ANALYSING LEADERSHIP AND ITS COMPONENTS: WHAT MAKES THE DIFFERENCE? Halia C. Silins School of Education The Flinders University of South Australia Abstract The strategy employed to identify two correlated dimensions of school leadership, transformational and transactional, and their correlated components is presented. Previous analyses in this area have used Canadian and American data bases and applied traditional factor analytic approaches that assume orthogonality of variables, often inappropriately. Factor analytic approaches to construction of scales and exploration of important educational constructs should make use of recent methodological advances, particularly since the scales are subsequently employed in forming latent variables in structural equation modelling. Oblimin factor analysis and LISREL is applied to data involving correlated leadership constructs. Distinctions are drawn between exploratory and confirmatory factor analysis, and between confirmatory and hierarchical confirmatory factor analysis. Five models of leadership are compared. Introduction The past decade of Australian literature on school leadership reveals a paradigm shift in the role of the principal. Societal change, political pressures, decentralisation and the establishment of self-managing schools have contributed to transforming the principal's role from school administrator to school leader and manager of change. At the same time, the debate around what constitutes effective leadership has intensified and the earlier perceptions of leadership and management as independent concepts have given way to viewing an effective principal as integrating the behaviours of leadership and management (Duignan, 1988; Silins, in press-b). There is an accumulating body of evidence based on Canadian and American data that school leaders can promote change more successfully through practices associated with leadership identified as transformational (Bass & Avolio, 1990; Leithwood & Jantzi, 1990; Silins, 1992). Bass and his associates (Bass, 1985; Bass & Waldman, 1987) have put forward a model of leadership for change using constructs of transformational and transactional leadership. The model of transformational and transactional leadership proposed by Bass (1985) advocates transformational leadership for successful organisational change and improved performance. Bass's model provided the conceptual framework for recently completed studies (Silins, 1992, 1993a, in press-a, in press-b) of the effectiveness of transformational leadership over transactional leadership in bringing about desired school outcomes such as improved teaching and learning. Data from Canadian schools were examined for consistency with Bass's explanation of the relationship between administrative leadership and organisational outcomes. Transformational leadership was found to be a useful construct to inform educational leaders in their striving for educational reform. The type of leadership required of school principals operating within a context of continual change in Australia has also been identified as transformational (Caldwell, 1991; Millikan, 1987; Sungaila, 1990). The influence of Canadian and American writers on Australian literature on leadership is evident. However, Australian empirical studies of transformational leadership seem rare. This study sets out to provide empirical evidence for hypothesised leadership concepts within the Australian education context. Transformational and transactional leadership Transformational leaders inspire their followers to transcend their own needs and to develop a commitment to an organisational vision and purpose (Campbell & Crowther, 1990; Sawatzki, 1989). They engage with others at an emotional, moral, as well as intellectual level (Bates, 1989). Silins (1993a) has identified practices such as providing a model for others, employing collaborative problem solving, encouraging teacher development and providing recognition and support as fundamental to transformational leadership in the school setting. Transformational leadership with its developmental orientation toward followers results in the followers performing beyond expectations (Bass, 1985). Transactional leadership engages followers in an exchange relationship that focusses on their basic needs and applies rewards and sanctions to achieve productivity and efficient management (Bass, 1985). Skills of planning, coordinating, scheduling and regularising have been associated with transactional leadership and the concept of leader as manager (Crawford, 1988). Transactional leadership may enable an organisation to operate effectively and efficiently, but it does not develop in followers the level of trust, loyalty and enthusiasm that is associated with transformational leadership. Beare et al. (1989) noted that although much of the principal's role involves transactional leadership, if excellence is the goal, then transformational leadership practices must prevail. Recent research using Canadian data (Silins, 1993b) has indicated that transformational and transactional leadership are related concepts supporting the view that a nexus between leadership and management is counterproductive. Ridden (1992) commented that the issue was not whether principals were administrators, managers, supervisors or leaders but whether principals looked for leadership opportunities rather than focussing on the daily tasks of running an organisation as an end in itself. Duignan (1988) has maintained that management is the daily practice of leadership when viewed as transactions to achieve organisational purposes and values. He contends that leadership finds its concrete expression within a cultural framework through the existing organisational structures and management processes and procedures. Effective principals use activities of management to accomplish the goals of leadership (Silins, 1993a). Recent Australian literature describes transformational leadership as involving the leader in the development of a school culture (Beare, 1987; Duignan, 1987). Other aspects commonly identified are: providing a vision and developing a commitment to the vision (Cahill, 1991; McCorley, 1988), developing structures and processes that allow for collaboration and participation (Angus, 1989), educating followers to become leaders (Boomer, 1991; Caldwell, 1992; Duignan & Macpherson, 1991), providing opportunities for critical self-reflection (Sungaila, 1990), and considering the needs of followers (Watkins, 1991). Transactional leadership has not had the same attention in recent times. However, transactional leadership has been associated with the traditionally accepted management techniques and procedures which are seen as incomplete and inadequate in today's environment of change (Duignan, 1987, 1988). Leaders are exhorted to develop new ways of thinking and acting. Sungaila (1990) identified first order changes as transactional implying transformational leadership is required for second order changes involving changing assumptions and beliefs and established organisational practices. In addressing the responsibilities of school principals, Wilkinson (1987) linked transactional administration skills to systems requirements and accountability. Angus (1986) and Bates (1982) differentiated between transformational and transactional leadership by describing transactional as rational and value free, whereas transformational was concerned with values, meanings and beliefs. Practices associated with transactional leadership are those related to the technical aspects of leadership focussing on efficiency and productivity. They have been described as involving value free, task-oriented transactions, logical rather than emotional in nature, providing rewards and sanctions and concerned with maintaining the status quo. When transactional leadership is defined in positive terms then it is advocated as a requirement for educational reform together with transformational leadership. When it is perceived as mechanistic, technical, task-oriented and without "soul", it is ignored or dismissed. The study reported here is the first of a series of Australian studies proposed to test transformational models of leadership using school data. While systematic attempts to explore the meaning and utility of such models in educational organisations have only recently begun (Leithwood & Jantzi, 1990; Sergiovanni, 1990) results to date are promising (Leithwood, Jantzi, Silins, & Dart, 1992; Silins, 1992). These results create the expectation that, in many policy implementation contexts, this body of theory will predict quite well the types of leadership initiatives that foster significant change in schools. Method A sample of 23 schools selected from the Adelaide metropolitan area provided 458 primary teachers' responses to the items representing transformational and transactional constructs. Teachers were requested to refer to their experience of school leadership in the context of change. This provided the data base for the analysis reported in this study. For the purposes of this study, a 50 item questionnaire was constructed based on recent Australian perspectives on transformational and transactional leadership found in the literature. The constructs used to generate items were: (1) providing a vision and/or inspiration, (2) creating a culture for success, (3) modelling behaviour, (4) fostering commitment to group goals, (5) providing individual support, (6) providing intellectural stimulation, (7) holding high performance expectations, (8) taking account of individual needs, (9) providing contingent rewards, (10) maintaining status quo (management-by-exception) and (11) employing a bureaucratic, task oriented approach to change. The questionnaire also included items that related to Bass's model of transformational and transactional leadership. A Likert scale with five response categories was employed with response options ranging from 'strongly disagree', 'disagree', 'agree', 'strongly agree', 'not applicable'. Analysis of Data Factor analysis can be used for exploratory or confirmatory statistical purposes. Exploratory factor analysis is a widely accepted technique in the early stages of experimentation or instrument development to detect and assess latent sources of variation and covariation in observed measurements. The variables themselves may or may not have been chosen with underlying processes in mind so that hypotheses about factors thought to reflect underlying processes creating the correlations among variables may be generated or developed (Tabachnick & Fidell, 1989). Confirmatory factor analysis is used in the advanced stages of the research process to test a theory about latent processes or to investigate hypothesised differences in latent processes between groups of subjects. In this usage, variables are carefully chosen to reveal underlying processes. This study used a confirmatory approach to test Bass's model of transformational and transactional leadership in the Australian Primary school context. An exploratory approach was subsequently used to develop latent processes in a set of measured variables generated from recent Australian literature on leadership. Confirmatory Factor Analysis Bass (1985) proposed a multidimensional concept of leadership based on American data from the corporate and military sectors using exploratory factor analysis to generate a model of leadership. Bass and Avolio (1990) demonstrated that responses to their Multifactor Leadership Questionnaire (MLQ) defined five distinct components of leadership behaviour that, they argued, could be categorised into two higher-order factors identified as transformational and transactional leadership. This implicit factor model forms the basis of an a priori model tested in this study. Bass's five construct model of leadership is a hierarchical model that identifies three factors representing transformational leadership: charisma/inspiration, intellectual stimulation and individual consideration; and two factors representing transactional leadership: contingent reward and management-by-exception. Table 1 Conceptual and operational definitions of Bass's five construct model of leadership TRANSFORMATIONAL LEADERSHIP Charisma/inspiration - degree to which leader creates enthusiasm in followers, sees what is really important, and transmits a sense of mission to the organisation, inspires loyalty and devotion, instils pride and faith, and commands respect. Followers place a great deal of trust and confidence in the leader's vision and values, perceive the leader as a role model and want to identify with him or her. 1. Shows us that there are high expectations for us as professionals 2. Instils a sense of belonging 3. Will not settle for second best in performance of our work 4. Symbolises success and accomplishment within our profession 5. Allows us to feel and act like leaders 6. Has both the capacity and judgement to overcome most obstacles 7. Evokes respect from everyone in the school 8. Excites us with visions of what we may be able to accomplish if we work together Intellectual Stimulation - degree to which the leader provides intellectual and problem- oriented guidance, arouses followers to think in new ways. Followers are encouraged to question their own and others' assumptions, beliefs and values, and develop independent problem- solving capabilities. 9. Stimulates me to think about what I am doing for my students 10. Provides information that encourages dialogue about change 11. Encourages me to reflect critically on the educative process 12. Helps us make personal meaning of change 13. Challenges me to re-examine some basic assumptions about my work with students Individual Consideration - degree to which the leader is concerned with the individual needs of followers, responds to individual differences in followers' needs for growth and development, elevating needs and abilities to higher levels when appropriate, and delegating projects to stimulate individual learning experiences. Followers are provided with coaching and teaching as needed. 14. Provides for extended training to develop my knowledge and skills 15. Treats me as an individual with unique needs and expertise 16. Offers personal encouragement for my good performance 17. Behaves in a manner thoughtful of my personal needs 18. Provides public recognition and acknowledgement when work is especially good 19. Is concerned for our feelings and ability to cope with change TRANSACTIONAL LEADERSHIP Contingent Reward - degree to which the leader makes clear what the follower must accomplish in order to be rewarded, provides rewards if followers perform in accordance with contracts or expend the necessary effort to meet performance standards. Leaders recognise what subordinates need and, through extrinsic rewards, energise followers to reach objectives. 20. Publicly praises work perceived as useful 21. Exchanges tangible rewards for substantive work 22. Assures me that I can get what I personally want in exchange for my efforts Management-by-exception - degree to which the leader provides negative feedback for failure to meet agreed-upon standards, avoids giving directions if the old ways are working and allows followers to continue doing their jobs as always if performance goals are met. 23. Asks no more of us than what is absolutely essential to get our work done 24. Is satisfied with our performance as long as the old ways work 25. Provides only the information that we have to know to do our jobs 26. Does not try to change anything as long as things are going all right 27. Lets us take initiatives but does not encourage us to do so Table 1 presents the items chosen by the researcher to represent Bass's five constructs of leadership based on previous discussions with Bass in 1991, Bass (1985) and the MLQ Manual (Bass & Avolio, 1990). Bass's model suggests a structure consisting of two broad constructs represented by transformational and transactional leadership that are further defined by three and two factors respectively. There is evidence to suggest that transformational and transactional leadership are related concepts and that their underlying processes are correlated (Silins, 1992, 1993b). A confirmatory factor analysis approach with Australian data was used to show the actual factor structure that emerged from a principal factors extraction with oblimin rotation. Principal factors extraction was chosen because it conforms to the factor analytic model in which common variance is analysed with unique and error variance removed. The decision to use oblique rotation was made because it allows a continuous range of correlations between factors (Carroll, 1983). Rotation into five and two factors was specified by the researcher and performed using SPSS Factor (SPSS Inc., 1990) on 27 items from the survey data. Table 2. Factor Loadings*, Eigenvalues, Percent Variance, and Correlations Between Factors for Principal Factors Extraction and Oblimin Rotation for Five and Two Factor Solutions for Leadership Variables on School Survey Items F1 F2 F3 F4 F5 Eigenvalues 9.51 1.16 0.80 0.56 0.41 Per cent Variance 36.6 4.4 3.1 2.2 1.6 Cumulative Percent Variance 36.6 41.0 44.1 46.3 47.8 Leadership Items - 5 Factor Solution Transformational Leadership Charisma/inspiration 1. Shows high expectations .57 2. Instils a sense of belonging .51 -.32 3. Will not settle for second best .41 4. Symbolises success .60 5. Allow us to feel and act leaders .37 6. Has capacity overcome obstacles .43 .43 7. Evokes respect in the school .45 .33 8. Excites us with visions .43 Intellectual stimulation 9. Stimulates me to think .52 10. Provides information dialogue .35 .38 11. Encourages me to critically reflect .64 12. Helps make meaning of change .43 13. Challenges basic assumptions .73 Individual consideration 14. Provides for training skills 15. Individual with unique needs .41 -.44 16. Offers personal encouragement -.52 17. Behaves thoughtful of needs .57 -.43 18. Provides recognition work is good -.58 19. Concerned for ability to cope .55 Transactional Leadership Contingent reward 20. Praises work perceived as useful - .61 22. Assures me I can get what I want Management-by-exception 23. Asks no more than essential 24. Satisfied as long as old ways work .47 25. Only the information have to know -.30 26. Not change as long as all right .62 27. Take initiatives but not encouraged Correlations between factors F1 F2 F3 F4 F5 Factor 1 1.00 Factor 2 -.11 1.00 Factor 3 .40 -.12 1.00 Factor 4 .33 -.27 .43 1.00 Factor 5 -.49 .00 -.49 -.34 1.00 Cont. Table 2 Continued F1 F2 Eigenvalues 9.45 1.06 Per cent Variance 36.3 4.1 Cumulative Percent Variance 36.3 40.4 Leadership Items - 2 Factor Solution Transformational Leadership Charisma/inspiration 1. Shows high expectations .57 2. Instils a sense of belonging .78 3. Will not settle for second best .39 4. Symbolises success .69 5. Allow us to feel and act leaders .69 6. Has capacity overcome obstacles .71 7. Evokes respect in the school .75 8. Excites us with visions .73 Intellectual stimulation 9. Stimulates me to think .64 10. Provides information dialogue .69 11. Encourages me to critically reflect .71 12. Helps make meaning of change .51 13. Challenges basic assumptions .48 Individual consideration 14. Provides for training skills .62 15. Individual with unique needs .79 16. Offers personal encouragement .74 17. Behaves thoughtful of needs .74 .34 18. Provides recognition work is good .59 19. Concerned for ability to cope .74 Transactional Leadership Contingent reward 20. Praises work perceived as useful .58 22. Assures me I can get what I want .33 Management-by-exception 23. Asks no more than essential .31 24. Satisfied as long as old ways work 25. Only the information have to know -.54 26. No change as long as all right .44 27. Lets initiatives but not encourage -.35 Correlations between factors F1 F2 Factor 1 1.00 Factor 2 -.07 1.00 *Only factor loadings > .30 are given Table 2 gives the factor loadings and structure for a five and two factor solution as well as the correlations among the factors. When forcing five factors with the full 27 items, oblimin failed to converge. When the item with the lowest communality was dropped (item 21), oblimin converged resulting in factors that did not correspond well to Bass's hypothesised factors. Factor 1 became a compendium of charisma/inspiration and individual consideration, Factor 3 adequately represented intellectual stimulation, Factor 4 represented a new construct described as leader's high expectations of followers, Factor 5 represented individual consideration with praise and recognition by the leader as an element, and Factor 2 was the remaining factor of transactional leadership, management-by-exception, with only two of the items loading on this factor. The actual factors emerging from the two factor solution confirm the transformational scale as Factor 1, which includes all of the items used for that constructed factor. It should be noted that the two remaining contingent reward items loaded positively on Factor 1. Also, two of the management-by-exception items (25, 27) loaded negatively on Factor 1 and item 24 failed to load on either factor leaving transactional leadership poorly defined. Eigenvalues and per cent variance indicate that together the five factors account for just under 50 per cent and the two factors just under 40 per cent of the variance that exists in the items defining the factors with the first factor, in each case, making the substantial contribution. Only three of Bass's five constructs are validated by these data: intellectual stimulation, individual consideration and, to a lesser extent, management-by-exception. The two higher-order factors are supported describing the transformational scale as largely a strong single factor whereas the transactional scale is comparatively weak and ill-defined. Contingent reward is not associated with transactional leadership. Previous factor analytic research using Canadian data also failed to expose the relationship of the contingent reward items to the construct of transactional leadership (Silins, 1992, 1993). These procedures are confirmatory only so far as they have been restricted to examining two and five factor solutions. They fail to be confirmatory in so far as the items are not constrained to the hypothesised factors. Nevertheless, this is the conventional confirmatory approach that has been used in the absence of the kind of approach that linear structural relations analysis (LISREL) permits (Jšreskog & Sšrbom, 1993). Exploratory Factor Analysis Exploratory factor analysis was used to examine the properties and the dimensionality of the set of leadership scale items generated from leadership constructs extracted from current Australian literature. Principal component analysis with varimax rotation has been used in the past to identify items associated with two or more orthogonal scales (Oppenheim, 1992). It was considered essential that scales should be orthogonal to avoid the problems of multicollinearity in subsequent explanatory analyses. However, recent developments in analytical techniques of latent variable construction do not demand that attitude scales be orthogonal. Carroll (1983) argued convincingly for the use of principal factoring procedure with oblique rotations when substantial correlations exist among the primary factors. These methods of analysis were used in this study to establish the probable factorial composition and structure of the set of measured variables obtained from teachers' responses to the leadership items. This enabled the selection of items that contributed to the description of leadership. Carroll (1983) pointed out that finding the "correct" number of primary factors for a set of data is one of the more difficult problems in factor analysis. It is important to note, however, that this procedure leaves unclear the exact structure of the interrelations between the characteristics and the constructs of transformational and transactional leadership. After considerable experimentation with different numbers of factors, a six and two factor solution exhibited most consistency and meaning using principal factors extraction with oblimin rotation. The two factors reflected the transformational and transactional leadership constructs. The six factors were interpreted as four transformational characteristics identified as: (1) goal achievement, (2) individual consideration and support, (3) intellectual stimulation and (4) collaborative problem solving; and two transactional characteristics identified as: (1) technical orientation and (2) management-by-exception. Table 3 presents the observed measures underlying these six constructs and their hypothesised conceptual definitions. Table 4 presents the structure of the observed variables as they relate to the hypothesised hierarchical model of six first-order and two second-order factors. A problem with exploratory factor analysis is that, after extraction, there are an infinite number of rotations available that account for the same amount of variance in the original data, but with slightly differently defined factors. The researcher makes a final choice among the alternative solutions based on an assessment of its interpretability and scientific utility. This can result in a proliferation of models that are not confirmed or disproved by subjecting them to more rigorous statistical techniques (Jšreskog & Sšrbom, 1989). Factor analytic techniques can be employed to bias results in favour of any one model (Gustafsson, 1984). In this study, a six factor solution with two second higher-order factors is hypothesised. However, other models can be derived from this data. There is evidence for several alternatives: (a) no subsets of leadership, (b) only two subsets, (c) six subsets, (d) the hierarchical model of one second-order factor and two first-order factors, or (e) the hierarchical model of two second-order factors containing four first-order transformational factors and two first-order transactional factors respectively. Table 3 Conceptual and operational definitions of a six construct model of leadership TRANSFORMATIONAL LEADERSHIP Goal Achievement - degree to which the leader motivates and energises followers, helps them see what is really important for themselves and the organisation, transmits a sense of purpose and organisational mission and builds a commitment to change and improved performance. 1. Shows us that there are high expectations for us as professionals 2. Encourages us regularly to evaluate our progress toward achievment of school goals 3. Gives us a sense of overall purpose 4. Encourages teachers to work toward the same goals 5. Will not settle for second best in performance of our work 6. Symbolises success and accomplishment within our profession 7. Stimulates me to think about what I am doing for my students Intellectual Stimulation - degreee to which the leader provides intellectual and problem- oriented guidance, encourages enquiry and shared construction of meanings, encourages questionning of one's own and others' assumptions, beliefs and values and promotes understanding. 8. Provides information that helps me think of ways to implement initiatives 9. Helps us make personal meaning of change 10. Provides the necessary resources to support my participation in change initiatives 11. Challenges me to re-examine some basic assumptions about my work with students 12. Assures me that I can get what I want in exchange for my efforts Individual Consideration and Support - degree to which the leader is concerned with the individual needs of followers, responds to individual differences in followers' needs and provides resources for growth and development, instils confidence in the individual's ability to perform, delegates projects to stimulate individual learning experiences and recognises effort and achievement. 13. Offers personal encouragement for my good performance 14. Treats me as an individual with unique needs and expertise 15. Behaves in a manner thoughtful of my personal needs 16. Has confidence in my ability to perform at my best 17. Publicly praises work perceived as useful 18. Instils a sense of belonging 19. Provides public recognition and acknowledgement when work is especially good Collaborative Problem Solving - degree to which leader encourages collaboration and participative decision making, develops goals by consensus, promotes collegiality and negotiation and encourages meaningful dialogue between all concerned. 20. Works toward whole staff consensus in establishing priorities for school goals 21. Uses problem solving with staff to generate school goals 22. Provides for our participation in the process of developing school goals 23. Provides information that encourages dialogue about change TRANSACTIONAL LEADERSHIP Technical Orientation - degree to which leader is task focussed and policy driven, reactive rather than proactive, approaches change independently of followers' concerns and needs and strives to attain administrative efficiency with the emphasis on management rather than leadership. 24. Allows administrative expedience to override the consultative process 25. Empowers us more by default than by design 26. Puts departmental policy requirements above school requirements 27. Exchanges tangible rewards for substantive work 28. Lets us take initiatives but does not encourage us to do so Management-by-exception - degree to which the leader provides negative feedback for failure to meet agreed-upon standards, avoids giving directions if the old ways are working and allows followers to continue doing their jobs as always if performance goals are met. The leader concentrates on maintaining the status quo rathan seizing opportunities for change. 29. Does not try to change anything as long as things are going all right 30. Is satisfied with our performance as long as the old ways work 31. Provides only the information that we have to know to do our jobs 32. Directs through existing rules and procedures Table 4 Factor Loadings*, Eigenvalues, Percent Variance, and Correlations Between Factors for Principal Factors Extraction and Oblimin Rotation for Six and Two Factor Solutions for Leadership Variables on School Survey Items F1 F2 F3 F4 F5 F6 Eigenvalues 10.48 1.56 0.99 0.77 0.59 0.43 Per cent Variance 32.8 4.9 3.1 2.4 1.8 1.4 Cumulative Percent Variance 32.8 37.6 40.7 43.1 44.9 46.3 Leadership Items - 6 Factor Solution Transformational Leadership Goal Achievement 1. Shows high expectations .63 2. Evaluate progress toward goals .58 3. Gives sense overall purpose .57 4. Work toward same goals .56 5. Will not settle for second best .41 6. Symbolises success .40 7. Stimulates me to think .39 Intellectual stimulation 8. Information implement initiatives - .65 9. Helps make meaning of change -.56 10. Resources change initiatives -.54 11. Challenges basic assumptions -.42 12. Assures me I can get what I want - .32 Individual consideration 13. Offers personal encouragement .65 14. Individual with unique needs .64 15. Behaves thoughtful of needs .59 16. Confidence in ability to perform .56 17. Praises work perceived as useful .56 18. Instils a sense of belonging .53 19. Provides recognition work is good .50 Collaborative problem solving 20. Staff consensus for school goals .73 21. Staff generate school goals .62 22. Participation developing goals .42 23. Provides information dialogue .33 Transactional Leadership Technical orientation 24. Admin override consultative .56 25. Empowers by default .45 26. Departmental policy above school .38 27. Exchanges rewards for work .32 28. Lets initiatives but not encourage .31 Management-by-exception 29. No change as long as all right .55 30. Satisfied as long as old ways work .44 31. Only the information have to know .37 32. Directs through rules procedures .33 Correlations between factors F1 F2 F3 F4 F5 F6 Factor 1 1.00 Factor 2 -.06 1.00 Factor 3 .47 -.13 1.00 Factor 4 .41 -.17 .46 1.00 Factor 5 -.42 .01 -.43 -.42 1.00 Factor 6 -.28 .35 -.18 -.25 .16 1.00 Cont. Table 4 Continued F1 F2 Eigenvalues 10.38 1.50 Per cent Variance 32.4 4.7 Cumulative Per cent Variance 32.4 37.1 Leadership Items - 2 Factor Solution Transformational Leadership Goal achievement 1. Shows high expectations .58 2. Evaluate progress toward goals .60 3. Gives sense overall purpose .74 4. Work toward same goals .65 5. Will not settle for second best .42 6. Symbolises success .68 7. Stimulates me to think .63 Intellectual stimulation 8. Information implement initiatives .68 9. Helps make meaning of change .61 10. Resources change initiatives .71 11. Challenges basic assumptions .55 12. Assures me I can get what I want .42 .30 Individual consideration 13. Offers personal encouragement .71 14. Individual with unique needs .72 15. Behaves thoughtful of needs .67 16. Confidence in ability to perform .64 17. Praises work perceived as useful .59 18. Instils a sense of belonging .73 19. Provides recognition work is good .60 Collaborative problem solving 20. Staff consensus for school goals .51 21. Staff generate school goals .61 22. Participation developing goals .50 23. Provides information dialogue .69 Transactional Leadership Technical orientation 24. Admin overrides consultative -.31 .58 25. Empowers by default .49 26. Departmental policy above school .42 27. Exchanges rewards for work .30 28. Lets initiatives but not encourage .37 Management-by-exception 29. No change as long as all right .32 30. Satisfied as long as old ways work 31. Only the information have to know -.40 .42 32. Directs through rules procedures .43 Correlations between factors F1 F2 Factor 1 1.00 Factor 2 -.23 1.00 *Only factor loadings > .30 are given Hierarchical Confirmatory Factor Analysis Recent advances in the application of confirmatory factor analysis (CFA) and hierarchical confirmatory factor analysis (HCFA) using LISWIN8 (Jšreskog & Sšrbom, 1993) test the ability of a priori factor models to fit empirical data rather than simply trying to interpret the factors generated by exploratory factor analysis to fit the model. Marsh (1991) points out that in exploratory factor analysis, the researcher is a passive interpreter of the empirical results in relation to theory and prior research because of limited control over the structure. In CFA a priori models are compared for their ability to fit the observed data. In CFA and oblique factor analyses, the correlations among factors are estimated freely. Using exploratory factor analysis, a second factor analysis of the correlations among factors in the first factor analysis can be conducted. A better approach, however, is to use HCFA since the first-order and second-order factors are tested simultaneously in a single analysis. Marsh (1991) has noted that as long as the first-order factors are at least moderately correlated, it is also possible to test for higher-order factors. In the second stage of this investigation, HCFA analysis using LISWIN8 is applied to the set of first-order and second-order factors incorporated in the models advanced in this paper. HCFA and CFA using LISWIN8 is also applied to the alternative models generated by the data. CFA and HCFA provide the ability to compare the goodness of fit for competing models and, therefore, overcome the failure of exploratory factor analysis to disconfirm alternative models. The LISWIN8 program is very simple to use and is built around SIMPLIS and PRELIS2. This LISREL program provides estimates of criteria used in assessing the overall goodness of fit of a model to the data: the value of chi-square relative to the degrees of freedom, the probability that the true chi-square value is larger than the obtained value, the adjusted goodness of fit index, the root mean square residual, and the number of fitted relative unconstrained residuals. Guidelines for interpretation of these criteria are provided by Tuijnman & Keeves (in press). Table 5 presents the results of applying LISWIN8 to the five models derived from this data. Table 5 Goodness-of-Fit Indices for Models Positing First-Order (FO) and Second-Order (SO) Factors Based on A Priori Structures. No of factors Goodness-of-fit indices Model FO SO df Ď2 Ď2/df GF AGF p MI (a)1 only 1 434 2652 6.1 0.69 0.64 <.001 78.3 (b)2 only 2 433 2391 5.5 0.72 0.68 <.001 78.5 (c)6 only 6 419 3343 8.0 0.66 0.60 <.001 none (d)1-2nested 2 1 402 1828 4.5 0.79 0.74 <.001 65.0 (e)2-4,2nested 4,2 2 381 1348 3.5 0.85 0.81 <.001 41.6 The modification indices obtained for these five models would not appear to warrant further testing of a posteriori models. However, it is clear that the 4 by 2 nested model has the lowest Ď2/df ratio and the highest goodness-of-fit indices and must be considered, at this time, the preferred model. Conclusion The present analysis has been carried out with a relatively small sample to explore leadership concepts within the Australian context. Some support has been found for a higher-order factor model of leadership consisting of two second-order factors containing four first-order transformational factors and two first-order transactional factors. However, this model of leadership is far from satisfactory. A further study of the components of leadership and their influence on school outcomes is currently underway. The procedures used here will be employed again using a broader conceptualisation of leadership. With a larger sample and more stable estimates, there is an expectation that further progress will be made toward developing a model of leadership that will contribute to identifying school leadership practices that make the difference. References Angus, L. (1986). Schooling, the school effectiveness movement and educational reform. Burwood, VIC: Deakin University Press. Angus, L. (1989). New leadership and the possibility of educational reform. In J. Smyth (Ed.), Critical perspectives on educational leadership. (pp. 63-92). Philadelphia, PA: The Falmer Press. Bass, B. M. (1985). Leadership and performance beyond expectations. New York, NY: The Free Press. Bass, B. M., & Avolio, B. J. (1990). Transformational leadership development: Manual for the Multifactor Leadership Questionnaire. Palo Alto, California: Consulting Psychologists Press. Bass, B. M., & Waldman, D. A. (1987). Transformational leadership and the falling dominoes effect. Group & Organization Studies, 12(1), 73-87. Bates, R. J. (1982). Towards a critical practice of educational administration. Commonwealth Council for Educational Administration. Studies in Educational Administration, no.27(September), 1-15. Bates, R. J. (1989). Leadership and the rationalization of society. In J. Smyth (Ed.), Critical perspectives on educational leadership. Philadelphia, PA: The Falmer Press. Beare, H. (1987). Metaphors about schools: The principal as a cultural leader. In W. S. Simpkins, R. A. Thomas, & B. E. Thomas (Eds.), Principal and change: The Australian experience. Armidale, N.S.W: University of New England. Beare, H., Caldwell, B. J., & Millikan, R. H. (1989). Creating an excellent school: Some new management techniques. London: Routledge. Boomer, G. (1991). Curriculum change - the principal and the effective school management: A discussion paper. Paper prepared for South Australian Primary Principals' Association Conference, Victor Harbor. Cahill, W. (1991). Entrepreneurship and leadership. In B. Caldwell & F. Crowther (Eds.), The entrepreneurial school. Gosford, N.S.W: Ashton Scholastic. Caldwell, B. J. (1991). The changing role of the school principal: A review of developments in Australia and New Zealand. Paper presented in a Symposium on 'The changing role of the school principal: International perspectives' at the Annual Meeting of the American Educational Research Association, Chicago. Caldwell, B. J. (1992). The principal as leader of the self- managing school in Australia. Journal of Educational Administration, 30(3), 6-19. Campbell, D., & Crowther, F. (1990). What is an entrepreneurial school. In D. Campbell & F. Crowther (Eds.), The entrepreneurial school. Gosford, N.S.W: Ashton Scholastic. Carroll, J. B. (1983). Studying individual differences in cognitive abilities: Through and beyond factor analysis. In R. F. Dillon & R. R. Schmeck (Eds.), Individual differences in cognition. New York, NY: Academic Press. Crawford, P. J. (1988). Executive leadership in a rapidly changing environment. The Practising Administrator, 9(1), 31-36. Duignan, P. (1987). Leaders as culture builders: Leadership as a vital element in the development of school culture. Unicorn, 13(4), 208-213. Duignan, P. A. (1988). Reflective management: The key to quality leadership . Armidale, NSW, University of New England, Centre for Administrative and Higher Education Studies. Duignan, P. A., & Macpherson, R. J. S. (1991). Rationalisation and the educative leader. The Practising Administrator, 13(4), 30-32. Gustafsson, J.-E. (1984). A unifying model for the structure of intellectural abilities. Intelligence, 8, 179-203. Jšreskog, K. G., & Sšrbom, D. (1989). LISREL7 A guide to the program and applications (2nd edition). Chicago, IL: SPSS Inc. Jšreskog, K. G., & Sšrbom, D. (1993). LISREL8: Structural equation modeling with the SIMPLIS Command Language. Hillsdale, NJ: Lawrence Erlbaum Associates. Leithwood, K., & Jantzi, D. (1990). Transformational leadership: How principals can help reform school cultures. School Effectiveness and School Improvement, 1(4), 249-280. Leithwood, K., Jantzi, D., Silins, H., & Dart, B. (1992, January). Transformational leadership and school restructuring. Paper presented at the International Congress for School Effectiveness and Improvement, Victoria, BC. Marsh, H. W. (1991). Multidimensional students' evaluations of teaching effectiveness: A test of alternative higher-order structures. Journal of Educational Psychology, 83(2), 285-296. McCorley, J. (1988). Leaders must have a vision - Towards a model. The Practising Administrator, 10(2), 12-15,46-47. Millikan, R. (1987). School culture: A conceptual framework. Educational Administration Review, 5(2), 38-56. Oppenheim, A. N. (1992). Questionnaire design, interviewing and attitude measurement. New York, NY: Pinter. Ridden, P. (1992). School management. A team approach. Gosford, N.S.W: Ashton Scholastic. Sawatzki, M. (1989). The role of the principal - An inside point of view. A.C.T. Principal. 1, 20-23. Sergiovanni, T. J. (1990). Value-added leadership: How to get extraordinary performance in schools. New York: Harcourt Brace Jovanovich. Silins, H. C. (1992). Effective leadership for school reform. The Alberta Journal of Educational Research, 38(4), 317-334. Silins, H. C. (1993a). Leadership for school reform: A test of a transformational and transactional model (Doctoral dissertation, University of Maryland, 1992). Dissertation Abstracts International, 53(2198A). Silins, H. C. (1993b). The relationship between school leadership and school improvement outcomes. Paper presented at the AERA Annual Meeting, Atlanta, GA. Silins, H. C. (in press-a). Leadership characteristics and school improvement. Australian Journal of Education. Silins, H. C. (in press-b). The relationship between transformational and transactional leadership and school improvement outcomes. School Effectiveness and School Improvement. SPSS Inc. (1990). SPSS Reference guide. Chicago, IL: SPSS Inc. Sungaila, H. (1990). The new science of chaos: Making a new science of leadership? Journal of Educational Administration, 28(2), 4-23. Tabachnick, B. G., & Fidell, L. S. (1989). Using multivariate statistics (2nd ed.). New York, NY: Harper & Row. Tuijnman, A. C., & Keeves, J. P. (in press). Path and linear structural relations analysis. In T. HusŽn & T. N. Postlethwaite (Eds.), International encyclopedia of education Oxford: Pergamon Press. Watkins, P. (1991). Devolving educational administration in Victoria: Tensions in the role and selection of principals. Journal of Educational Administration, 29(1), 22-38. Wilkinson, V. (1987). Managing change - the principal meets the future: 'The Victorian experience". WA Primary Principal, 3(2), 5-37.