, P. and Dorfman, P. (2002). Understanding cultures
and implicit leadership theories across the globe: An introduction to project
GLOBE. Journal of World Business, 37, 3-10.
Kabnoff, B. (1981). A critique of Leader Match and its implications for leadership
research. Personnel Psychology, 88, 60-77.
Kelly, G. (1955).Principles of Personal Construct Psychology. New York: Norton.
Kline, P. (2000). Handbook of Psychological Testing. Routledge: London.
Kouzes, J. M. and Posner, B.Z. (1992). Ethical leaders: An essay about being in love.
Journal of Business Ethics, 11,479-485.
Landy, F.J. (1989). Psychology of work behaviour. Pacific Grove, CA: Brooks/Cole.
Latham, G.P., & Wexley, K.N. (1993). Increasing Productivity through Performance
Appraisal. Reading, MA: Addison-Wesley. In Latham, G.P., Skarlicki, D., Irvine, D.,
& Sigel, J.P. (1993). The increasing importance of performance appraisals to
employee effectiveness in organisational settings in North America. International
review of Industrial and Organisational Psychology, 8, 87 - 132.
Liden, R.C., & Graen, G.B. (1980). Generlizability of the vertical dyad linkage model of
leadership. Academy of Management Journal, 23, 451-465.
Lord, R.G., DeVader, C.L., & Alliger, G.M. (1986). A meta-analysis of the relation
between personality traits and leadership perceptions: An application of validity
generalisation procedures. Journal of Applied Psychology, 71, 402-410.
Miner, J.B. (1978). Twenty years of research on role-motivation theory of managerial
effectiveness. Personnel Psychology, 31, 739-760.
Nunnally, J. O. (1978). Psychometric theory. New York:McGraw-Hill.
Posner, B.Z. and Kouzes, J. M. (1988). Development and validation of the leadership
practices inventory. Educational and Psychological Measurement, 48, 483-496
Saal, F.E., & Knight, P.A. (1995). Industrial/Organizational Psychology: Science and
Practice (2
nd
ed.). Brooks/Cole Publishing Company. California.
23
Sashkin, M., (1981). Appraising appraisal: Ten lessons from research for practice.
Organisational Dynamics, 9, (3), 37-50.
Stogdill, R.M. (1948). Personal factors associated with leadership: A survey of the
literature. Journal of Psychology, 25, 35-71.
Stogdill, R.M. (1974). Handbook of leadership. New York: Free Press.
Taylor, P.J., & O’Driscoll, M.P. (1995). Structured Employment Interviewing. Britian:
Gower.
Vecchio, R.P. (1983). Assessing the validity of Fiedler’s contingency model of leadership
effectiveness: A closer look at Strube and Garcia. Psychological Bulletin, 93, 404-
408.
Yukl, G.A. (1989). Leadership in organisations (2
nd
ed.). Englewood Cliffs, NJ: Prentice-
Hall.
Zaccaro, S.J., Foti, R.J., & Kenny, D.A. (1991). Self-monitoring and trait-based variance
in leadership: An investigation of leaders flexibility across multiple group situations.
Journal of Applied Psychology, 76, 308-315.
