The Emerald Research Register for this journal is available at www.emeraldinsight.

com/researchregister

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0268-3946.htm

Assessing leadership styles and organisational context

Victor Dulewicz and Malcolm Higgs
Henley Management College, Greenlands, Henley-on-Thames, Oxfordshire, UK
Abstract
Purpose To investigate the new leadership dimensions questionnaire (LDQ) and a related framework for assessing an individuals leadership style in relation to the context in which the leader works; the three new LDQ sub-scales designed to measure organisational context, follower commitment and leader performance; and the relationship between personality and leadership. Design/methodology/approach Research is reported on LDQ data from a large sample of leaders and managers (n 222) from a range of public and private organisations. A style score was calculated and then related to data on respondents biographical job function, gender, sector and nationality and FFM personality data. Findings Results show a reasonably even allocation across all three leadership styles and that the styles are independent of the four important biographical variables. They also show that the ve FFM personality factors do not account for any additional variance on any of the styles at a signicant level. Results on the factor structure of the organisational context, follower commitment and leader performance scales show them to be reliable scales. Research limitations/implications A majority of the sample were from the UK, from the private sector and were male. This study did not incorporate measures of job performance or investigate the style and context link. The self-assessed, not the 3608 version of LDQ was used. Practical implications Some support is provided for the LDQs use for leadership assessment and development, and for identifying potential, in both public and private sector organisations, with a standardisation sample of more than 1,000 now available. Results also show that the LDQ can be used without losing signicant personality-related variance. Originality/value LDQ provides a unique opportunity for managers to relate leadership dimensions to three different leadership styles engaging, goal-oriented and involving and, in turn, to the degree of organisational volatility faced by the leader, thus enabling respondents to identify the most appropriate style. Leader performance and follower commitment sub-scales should facilitate further research by academics into leadership performance. Keywords Leadership, Job commitment Paper type General review

Assessing leadership styles

105
Received March 2004 Revised September 2004 Accepted September 2004

Introduction Organisations and researchers have been obsessed over the last four decades with leadership, and attempts to deconstruct the phenomenon into a universal set of measures (Kets De Vries, 1993; Goffee and Jones, 2000; Higgs, 2003; Conger and Toegel, 2002). More recently a dominant approach to studying leadership has emerged. This is based around the model of Transformational and Transactional leadership developed by Bass (1985) and operationalised by Bass and Avolio (1995). A strength of the model has been the distinction between sets of leadership behaviours required in two distinct contexts. This paper builds on the literature on Transformational leadership and explores an extended range of contexts. It presents a new framework for assessing leadership

Journal of Managerial Psychology Vol. 20 No. 2, 2005 pp. 105-123 q Emerald Group Publishing Limited 0268-3946 DOI 10.1108/02683940510579759

JMP 20,2

competencies from which style proles are drawn and then linked to the context in which the leader operates. Sub-scales are also incorporated which enable the user to review leader performance and commitment of the followers. A new model of leadership Having reviewed the development in thinking on the nature of effective leadership and, in particular, having looked at the literature from a sense making rather than discovery perspective (Weick, 1995), a pattern is beginning to emerge. From this emerging literature it is proposed that effective leaders are differentiated from other leaders through the exercise of a relatively small range of skill or competence areas (Kouznes and Posner, 1998; Goffee and Jones, 2000; Higgs and Rowland, 2001; Hogan and Hogan, 2001). The way in which these skills and competencies are exercised is not prescribed, but is the function of the underlying personality of the leader (Hogan, 2002; Hogan and Hogan, 2001). Indeed, this combination is implied by Goffee and Jones (2000) in their statement that effective leadership requires being yourself, with skill. This relatively simple statement has signicant implications for the way in which we view leadership, although it does challenge the view of some (Hogan and Hogan, 2001; Collins, 2001). As Collingwood (2001) points out, it is becoming evident that leadership is personal. The personality of the leader plays an important part in the exercise of leadership. The areas of effectiveness (the skills) need to be exercised in a way which is congruent with the underlying personality of the leader. Building on this view, it is possible to suggest a model (Higgs, 2003) that reects the research and thinking on leadership emerging from a sense making paradigm. The elements in this model are explored briey below: (1) Competence areas: . Envision the ability to identify a clear future picture, which will inform the way in which people direct their efforts and utilise their skills. . Engage nding the appropriate way for each individual to understand the vision and, hence, the way in which they can contribute. . Enable acting on a belief in the talent and potential of individuals, and creating the environment in which these can be released. . Inquire being open to real dialogue with those involved in the organisation and encouraging free and frank debate of all issues. . Develop working with people to build their capability and help them to make the envisioned contribution. (2) Personal characteristics . Authenticity being genuine and not attempting to play a role; not acting in manipulative way. . Integrity being consistent in what you say and do. . Will a drive to lead, and persistence in working towards a goal. . Self-belief a realistic evaluation of your capabilities and belief that you can achieve required goals. . Self-awareness a realistic understanding of who you are; how you feel and how others see you.

106

This model allows for the exercise of leadership in different ways and thus, for the impact of organisational context on the way in which leadership is observed. Although the model is not prescriptive, it does appear to miss an important element of leadership. The literature on leadership has consistently ignored cognitive elements for the last few decades. This is due in part to the dominance of quantitative studies and the associated tendency not to measure cognitive elements in such studies. However, recently Kets de Vries and Florent-Treacy (2002) in a qualitative study of global leaders proposed that effective leadership required a combination of behavioural, cognitive and personality factors. This suggestion does not make explicit where the leaders ability to learn sits (or indeed if it is signicant). Higgs and Rowland (2001) in a study of change leaders identied the signicance of leader learning. These two sets of ndings may suggest that a leadership framework combining cognitive, behavioural, personality and learning factors may be a useful one within which to study leadership. However, as stated earlier it is clear that there is a dynamic relationship between the leaders and the organisational context. It may be reasonable to see the organisational factors which impact on leadership comprise its strategy, culture, policies and practices and its ability to learn as an organisation (Senge et al., 2000). The relationship between leader and organisation is potentially a dynamic one. For example, different organisational strategies may require changes in leadership behaviours. However, it is also feasible that a change in leadership behaviour may lead to a different strategic approach being adopted by the organisation. This dynamic relationship is described in detail by Higgs and Dulewicz (2002). The essence is that changes in context require changes in the way in which leaders operate in the organisation. Leadership styles and context From the leadership literature there is an emerging consensus that there is no single prescription for effective performance (Goffee and Jones, 2000; Gill, 2001; Higgs, 2003; Higgs and Rowland, 2003). The relationship between the approach of leaders (or their leadership style) and the context in which they operate is seen to be important. This is by no means a new thought and is rooted in contingency theory (Fiedler, 1964; Hersey and Blanchard, 1969, 1993). However, more recently the contextualisation implied by the Transformational school (Bass and Avolio, 1996) has moved from a largely internal leader: follower focus to a broader, and often external one (Higgs and Rowland, 2003). In particular, there is an increasing focus on the efcacy of different leadership behaviours in differing contexts of change (Wheatley, 2000; Jaworski, 2001; Senge, 1997; Higgs and Rowland, 2003). Whilst a diverse range of behaviours is described within this literature, it is feasible to group them into three broad categories. (1) Goal-oriented. A set of behaviours in which the leader sets direction and behaves in a way in which he/she plays a signicant role in directing others to achieve the key goals required to attain the performance required. This is not to suggest an authoritarian approach but rather behaviours which are strongly leader-centric. (2) Involving. A somewhat less leader-centric set of behaviours. In this category the leaders focus remains on providing a strong sense of direction. However, there is a more signicant focus on involving others in both setting direction and, to a larger extent, in determining how goals will be achieved.

Assessing leadership styles

107

JMP 20,2

(3) Engaging. Leader behaviours in this category are focused on facilitating others in achieving both nature of the direction and means of achieving the necessary goals. The leader is more concerned with developing the capability of others to achieve than with the close direction of the enterprise. Table I illustrates the way in which the research and writing of some of the authors working with this contextualised view of leadership may be mapped onto these three categories. Turning to context, and in particular the context of change, the Transformational school (Bass and Avolio, 1996; Alimo-Metcalfe and Alban-Metcalfe, 2001) posit two distinct contexts: transformational change and steady-state. However, within the change literature it is evident that there is a diverse range of change contexts (Buchanan et al., 1999; Beer and Nohria, 2000; Higgs and Rowland, 2003). Reviewing this range it is feasible to identify the following key contextual variables (Higgs and Rowland, 2003): (1) magnitude and scale of change; (2) source of change (i.e. internal or external); (3) impact of the change on those in the business; and (4) timescales and speed of change Overall, it is perhaps useful to consider the change context in terms of volatility and complexity and to see it as a continuum rather than, as is done by the Transformational school, a bipolar variable. The relationship between leadership behaviours and change context is evident in the Transformational school. However, until recently little empirical work had been conducted which related leadership behaviours to a more elaborate operationalisation of change context. Higgs and Rowland (2003) report the results of a study of over seventy change stories from ten organisations in which leadership behaviour categories were examined for effectiveness in a broad range of change contexts. They found that as complexity of the context increased, a more facilitative style of leadership became necessary for success. A leader-centric or directive style was found to be inappropriate and ineffective in such context. However, such a style was found to be more common (and indeed dominant) in relatively simple and straightforward contexts. From the above it is feasible to suggest a relationship between leadership style and context along the lines presented in Table II. Follower commitment and leader performance The term organisational commitment (OC) covers a range of affective factors concerning followers attachment, loyalty, involvement and identication with their organisation and also some cognitive factors relating to, for example, understanding
Hersey and Goffee and Bass and Higgs and Blanchard Wheatley Jaworski Fiedler Jones (2000) Avolio (1996) Rowland (2003) (1969; 1993) (2000) (2001) (1964) _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

108

Style

Table I. Mapping the three styles on the relevant literature

Goal Oriented Involving Engaging

change and calculating benets to self. Indeed, research studies have found there to be two important and fairly independent aspects of OC, an attitudinal/affective and a rational/calculative type (Cook and Wall, 1980; Mathieu and Zajac, 1990). Cacioppe (1997) found that quality of work experience is related to the attitudinal/affective type and so interacting with effective leaders is likely to contribute to followers OC. Mathieu and Zajac (1990), from a meta-analysis, showed that various leader relations with followers such as leader consideration, communication and participative behaviour were antecedents of OC. Bass and Avolios (1995) MLQ questionnaire includes three scales designed to measure outcomes, specically followers: . extra effort; . leader satisfaction; and . leader effectiveness. Since MLQ is a 3608 instrument, data collected includes not only self-assessments but also assessments by others, especially subordinates. Research on these has shown that transformational leadership is more highly correlated with these outcome variables than transactional. Furthermore, management by exception and Laissez-Faire behaviour has been found to be negatively related or not related (Bass, 1985). Kaipiainens (2004) results conrmed that transformational leadership produces signicantly greater OC amongst followers than transactional and also leads to signicantly higher job performance of leaders. Next, this paper will describe how research ndings into emotional intelligence and senior management (leaders) competencies has been developed into a framework for assessing an individuals leadership style within the organisational context, which concurs with the conclusions drawn from the brief review of the leadership literature above. Links between the leadership literature and EQ, IQ and MQ competencies In looking more broadly at leadership and, in particular, the future nature of leadership, a number of authors have identied the growing signicance of emotional intelligence (Cacioppe, 1997; Sosik and Magerian, 1999; Chaudry, 2001). In part, this shift in focus from the rational to emotional aspects of leadership represents the continuation of a trend encountered more broadly in thinking on organisational behaviour and leadership (Fineman, 1997; Goffee and Jones, 2000). Indeed, although not explicitly surfaced, much of the literature on transformational leadership implies that leaders require emotional intelligence (Higgs and Rowland, 2001). The authors review of the leadership literature (above) from the Transformational period onwards focused on models that contain clearly dened,

Assessing leadership styles

109

Leadership styles Goal oriented Involving Engaging

Low change _ (_)

Change context Moderate change (_) _ (_)

High change (_) _

Notes: _ good t; ( _ ) a degree of t

Table II. Matching leadership style and change context

JMP 20,2

110

behavioural constructs. On the basis of a content analysis of these constructs, there appears to be strong indications of a linkage between leadership and emotional intelligence. Dulewicz and Higgs (2003) provide a map of some of the key leadership models and their potential relationship to the elements of emotional intelligence as dened by Higgs and Dulewicz (2002). The key themes were propounded by eminent authorities on the subject of leadership reviewed above (Bass and Avolio, 1995; Alimo-Metcalf and Alban-Metcalf, 2001; Goffee and Jones, 2000; Kouznes and Posner, 1998; Kotter, 1990; Bennis, 1998; and Goleman et al., 2002). Their results show that the large majority of cells in the matrix are lled by at least one construct. The EQ element self-awareness appears to be least well covered, but still gains support from three authors. Dulewicz and Higgs (2003) also conducted a similar mapping exercise on the key themes propounded by same eminent authorities onto the IQ and MQ dimensions required for effective leadership. Their results show that the large majority of cells in the matrix are lled by at least one construct. The Bass and Avolio (1995) and Alimo-Metcalfe and Alban-Metcalf (2001) models provide support for all eight IQ and MQ dimensions while the lowest support is provided by the Goffee and Jones (2000) model. However, the other ve models have fewer constructs and so inevitably provide less wide coverage of the domain. Nevertheless, they all provide links with at least ve of the eight IQ and MQ dimensions. Critical analysis and judgement appears to be least well covered but nevertheless gains support from three authors. Finally, Dulewicz and Higgs (2003) linked the 15 dimensions measured by the LDQ to the components of the emerging model of leadership, embracing both competencies and personal characteristics. They found that each component was linked to at least one LDQ dimension and that each dimension is linked to at least one component of the emerging model apart from critical analysis and judgement. The results of these three mapping exercises provide strong evidence to support the content validity of the LDQ. In the next section, the dimensions will be described and data on the reliability and criterion, content and construct validity of the LDQ instrument will be presented. The leadership dimensions questionnaire Two pilot studies were conducted to rene the questionnaire using item analysis, as described by Dulewicz and Higgs (2003). The nal version of the LDQ contains 70 EQ items and 80 IQ and MQ items, 10 per scale, and was rened from the original 175 trial items. The nal version of the LDQ contains 15 scales, titles and short denitions of which appear in Table III. Scale scores and biographical data Dulewicz and Higgs (2003) analysed scores on the 15 dimensions in relation to a number of respondents biographical variables: age; level of qualication, gender, function, sector and qualications attained. Taking all of their results into account, they concluded that scores on most, if not all of the 15 LDQ dimensions appear to be largely independent of the gender, job function, qualication level and sector of their respondents. However, there were a few exceptions: males had higher scores on the critical analysis and judgement scale; general managers had higher scores, and those with degrees or professional qualications had lower scores, on self-awareness; and age was signicantly negatively correlated with four dimensions out of 15.

Intellectual dimensions (IQ) (A) Critical analysis and judgement

(B) Vision and imagination

(C) Strategic perspective.

A critical faculty that probes the facts, identies advantages and disadvantages and discerns the shortcomings of ideas and proposals. Makes sound judgments and decisions based on reasonable assumptions and factual information, aware of the impact of any assumptions made Imaginative and innovative in all aspects of ones work. Establishes sound priorities for future work. Clear vision of the future direction of the organisation to meet business imperatives. Foresees the impact of changes on ones vision that reect implementation issues and business realities Sees the wider issues and broader implications. Explores wide range of relationships, balances short- and long-term considerations. Sensitive to the impact of ones actions and decisions across the organisation. Identies opportunities and threats. Sensitive to stakeholders needs and the implications of external factors on decisions and actions Plans ahead, organises all resources and coordinates them efciently and effectively. Establishes clear objectives. Converts long-term goals into action plans. Monitors and evaluates staffs work regularly and effectively, gives sensitive, honest feedback A lively and enthusiastic communicator, engages others and wins support. Clearly communicates instructions and vision to staff. Communications are tailored to the audiences interests and focused. Communication style inspires staff and audiences, conveys approachability and accessibility Gives staff autonomy, encourages them to take on personally challenging demanding tasks. Encourages them to solve problems, produce innovative ideas and proposals and develop their vision and a broader vision. Encourages a critical faculty and a broad perspective, and encourages the challenging of existing practices, assumptions and policies Believes others have potential to take on ever more-demanding tasks and roles, encourages them to do so. Ensures direct reports have adequate support. Develops their competencies, and invests time and effort in coaching them so they contribute effectively and develop themselves. Identies new tasks and roles to develop others. Believes that critical feedback and challenge are important Willing to make decisions involving signicant risk to gain an advantage. Decisions are based on core business issues and their likely impact on success. Selects and exploits activities that result in the greatest benets to the organisation and its performance. Unwavering determination to achieve objectives and implement decisions (continued )

Assessing leadership styles

111

Managerial dimensions (MQ) (D) Resource management

(K) Engaging communication

(L) Empowering

(M) Developing

(P) Achieving

Table III. Denitions of the scales of the leadership dimensions questionnaire

JMP 20,2

Emotional and social dimensions (EQ) (E) Self-awareness

112

(F) Emotional resilience

(G) Intuitiveness

(H) Interpersonal sensitivity

(J) Inuence

(N) Motivation

(Q) Conscientiousness

Table III.

Awareness of ones own feelings and the capability to recognise and manage these in a way that one feels that one can control. A degree of self-belief in ones capability to manage ones emotions and to control their impact in a work environment Performs consistently in a range of situations under pressure and adapts behaviour appropriately. Balances the needs of the situation and task with the needs and concerns of the individuals involved. Retains focus on a course of action or need for results in the face of personal challenge or criticism Arrives at clear decisions and drives their implementation when presented with incomplete or ambiguous information using both rational and emotional or intuitive perceptions of key issues and implications Is aware of, and takes account of, the needs and perceptions of others in arriving at decisions and proposing solutions to problems and challenges. Builds from this awareness and achieves the commitment of others to decisions and action. A willingness to keep open ones thoughts on possible solutions to problems and to actively listen to, and reect on, the reactions and inputs from others Persuades others to change views based on an understanding of their position and a recognition of the need to listen to this perspective and provide a rationale for change Drive and energy to achieve clear results and make an impact. Balances short- and long-term goals with a capability to pursue demanding goals in the face of rejection or questioning Displays clear commitment to a course of action in the face of challenge and to match words and deeds in encouraging others to support the chosen direction. Shows personal commitment to pursuing an ethical solution to a difcult business issue or problem

Reliability The results of a reliability analysis conducted on the 15 dimensions of LDQ are published by Dulewicz and Higgs (2003, 2004). All 15 LDQ scale co-efcient reached an acceptable level, being above 0.7 (Nunnally, 1978), with vision and imagination, engaging communication, managing resources and developing being particularly highly reliable. The relatively weakest dimensions were critical analysis and judgement, empowering and achieving but they still reached an acceptable level. Validity Three main types of validity of the LDQ construct, content and criterion (concurrent) will be summarised. Dulewicz and Higgs (2003) claim that Content validity of LDQ was derived from the rigour with which all facets of leadership were reviewed and mapped onto the 15 dimensions. Furthermore, they reference the extensive

literature on personal competences of senior managers in leadership roles. Clear links were drawn with items in manager and director personal competences surveys that appeared to be related to leadership. This work enabled them to write LDQ items based upon a comprehensive set of constructs considered by many leading authors in the eld to relate to leadership requirements, and then in turn to link these to personal competences. Construct validity deals with how well, or to what degree, the test measures the target trait or construct. Correlation co-efcients between LDQ dimensions and 16PF (Cattell et al., 1970) personality factors are presented by Dulewicz and Higgs (2003). They state that one would expect some personality factors to be related to some LDQ dimensions, specically EQ and MQ dimensions, but not others (IQ), because they are closer to personality-type constructs. The authors tested out a number of hypothesised relationships and found that most of the LDQ dimensions related to interpersonal behaviour and emotional adjustment behaviours had construct validity (Dulewicz and Higgs, 2003). Young and Dulewicz (2003) present correlations between the 15 LDQ dimensions and scores on the OPQ personality questionnaire (Saville et al., 1993). They hypothesised areas of convergence and divergence between the two sets of scales and found a high proportion of their predictions were borne out. In a recent study of the convergent and divergent construct validity of the eight LDQ dimensions derived from the social and emotional (EQ) competencies, Kaipiainen (2004) found evidence of convergent validity against scales of social intelligence, empathy and leadership derived from Basss MLQ. She found divergent validity with Machiavellianism and Tyrannical Leadership scales. Criterion (Concurrent) validity refers to the degree to which test scores can predict job performance. Young and Dulewicz (2003) present correlations between the 15 LDQ dimensions and job performance ratings on 261 Naval Ofcers from the formal appraisal system. They found that 11 of the 15 dimensions of LDQ (73 per cent) were signicantly related to current job performance. Kaipiainens (2004) results also showed signicant correlations between the EQ dimensions of LDQ and job performance, particularly with EQ assessments from peers and subordinates. Furthermore, she showed that EQ dimensions added incremental validity over transformational leadership from the MLQ, using hierarchical regression analysis with job performance as the dependent variable. Research into leadership styles and context Further analyses were conducted on leadership style and organisational context using the nal version of LDQ, on data from the combined sample from the two pilot studies whose biographical details are presented below. Sample The sample consisted of 222 managers and ofcers whose biographical details are summarised here. The average age of the subjects was 38.3 years (SD 7.3); and time spent in current job was 31.3 months (SD 23.9). A large majority of the sample was male (70.2 per cent) and worked in the private sector (81.8 per cent). Looking at the functions in which they worked, 32 per cent were general managers, 18 per cent were in sales/marketing, 15 per cent in nance, 10 per cent in technical and the remainder in

Assessing leadership styles

113

JMP 20,2

other functions. As regards their highest qualications, 37 per cent had a rst degree and a further 48 per cent had higher degrees and/or professional qualications. Turning nally to the nationality of the respondents, 58 per cent were from the UK, 22 per cent were from the rest of Europe and the remaining 20 per cent were from all other parts of the world. The three leadership styles The authors contend that effective leadership is increasingly being seen in terms of a combination of: . personal characteristics which are required to enable an individual to engage in a leadership role in an effective manner; . a range of skills and behaviours which need to be in place to provide effective leadership; . a range of styles related to the context in which leadership is exercised; and . a range of ways in which the leadership behaviours may be exercised in a way that matches the personal style of the individual leader. In addition, it is quite widely accepted that leadership may be exhibited at many levels in an organisation (Alimo-Metcalfe, 1995). Based on their review of the literature covering different leader behaviours in differing contexts of change (see Introduction above), the authors identied three distinct leadership styles. . Engaging leadership. A style based on a high level of empowerment and involvement appropriate in a highly transformational context. Such a style is focused on producing radical change with high levels of engagement and commitment. . Involving leadership. A style that is based on a transitional organisation which faces signicant, but not necessarily radical changes in its business model or modus operandi. . Goal leadership. A style that is focused on delivering results within a relatively stable context. This is a Leader-led style aligned to a stable organisation delivering clearly understood results. The prole for each style, based upon the range (high, medium, or low) of scores obtained on the 15 LDQ dimensions, is presented in Table IV. These proles were developed from a content analysis of the literature on leadership. Initially, this focused on the transformational and transactional behaviours which were context-based (Bass, 1985, 1990; Bass and Avolio, 1995) and subsequently expanded to encompass the change leadership and broader change literature. The engaging style was informed by authors working in the transformational and change leadership elds (Bass, 1995; Higgs and Rowland, 2003; Kouznes and Posner, 1998). The traditional and the transactional leadership literature informed the development of the goal-oriented style (Bass, 1995; Kotter, 1990; Hersey and Blanchard, 1993). The involving style was again informed by both more traditional and change leadership literature (Bennis, 1998; Chaudry, 2001; Gill, 2001; Hersey and Blanchard, 1993; Higgs and Rowland, 2003).

114

LDQ dimension Critical analysis and judgement Vision and imagination Strategic perspective Engaging communication Managing resources Empowering Developing Achieving Self-awareness Emotional resilence Motivation Interpersonal sensitivity Inuencing Intuitiveness Conscientiousness

Low E G

Medium EI E EI GI I I GI EI G GI G GI

High G GI G E G E E G EI GEI GEI E EI E GEI

Assessing leadership styles

115

Note: Proles of the three leadership styles showing their typical scores on the leadership dimensions questionnaire

Table IV. Three leadership style proles: goal oriented (G), engaging (E) and involving (I)

Results of study 1 (styles) The difference measure (d) To assess the degree of t between an individuals prole and the designated prole for each leadership style, it was necessary to calculate difference scores. Wohlers and Londons (1989) d measure, widely used in this eld for prole comparisons, was computed for each individual for each of the three styles, using their formula. To calculate each persons d score, dimension scores were calculated, based on the per centage of the maximum possible raw score (50). Three ranges were chosen, to denote high, medium and low scores. The benchmarks used for the style ranges were based on a review of the frequency distributions for all 15 scales. In order to obtain maximum separation, the benchmark selected for the low range was 60 per cent, and for the high range, 80 per cent. The benchmark selected for the medium range was the midpoint between the other two, i.e. 70 per cent. Descriptive statistics for the d measures appear in Table V. Although the means, SDs and ranges for all three scales are similar, the variables were all standardized so
Goal-oriented Mean Std. deviation Range Minimum Maximum 13.2 2.8 15.4 7.0 22.4 Involving 12.9 2.9 15.6 5.7 21.4 Engaging 13.4 3.1 18.3 4.9 23.2 Missing 1 Sub-total 141 Poor t 81 36 Total 222 Total 222 Table V. Descriptive statistics for d scores for the three leadership styles

N 221 221 221 N tted to each style: (d z-score , 0.00, bottom 50 per cent) Lowest d z-score Frequency 44 39 58 Per cent 31 28 41

JMP 20,2

116

that exactly comparable scores for each style could be used to assess the degree of t for each individual, and between individuals. Taking the individuals lowest d z-score as their current style, would have meant that some individuals were not closely tted to any of the styles and so the analysis focussed on individuals with d z-scores below 0.00 (i.e. the lowest 50 per cent). Data on these individuals (n 141) are presented in Table V and show the following breakdown: 31 per cent are goal-oriented, 28 per cent are involving; but the most common style is engaging with 41 per cent. The subsequent analyses reported below are based on this sub-sample of individuals closely tted to at least one style. Styles and biographical data The styles best t results just mentioned were analysed to determine whether there were any statistically signicant differences on biographical variables, using the Chi-square test. A comparison of current appropriate styles by gender showed no differences between males and females (Chi square 2:05; df 2; Not sig.). Furthermore, no differences were found when comparing styles according to the sector in which the manager works Private and Public/Not for Prot which were combined to ensure adequate cell-size (Chi square 2:04; df 2; Not sig.). Turning to job function and nationality, once again no differences were found between the three styles. Chi-square analysis was also conducted across four job categories General Management, Marketing, Finance and Technical (All others were combined to ensure adequate cell-size; Chi square 5:96; df 2; Not sig.). A comparison between managers from the UK and from all other countries combined (the numbers from many other countries was too small to treat them separately) produced a Chi square 3:38; df 2; Not sig. So, overall, there were no differences between styles on these four important biographical variables. LDQ dimensions, styles and personality An important issue arising from the literature is the relationship between leadership and personality. In order to explore this, the relationships between respondents LDQ dimensions, three leadership styles and personality characteristics from the 16PF questionnaire (Cattell et al., 1970) were investigated. The rst-order factors were mapped onto the ve-factor model (FFM) according to the classication framework developed by Salgado (2003) for his meta-analysis of the FFM. From a comparison of the respective denitions, the LDQ intellectual IQ competencies and managerial MQ competencies do not appear to overlap with the FFM personality factors while; in contrast, some of the social and emotional EQ competencies do. In order to explore this area, correlations were conducted between four pairings for which relationships were hypothesised: emotional resilience and FFM emotional stability (r 0:62; n 107; sig. at 0.01 level); inuence and FFM agreeableness ( r 0:12; n 107; not sig.); interpersonal sensitivity and FFM agreeableness (r 20:09; n 107; not sig.) and conscientiousness and FFM conscientiousness (r 0:51; n 107; sig. at 0.01 level). Therefore, two FFM personality factors are closely related to LDQ dimensions and two are not. Initially, stepwise multiple regression was used to explore the relationships between personality and the three leadership styles. The model produced for the engaging style showed that only one FFM factor, extraversion, contributed to the variance (R 0:221; R2 0:049; standardised b 2 0.221; sig. at 0.02 level), only 4.9 per cent of which was

explained. The relationships between the FFM factors and the other two styles were so weak that stepwise models could not be produced. In order to determine whether the ve FFM personality factors contributed signicant additional variance over and above the 15 LDQ dimensions, hierarchical regression analyses were conducted on each of the three styles. The results, which appear in Table VI, show that the ve FFM personality factors do not account for any additional variance on any of the styles at a signicant level. Therefore, personality does not seem to add signicantly to the leadership competencies which dene the three LDQ styles. Results of study 2 As noted in the Introduction, the context within which leaders operate is a major factor mediating their performance. From the literature review, the authors concluded that the different styles, matched to the degree of contextual volatility, would be important in determining both appropriateness and effectiveness. Therefore, an organisational context scale was designed (to become section II of the LDQ) to examine the degree and nature of change and volatility in their working environment that respondents perceive they face in their role as a leader. The pilot scale was developed from a review of the change literature (Higgs and Rowland, 2003) and contained 27 items relating to a wide range of contextual variables faced by leaders within organisations. After item-analysis involving part-whole correlations, six items which did not correlate signicantly with the overall scale score were dropped. The nal context scale was thus reduced to 21 items, relating to various aspects of change being faced by the respondent. In order to obtain a better understanding of the nature and structure of this new scale, a factor analysis was conducted. It showed (Table VII) that the scale is made up of ve separate components: a general fundamental need to change, fundamental change of the organisation/business, the need for followers to change, specic pressures from the business environment; and an unstable context. Cronbach reliabilty analysis showed that the overall scale has high reliability (a 0:9). In order to enable users to relate scale scores to leadership style, a frequency distribution was produced and score ranges identied for three equal-sized groups, reecting the degree of change perceived. Scores of 58 or below were selected to reect relative stability; 59-73 signicant change; and 74 and above radical, transformational change.
Change statistics F change df1 df2 4.330 0.677 6.556 0.772 7.267 1.443 15 5 15 5 15 5 91 86 91 86 91 86

Assessing leadership styles

117

Model

R2 0.416 0.439 0.519 0.540 0.545 0.580

R2 change 0.416 0.022 0.519 0.021 0.545 0.035

Sig. F change 0.00 0.64 0.00 0.57 0.00 0.22 Table VI. Hierarchical regression on three LDQ styles (DVs) and LDQ dimensions plus FFM personality scores as IVs

DV goal oriented style LDQ 0.645 LDQ FFM 0.662 DV involving style LDQ 0.721 LDQ FFM 0.735 DV engaging style LDQ 0.738 LDQ FFM 0.762

JMP 20,2

Item C8 C23 C2 C36 C3 C13 C9 C7 C12 C22 C19 C6 C26 C27 C16 C31 C28 C11 C18 C14 C1

1 0.88 0.87 0.76 0.72 0.59 0.58 0.56 0.46 0.42 0.22 0.22 0.21 0.20 0.20 0.16 0.15 0.14 0.12 0.11 0.11 2 0.27

2 0.13 0.06 0.25 2 0.06 0.33 0.40 0.06 0.38 0.67 2 0.31 0.40 0.75 0.16 0.58 2 0.04 0.07 0.25 0.49 0.11 0.00 0.63

Component 3 0.09 0.15 0.20 0.20 0.16 0.03 0.33 0.12 0.11 0.58 0.57 2 0.04 0.19 0.29 0.69 0.77 0.71 2 0.21 0.25 2 0.04 0.17

4 2 0.04 0.16 0.07 0.14 2 0.06 2 0.01 0.33 2 0.12 0.19 0.12 2 0.03 2 0.24 0.08 0.29 0.13 2 0.18 0.23 0.37 0.67 0.61 0.47

5 0.01 2 0.01 0.16 0.14 0.40 0.34 0.13 2 0.01 2 0.11 0.43 0.26 0.14 0.80 0.05 0.07 0.00 2 0.07 0.38 2 0.36 0.21 0.12

118

Table VII. Rotated component matrix on LDQ change scale (21 items)

Leader performance and follower commitment As noted in the introduction above, some leadership questionnaires contain integral leader performance and follower commitment scales (Bass and Avolio, 1995). In order to facilitate further research into LDQ, it was decided to include two such scales in the second part of the LDQ. The rst scale provides a self-assessment of leadership performance. It contains six items, covering followers effort, capability and exibility and overall team performance and impact. A factor analysis revealed two components, broadly reecting followers individual contributions and team output, respectively. Detailed results appear in Table VIII. Further analysis showed the overall scale to have acceptable reliability (Nunnally, 1978), with a 0:7. It was stated in the Introduction that the term OC covers a range of affective factors concerning followers attachment, loyalty, involvement and identication with their organisation and also some cognitive factors relating to, for example, understanding
Performance items (Rotated component matrix) Component Item 1 2 P1 P2 P3 P5 P6 P4 0.415 0.020 0.283 0.124 0.882 0.895 0.186 0.862 0.510 0.743 0.114 0.053 Commitment items (Component matrix) Component 1 0.761 0.612 0.696 0.644 0.654

Item Com Com Com Com Com 1 2 3 4 5

Table VIII. Component matrices for performance and commitment scales

change and calculating benets to self. Indeed, research studies have found there to be two important and fairly independent aspects of OC, an attitudinal/affective and a rational/calculative type (Cook and Wall, 1980; Mathieu and Zajac, 1990). Therefore, an OC scale was designed to assess the degree of commitment that followers show to the organisation and to the team in which they work, covering job satisfaction, realism, commitment to requisite change and to the organisation and understanding the need for change. It contains ve items and only one component emerged from a factor analysis (Table VIII). The overall scale also shows acceptable reliability (a 0:7). Discussion According to its authors (Dulewicz and Higgs, 2003), the nal version of the LDQ has beneted from rigorous trials involving two pilot studies. Their overall aim of having 10 items per scale has given broad coverage of the complex dimensions measured by the LDQ and has at the same time produced respectable a reliability coefcients. Its authors also demonstrated construct validity, nding a large number of signicant correlations between LDQ and personality characteristics (from 16PF), with the general nding that managers higher on relevant dimensions generally tend to be more extraverted and emotionally well adjusted, and specically to have greater strategic perspective and conscientiousness. In addition, Young and Dulewicz (2003) showed construct and criterion validity from their research based on personality data (from the OPQ) and job performance ratings of Naval ofcers. One important feature of the LDQ is its ability to relate proles of the scores across the 15 dimensions to three different leadership styles. The relevance of each one is dependent on the context within which leadership is exercised. Analyses of leadership styles, using a reasonably close t score (capturing 64 per cent of the total sample of 222), showed a reasonably even allocation across all three styles. Furthermore, data in this paper has shown that the styles are independent of four important biographical variables the breakdown did not vary according to the gender, sector, function, or nationality of the respondent. Turning to personality, the possible overlap between the social and emotional EQ competencies from LDQ and the FFM personality factors were explored. Whilst correlations between inuence and FFM agreeableness and between interpersonal sensitivity and FFM agreeableness were not signicant, the relationships between emotional resilience and FFM emotional stability, and conscientiousness and FFM conscientiousness were highly statistically signicant. However, Salgado (2003) reports in his review of meta-analyses of the FFM that only these two factors have been shown to predict job performance in a large number of studies. Therefore, any such overlap should help to improve the criterion validity of the LDQ. Of the three styles, only the extraversion FFM personality factor accounted for additional style variance, and only to one style engaging. Furthermore, the FFM personality characteristics did not add any incremental variance over and above the LDQ dimensions. Moreover, a concurrent study by Young (2004), which investigated the link between FFM personality factors and leadership performance, found that personality does not explain any additional variance over the LDQ dimensions whereas the latter do explain extra variance compared to personality factors alone. From these results it would appear that personality is not a moderator variable between leadership competencies on the one hand and leadership style and

Assessing leadership styles

119

JMP 20,2

120

performance on the other. This does tend to challenge the personality view put forward by some authors (Hogan and Hogan, 2001). Leaders with different personality proles appear to be able to adopt similar styles and perform at different levels. This lends credence to the propositions presented in the emerging literature (Goffee and Jones, 2000). Nevertheless, further research in this area would be desirable. The literature strongly suggests that the situation or context is highly relevant to leadership style. In addition, many assert that leadership nowadays is largely about dealing with, and leading, change and that therefore, the efcacy of different leader behaviours in different change contexts needs to be assessed. The organisational context scale has therefore been designed to cover these requirements and has been shown to be a reliable scale, covering ve different aspects of leadership. The ve factors measured by the scale correspond to the relevant contextual change variables magnitude, sources, impact on followers, and timescale/speed outlined by Higgs and Rowland (2003) and noted above. Further work should focus on interactions between style and context, and links to leader performance. One way of measuring leader performance is by self-assessment. To facilitate this, the LDQ includes a six-item scale that, after item analysis, has been shown to have acceptable reliability. The scale covers both the performance of individual followers and the overall output of the group being led. A 3608 version of LDQ has been designed which will also provide performance assessment of the leader by the followers. However, assessments of performance through the organisations appraisal system are to be encouraged and current work is underway to investigate the relationship between self, follower and formal (boss) appraisals. Another angle for assessing leader performance is via the commitment and job satisfaction of the followers. Therefore, a commitment scale has also been built into the LDQ. This has been shown to be reliable and covers commitment to change and to the organisation, as well as job satisfaction, realism and commitment to change. Further research should investigate the interactions between follower commitment and leader performance via self and 3608 appraisal, and also the dimensions of leadership which are most closely linked to these various aspects of commitment. Three recent studies have included research into the relationship between job performance and the seven EQ dimensions of LDQ. A study of team leaders (Dulewicz and Higgs, 2000) provided clear evidence of a link with measures of current performance on six of the dimensions self-awareness, emotional resilience, motivation, inuence, intuitiveness and conscientiousness, using both self and 3608 appraisal. Dulewicz et al. (2003) reported the results of a study on middle-managers which investigated the degree to which EQ dimensions are related to performance. Signicant relationships were found with current job performance on three dimensions self-awareness, emotional resilience and motivation. Kaipiainens (2004) recent study, noted above, included the EQ dimensions and her results conrmed signicant relationships between these scales and commitment amongst followers, particularly aspects of loyalty and identity; and scores on MLQ (Bass and Avolio, 1995) outcome scales relating to followers extra effort, leader satisfaction and leader effectiveness. Current research studies underway will show to what extent these two subscales provide support for the validity of LDQ and the extent to which style-context match is relevant to both performance and follower commitment.

Further studies are planned or underway on a wide variety of leaders from a number of different countries. Correlations with other leadership questionnaires are also planned, to produce further data on the construct validity of the LDQ, to broaden understanding of the 15 dimensions. Replication of other studies that have incorporated measures of context, performance, commitment and job satisfaction as criterion measures will also provide further relevant information on validity. Finally, work is currently underway to investigate the link between style and context by analysing job performance of different style ts. Conclusion In this paper, the technical properties of the LDQ have been summarised. The results of this study provide some support for its use for leadership assessment and development, and for identifying potential, in both public and private sector organisations. These results suggest that the selection of leaders could become more accurate and suggested development actions contained in the LDQ report more focused and relevant. Furthermore, standardisation data are available, based on a sample of more than 1,000 managers and senior ofcers, with 50 per cent from the public sector and 24 per cent female. The norms produced constitute a comprehensive sample of middle and senior managers from around the world (Dulewicz and Higgs, 2003). A unique feature of the LDQ is the opportunity it provides to relate proles of the scores across the 15 dimensions to three different leadership styles and in turn to the degree of organisational volatility faced by the leader. The relevance of each style is dependent on the context within which leadership is exercised. The context scale in part II of the questionnaire provides a reliable measure to help respondents identify the style that is most appropriate for their role in the current organisational context. New sub-scales designed to measure leader performance and follower commitment have also been produced to facilitate further research into the LDQ questionnaire.
References Alimo-Metcalfe, B. (1995), An investigation of female and male constructs of leadership, Women in Management Review, Vol. 10 No. 2, pp. 3-8. Alimo-Metcalfe, B. and Alban-Metcalfe, R.J. (2001), Development of the TLQ, Journal of Occupational & Organizational Psychology, Vol. 74 No. 1, pp. 1-24. Bass, B.M. (1985), Leadership and Performance Beyond Expectations, Free Press, New York, NY. Bass, B.M. (1990), Bass and Stodghill Handbook of Leadership: Theory, Research and Applications, Free Press, New York, NY. Bass, B.M. and Avolio, B. (1995), The Multifactor Leadership Questionnaire, Mind Garden, Palo Alto, CA. Bass, B.M. and Avolio, B.J. (1996), Postscripts: Recent Developments for Improving Organisational Effectiveness, Sage, London. Beer, M. and Nohria, N. (2000), Breaking the Code of Change, Harvard Business School Press, Boston, MA. Bennis, W. (1998), On Becoming a Leader, Hutchinson, London. Buchanan, D., Claydon, T. and Doyle, M. (1999), Organisation development and change: the legacy of the nineties, Human Resource Management Journal;, Vol. 9 No. 2, pp. 20-37. Cacioppe, R. (1997), Leadership moment by moment!, Leadership & Organization Development Journal, Vol. 18 No. 7, pp. 335-46.

Assessing leadership styles

121

JMP 20,2

Cattell, R.B., Eber, H.W. and Tatsuoka, M.M. (1970), Handbook for the 16PF, IPAT, IL. Chaudry, S. (2001), Management 21C, Pearson Education, London. Collingwood, H. (2001), Personal histories, Harvard Business Review, December, pp. 27-38. Collins, J. (2001), Level 5 leadership. the triumph of humility and erce resolve, Harvard Business Review, pp. 67-76, January-February.

122

Conger, J. and Toegel, G. (2002), A story of missed opportunities: qualitative methods for leadership research and practice, in Parry, K.W. and Meindl, J.R. (Eds), Grounding Leadership Theory and Research, Information Age Publishing, Greenwich, CT, pp. 175-98. Cook, J. and Wall, T. (1980), New work attitude measures of trust, organisational commitment and need non-fullment, Journal of Occupational Psychology, Vol. 53, pp. 39-52. Dulewicz, V. and Higgs, M.J. (2000), 360-degree assessment of emotional intelligence: a study, Selection & Development Review, Vol. 16 No. 3. Dulewicz, V. and Higgs, M.J. (2003), Design of a new instrument to assess leadership dimensions and styles. Henley Working Paper 0311, Henley Management College, Henley, available at: www:henleymc.ac.uk Dulewicz, V. and Higgs, M.J. (2004), A new instrument to assess leadership dimensions and styles, Selection & Development Review, Vol. 20 No. 2, pp. 7-12. Dulewicz, V., Higgs, M.J. and Slaski, M. (2003), Emotional intelligence: construct and concurrent validity, Journal of Managerial Psychology, Vol. 5, p. 18. Fiedler, F. (1964), A contingency model of leadership effectiveness, in Berkowicz, L. (Ed.), Advances in Experimental and Social Psychology, Academic Press, New York, NY. Fineman, S. (1997), Emotion and management learning, Management Learning, Vol. 28 No. 1, pp. 13-25. Gill, R. (2001), Towards an integrated theory of leadership, paper presented at the EIASM Leadership Conference, Oxford, December 2002. Goffee, R. and Jones, G. (2000), Why should anyone be led by you?, Harvard Business Review, September-October, pp. 63-70. Goleman, D., Boyatzis, R. and McKee, A. (2002), The New Leaders, Harvard Business School Press, Boston, MA. Hersey, P. and Blanchard, K.H. (1969), Life cycle theory of leadership, Training & Development Journal, Vol. 23 No. 5, pp. 26-34. Hersey, P. and Blanchard, K.H. (1993), Management of Organisational Behaviour: Utilising HR, Prentice-Hall, Englewood Cliffs, NJ. Higgs, M.J. (2003), Developments in leadership thinking, Leadership & Organization Development Journal, Vol. 24 No. 5, pp. 273-84. Higgs, M.J. and Dulewicz, V. (2002), Making Sense of Emotional Intelligence, 2nd ed., NFER-Nelson, Windsor. Higgs, M.J. and Rowland, D. (2001), Developing change leaders: assessing the impact of a development programme, Change Management Journal, Vol. 2 No. 1. Higgs, M.J. and Rowland, D. (2003), Is change changing? An examination of approaches to change and its leadership, Henley Working Paper 0313, Henley Management College, Henley, available at: www: henleymc.ac.uk Hogan, R. (2002), Leadership: what do we know?, presentation for Leadership Development Centre, Wellington.

Hogan, R. and Hogan, J. (2001), Assessing leadership: a view from the dark side, International Journal of Selection and Development, Vol. 9 No. 1/2, pp. 40-51. Jaworski, J. (2001), Synchronicity, Berrett-Koehler, New York, NY. Kaipiainen, S. (2004), The relationship of emotional intelligence with leadership and self-awareness in predicting organisational outcomes, unpublished PhD thesis, University of London, London. Kets De Vries, M. (1993), Leaders, Fools, Imposters, Jossey-Bass, San Francisco, CA. Kets De Vries, M. and Florent-Treacy, E. (2002), Global leadership from A to Z: creating high commitment organisations, Organisation Dynamics, pp. 295-309. Kotter, J.P. (1990), What leaders really do, Harvard Business Review, May-June, pp. 37-60. Kouznes, J.M. and Posner, B.Z. (1998), Encouraging the Heart, Jossey-Bass, San Francisco, CA. Mathieu, J.E. and Zajac, D.M. (1990), A review and meta-analysis of the antecedents, correlates and consequences of organisational commitment, Psychological Bulletin, Vol. 108, pp. 171-94. Nunnally, J.C. (1978), Psychometric Theory, McGraw-Hill, New York, NY. Salgado, J.F. (2003), Predicting job performance using FFM and non-FFM personality measures, Journal of Occupational & Organizational Psychology, Vol. 76 No. 3, pp. 323-46. Saville, P., Holdsworth, R., Nyeld, G., Cramp, L. and Mabey, W. (1993), Occupational Personality Questionnaire Manual, SHL Group, Esher. Senge, P.M. (1997), Communities of leaders and learners, Harvard Business Review, Vol. 75 No. 5, pp. 30-1. Senge, R., Kleiner, A., Roberts, C., Ross, R., Roth, G. and Smith, B. (2000), The Dance of Change, Nicholas Brealey, New York, NY. Sosik, J.J. and Magerian, L.E. (1999), Understanding leader emotional intelligence and performance, Group and Organisation Management, Vol. 24 No. 3, pp. 367-91. Weick, K.E. (1995), Sensemaking in Organisations, Sage, London. Wheatley, M. (2000), Turing to One Another, Berret- Koehler, San Francisco, CA. Wohlers, A.J. and London, M. (1989), Ratings of managerial characteristics-evaluation difculty, co-worker agreement and self awareness, Personnel Psychology, Vol. 42 No. 2, pp. 235-61. Young, M. (2004), Command, leadership and management competencies predicting superior performance in the Royal Navy, unpublished DBA thesis, Henley Management College/Brunel University, London. Young, M. and Dulewicz, V. (2003), Command, leadership and management competencies predicting superior performance in the Royal Navy, Henley Working Paper 0317, Henley Management College, Henley, available at: www: henleymc.ac.uk Further reading Fiedler, F. (1967), Theory of Leadership Effectiveness, McGraw-Hill, New York, NY. Higgs, M.J. and Rowland, D. (2000), Building change leadership capability: the quest for change competence, Change Management Journal, Vol. 1 No. 2, pp. 116-31. Kets de Vries, M. (1994), The leadership mystique, Academy of Management Executive, Vol. 8 No. 3, pp. 73-5. Kets de Vries, M. (1995), Life and Death in the Executive Fast Lane, Jossey-Bass, San Francisco, CA. Owen, H. (2000), In Search of Leaders, Wiley, New York, NY.

Assessing leadership styles

123