Development of a measurement for transformational and transactional leadership in schools taking on a school-based intervention

Abstract

This article describes the development and validation of a measure for transformational and transactional leadership that may be used as part of a more comprehensive instrument to measure a school’s capacity to implement school-based interventions. Data used for this study were gathered from the reports of 1144 teachers from 45 Norwegian schools who were taking on a school development intervention aimed at preventing and reducing problem behaviours among students. Using exploratory factor analysis and confirmatory factor analysis, measurement structures with two factors at the teacher level and one factor at the school level provided a good fit to the data. Although there is some overlap between transactional and transformational leadership, the two concepts do measure distinct features. Transformational and transactional leadership was eventually measured by teacher reports on two scales, each consisting of four items. As part of the validation, scores on the scales were related to other aspects relevant for organisational capacity. As expected, the transformational and transactional leadership scales both correlated significantly negatively with staff freedom and positively with innovation climate, affiliation, principal positions in schools, and collaborative activity at both the teacher and the school levels.

Keywords

Transformational leadership transactional leadership scale development school improvement interventions

Introduction

The aim of this article is to present a measure for leadership that may be used as part of a more comprehensive instrument to measure a school’s capacity to implement school-based interventions. Being able to measure the constructs of both transformational and transactional leadership can provide valuable information about the type of leadership being practiced in a school and where support may be given in order to improve leadership and, consequently, capacity to change at the school level. Assessing capacity to change is important before a school-based programme is implemented, as a mismatch between the school’s capacity and the capacity required to successfully implement the intervention may lead to implementation failure (Flaspohler et al., 2008). By identifying schools where implementation failure is most likely to occur before a change initiative is introduced, we may be able to provide these schools with the assistance they need to build internal capacity (Gingiss et al., 2006).

School-based interventions that are complex in both content and range of participants will demand a strong focus on leadership, especially with respect to the principal’s ability to promote collective leadership among staff members (Ertesvåg and Roland , 2011). This type of leadership is a necessity, but also a great challenge, as it will demand a strong formal leader who creates a culture and structures that enable the development of sustainable leadership on many levels within the school (Ertesvåg and Roland , 2011). We argue that a combination of transformational and transactional leadership may provide the necessary leadership practices required to successfully implement complex interventions. Research shows that transformational leadership practices can be widely distributed across the organisation (Leithwood et al., 2004). Meanwhile, transactional leadership may provide the predictable, orderly and positive school environment necessary for teachers to be able to play a role in the decision-making process (Stigler and Hiebert, 1999). Based on the previous discussion, we support the premise that a principal’s ability to skilfully balance a combination of transformational and transactional leadership styles may result in greater capacity and a more successful school. Furthermore, we argue that the principal, in conjunction with the use of collaborative efforts and shared decision-making by stakeholders towards common visions and goals, may produce a more effective school (Sergiovanni, 2007).

Theoretical framework

We are in the process of developing an instrument designed to measure a number of key organisational factors that may influence a school’s capacity to change (see Oterkiil and Ertesvåg, 2012). While leadership is only one of these factors, it may be the most prominent one because of its ability to influence the other factors and to bring about the potential capacities that exist within a school. Due to the complex nature of leadership, the concept of leadership is difficult to fully grasp and define. As a consequence, a number of different leadership theories have emerged throughout the years, and several instruments to measure these concepts are now available. One of the most prominent theories in relation to school improvement is transformational leadership, as it has been well studied within educational settings and provides an empirically grounded theory on the role leadership plays in supporting or inhibiting organisational change (Bass and Avolio, 1994). During the last 20 years, numerous studies have investigated the constructs of transformational and transactional leadership theory, and, consequently, we are now left with a substantial body of research supporting as well as critiquing these constructs.

The transformational and transactional leadership styles were first characterised by Burns (1978) and originally proposed for use in political settings. Based on Burn’s conceptualisation, though with a few modifications and elaborations, Bass (1985) later developed a model of transformational leadership for use in a number of organisational contexts. Contrary to Burns (1978), Bass (1999) argued that rather than representing opposite ends of a single continuum, transformational and transactional leadership are conceptually independent, thus suggesting that a leader can demonstrate low or high levels in either or both styles simultaneously. Bass further claimed that the best leaders are those who are both transformational and transactional, as transformational behaviours augment the effects of transactional behaviours and because the two leadership styles focus on different aspects of the organisation. In support of this claim, Midthassel and Erstesvag (2008) found that those schools who successfully implemented the interventions had leaders who were both transformational and transactional. Unsuccessful schools, on the other hand, had leaders who failed to execute practices related to one or both of these leadership dimensions. Furthermore, Larsen and Samdal (2007) observed that principals who skilfully balanced both leadership and management strategies were able to motivate teachers to use the intervention that was being implemented, an essential factor behind any systematic effect. In their systematic reviews of the literature, Leithwood and colleagues found that transformational leadership has a significant effect on teachers’ emotions, beliefs and commitments to change (Leithwood and Sun, 2009; Leithwood et al., 1993, 1994), as well as on organisational improvement and on the organisation’s culture and climate (Leithwood et al., 1996). These findings provide further support to our argument that a contextually sensitive set of transformational leadership practices designed to promote a collective culture, in combination with transactional leadership practices essential in providing a positive learning environment for both teachers and students, may lead not only to successful change initiatives but also to sustainable change.

As previously mentioned, the scales presented in this study are developed as part of a more comprehensive instrument to measure several factors that may influence a school’s capacity to implement a school-based intervention, with leadership being only one of those factors (see Figure 1). As a basis for this development process, we make use of the Burke-Litwin Causal Model of Organization Performance and Change (Burke and Litwin, 1992) adapted by us to specifically guide the assessment of readiness and the capacity to implement school-based interventions (see Oterkiil and Ertesvåg , 2012). The top half of the model consists of the mission, strategy, leadership and culture, all of which are defined as transformational factors that most directly respond to external environmental dynamics. The lower half is made up of management practices, structure, resources, climate, motivation, skills, knowledge, beliefs and values, all of which are transactional factors. According to Burke (2008), a comparison between the transactional categories in the top half of the model and the transactional categories in the bottom half is the key to understanding change.

Figure 1.

A model to guide the assessment of readiness and capacity to implement evidence-based prevention programmes in schools, adapted from the Burke-Litwin model of organisational change (Burke and Litwin, 1992).

We believe that an instrument based on the theoretical framework presented above may help us identify vital organisational variables that, based on previous research, have been found to influence an organisation’s capacity and readiness to improve. Moreover, the model provides a systematic way to collect, interpret and categorise the organisational data, and it helps to identify key strengths in the organisation as well as issues that need attention (Burke, 2008). Although complex, the model is still a simplification of reality and will never be able to account for all factors affecting an organisation’s capacity to change. The concepts of transformational and transactional leadership closely resemble what Burke (2008) refers to in the model as leadership and management. Burke’s definition of leadership and Bass’s concept of transformational leadership both describe leadership practices that provide direction, inspiration and motivation and bring about organisational change. Management and transactional leadership, on the other hand, entail leadership practices that involve planning, controlling, allocating resources, establishing predictability and maintaining the status quo. Our understanding of transactional and transformational leadership is based on the notion that transformational leadership is focused on change while transactional leadership is focused on management, e.g. maintaining the day-to-day activities associated with the intervention.

To produce successful and sustainable change, the principal as a key resource must be able to draw from a number of effective leadership and management practices in order to foster a school culture and climate where teachers feel ownership and support for the change. We agree with Leithwood and Sun’s (2012) suggestion that more attention should be devoted to the impact of specific leadership practices and less to leadership models. In their meta-analytic review of unpublished studies about the nature of transformational leadership, Leithwood and Sun (2012) compared their findings with results of earlier systematic reviews and found that a number of the most widely advocated models of effective educational leadership actually contain several of the same practices. In our development of the instrument presented here, we have therefore, as suggested by Leithwood and Sun (2012), focused more on producing a set of items that capture the range of specific leadership practices proven to be the most effective when implementing school-based interventions, rather than adhering to one specific leadership model. We do, however, categorise these practices as being either transformational or transactional, based on where they fit within our definition of these two leadership styles and our conceptual model based on the Burke-Litwin model of organisational change. Furthermore, based on an extensive synthesis of research, Leithwood et al. (2008) make the claim that most successful leaders have one thing in common: they all draw on the same set of four core leadership practices. These practices include building a vision and setting directions, understanding and developing people, redesigning the organisation, and managing the teaching and learning programme (Leithwood et al., 2008), all of which are part of a transformational or transactional leadership style. The first three leadership practices closely resemble what Burke and Litwin (1992) refer to as ‘leadership’ (transformational) in their model of organisational change, while the fourth practice resembles what they call management (transactional). Although the transformational leadership model developed by Leithwood and colleagues in its initial formulations is influenced by the work of Burns and Bass, as well as their own comprehensive research in schools, they have chosen to add, eliminate or give new meaning to some of the constructs of transformational and transactional leadership (Leithwood and Jantzi, 2006). For instance, one of the greatest differences between Leithwood and Jantzi’s (2006) understanding of transformational leadership and that of Bass is that the managerial dimension of school leadership is one component of transformational leadership rather than representative of another leadership style (i.e. transactional leadership) opposite that of transformational leadership. We further elaborate on this notion as we, unlike Bass, regard transactional leadership to be more of a complement to the transformational leadership style rather than to represent an inferior and somewhat negatively charged leadership style. However, contrary to Leithwood and colleagues and in accordance with our theoretical framework, we do not include the transactional leadership practices as part of the concept of transformational leadership. Rather, transformational leadership consists of leadership practices related to Leithwood and colleagues’ concepts of building a vision and setting directions, understanding and developing people, and redesigning the organisation. We have included their fourth concept of managing the teaching and learning programme as part of our perception of transactional leadership.

Although influenced in its initial formulations by well-known views both within and outside of school contexts (e.g. Bass and Avolio, 1994; Leithwood and Sun, 2012; Leithwood et al., 2001), we also draw on the Burke-Litwin model (Burke, 2008; Burke and Litwin, 1992), highlighting transformational and transactional factors that affect organisational change. The model distinguishes between the set of variables that influences and is influenced by the organisational climate (everyday transactional level) and those of the organisational culture (fundamental transformational level). Based partly on the theoretical outline above and partly on our previous work on school-based interventions (e.g.Ertesvåg, Roland, Vaaland, Størksen and Veland , 2010; Oterkiil and Ertesvåg 2012;, Midthassel and Erstesvag, 2008), this distinction seems important to capture when a school is in need of change, as the nature of the change initiative may indicate on which of these two levels the main focus should be. The application of the Burke-Litwin model to school-based interventions and the understanding of the transformational and transactional factors related to the application, including the leadership aspects, are discussed in detail elsewhere (see Oterkiil and Ertesvåg, 2012). Accordingly, some of the central features associated with transformational leadership by others either are not part of our consideration (e.g. charisma; see, e.g., Groves, 2005) or are awarded somewhat different significance (e.g. transactional practices; see Bass, 1985). As our measure is developed based on a distributed/shared leadership perspective, our aim will be to measure leadership practices performed by several people within the school, not solely by the principal. Hence, our measure does not include individual traits traditionally associated with the sole ‘charismatic and inspirational leader’, as they have proven to be difficult to both conceptualise and measure (see Hunt and Conger, 1999; Van Knippenberg and Sitkin, 2013). Instead, we focus on collaboration, inclusion and building relationships as means to inspire and motivate staff (see, e.g., Carter et al., 2012). This focus also means that as our perspective includes dimensions of practices that need to be specifically addressed in school-based interventions, it may also demand a stronger emphasis in this context compared to business as usual. An example may be transactional aspects that, by our definition and consistent with, e.g. the Burke-Litwin model, emphasise the need to secure adequate resources for the intervention, to plan and clarify the work and to ensure progress in the project. Hence, we have chosen to include these aspects of transactional leadership while excluding others (see Bass, 1999).

As part of the examination of the construct validity of transformational and transactional leadership, we investigate how these constructs relate to other aspects of a school’s capacity to improve. Based on research as well as theory, it is reasonable to assume that innovation climate and teacher affiliation are positively related to the two aspects of leadership. Leadership that promotes a shared vision and collaborative decision-making implicitly promotes a positive innovation climate (e.g. Gil et al., 2005; McMurray et al., 2010). By fostering motivation and a collective orientation towards group performance, transformational leadership can help build a collaborative working environment that supports high performance (McMurray et al., 2012). A climate that supports innovation is thus related to leadership in that leadership is described by several researchers as characteristic of an innovative school (e.g. Bülbül, 2012; Cormican and O’Sullivan, 2004; Smith et al., 2008; Tidd et al., 2005). Moreover, Moolenaar et al. (2011) found that the more densely connected teachers are with respect to work-related and personal issues, the more they perceive their school climate to be supportive of innovation. The results suggest that the positive relationship between density of work-related advice networks and innovation-supportive school climate could be partially explained by increased shared decision-making that may well be promoted by transactional leadership in an innovative school. Given the relationship between leadership and innovation climate in schools, it is reasonable to assume that leadership may also be associated with teachers’ perceptions of affiliation among colleagues. It is also reasonable to expect that transformational and transactional leadership styles are positively related to the measurement of principal leadership as leadership style has been found to be related to teacher affiliation and school innovation climate. Staff freedom is expected to be negatively associated with leadership, given that the measurements tap the degree to which teachers are free of set rules, guidelines and procedures as well as the supervision necessary to ensure compliance with the rules (Fisher and Fraser, 1991). Schools that lack a collective orientation, and/or have an individualistic orientation to change will, most likely, struggle to implement interventions (e.g. Leithwood et al., 2008) Previous studies regarding transformational and transactional leadership styles (e.g. Bycio et al., 1995; Judge and Piccolo, 2004) indicate that measures of these two aspects of leadership are highly correlated. Based on this high correlation and the rationale presented above, it is expected that our transformational and transactional leadership scale may be positively correlated to the different organisational factors discussed herein, apart from staff freedom which is expected to correlate negatively.

Instruments to measure leadership

The literature presents a number of leadership survey instruments developed specifically with the aim of measuring transformational and transactional leadership (Leithwood and Sun, 2012). Some instruments measure elements of both transformational and transactional leadership, such as the Multifactor Leadership Questionnaire (MLQ) (Bass and Avolio, 1995), the Transformational Leadership Behavior Inventory (Podsakoff et al., 1990) and the Nature of School Leadership Survey (NSL) (e.g. Leithwood et al., 2001), while others, such as the Leadership Practices Inventory (LPI) (Kouzes and Posner, 1995) and the Leadership Behavior Questionnaire (LBQ) (Sashkin, 1990), only focus on the transformational leadership factor. The most commonly used instrument to measure the various aspects of transformational and transactional leadership is the MLQ. Despite its popularity, however, questions have been raised about the factor structure of the transformational model and about the discriminant validity of the components of the model with each other (Avolio et al., 1999; Bycio et al., 1995; Carless, 1998). As a consequence of the mixed empirical support for the transformational model, several authors have suggested that the individual components of the model actually measure one higher-order factor (Carless, 1998; Conger and Kanungo, 1998; Tepper and Percey, 1994). Hence, it may seem that the transformational model fails to measure transformational leadership behaviours and instead measures one single hierarchical construct of transformational leadership. Conger and Kanungo (1998) further argue that transformational leadership behaviours are interrelated and interconnected and therefore conceptually indistinct. Taken together, the evidence basically indicates that there is no point in using individual subscale scores when, in practice, teachers are not able to distinguish between these transformational leadership behaviours. In addition to the measurement problems related to the MLQ, the short version of the MLQ includes 45 items (MLQ-5X Short). Other instruments, such as the NSL (Leithwood et al., 2001) and the Transformational Leadership Behavior Inventory (Podsakoff et al., 1990), are also fairly comprehensive as they are specifically designed to measure only the construct of leadership. Especially in studies of complex interventions where several aspects of a school’s capacity are being assessed, it is preferable to make use of a leadership questionnaire that is not too comprehensive or time consuming. Therefore, as the measures that already exist may be too exhaustive for our use, our aim will be to include what we believe, based on our review of the literature, to be the most important and distinguishing practices related to the two leadership styles. Although our measure will not account for all dimensions of the concepts of transformational and transactional leadership, it will provide us with some valuable information regarding the type of leadership that is being practised at a specific school. This information, in combination with data collected from other measures in our framework such as school culture, school innovation climate and teacher motivation, may expand our understanding of a school’s capacity to change and improve.

A third reason for not using the MLQ is that transactional leadership, as defined and measured by Bass (1998), is quite negatively charged compared to how it is conceptualised in our theoretical framework as defined by Burke and Litwin (1992). We agree with Burke (2008: 235) when he notes that ‘an organization can survive for some period of time (not long, however) without leadership but can hardly survive at all without management’. We regard transformational and transactional leadership to be equally important, as an imbalance between the two leadership styles may produce unsuccessful change. For example, simply because a principal provides directions does not mean that action will follow, in the same way that action without direction will most likely fail to lead in the right direction. Although both leadership styles are necessary, problems may occur if they are not appropriately balanced to suit the individual school context and situation in which they are implemented. Too much emphasis on transactional leadership may discourage risk taking, while too much emphasis on transformational leadership may result in chaos and impractical change (Yukl and Lepsinger, 2005). Contrary to this argument, Bass and Avolio suggest, based on a summary of research evidence, that transformational leadership is superior to transactional leadership (Avolio et al., 1999; Bass, 1998). We understand how this may be the case with respect to the two components of Bass’s (Bass and Avolio, 2000) transactional leadership factor: contingent reward and management-by-exception. According to Bass’s definition, contingent reward is based on the leader clarifying what is expected from followers and what the followers will gain if they meet the expectations, while management-by-exception focuses on monitoring the performance of tasks for problems that may occur and correcting these problems to maintain current performance levels (Avolio et al., 1999). Hence, contingent reward is related to constructive and positive exchanges with followers, while management-by-exception is related to a corrective orientation towards exchanges (Avolio et al., 1999). A number of studies have found high correlations between contingent reward and transformational leadership, indicating that contingent reward is part of a transformational leadership dimension (Judge and Piccolo, 2004; Leithwood et al., 1996). Based on these notions, we have chosen to include contingent reward as part of our transformational leadership dimension. Furthermore, contingent reward, as conceptualised in the MLQ, has been found to work better in business settings than in organisations such as schools (Judge and Piccolo, 2004). The reason for this may be that leaders of business organisations have access to more resources in the form of tangible bonuses in exchange for completed tasks than do school leaders. Hence, based on our perception of contingent reward, we have chosen to focus less on these types of tangible rewards and more on school leaders providing positive feedback and showing appreciation for individual contributions when tasks are completed.

Methods

Sample and procedure

Data used for this study were gathered from a study of teachers at schools implementing a school development intervention aimed at preventing and reducing problem behaviour. Baseline data were used in the development and testing of the measurements. A web-based questionnaire was administered in May, before the school implemented the school development intervention programme, which was introduced in the beginning of the following school year. Consent was obtained by voluntary participation based on a written description of the project according to standards prescribed by the Norwegian Data Inspectorate. Teachers (N = 1144) from 45 Norwegian primary, secondary or combined (both primary and secondary) schools participated in the study.

Measurement development

The scales were developed on the basis of the theoretical framework and empirical evidence previously outlined herein. As part of a larger study aiming to explore school readiness and capacity to improve, 27 items were previously developed for a Q-study that investigated school staff perceptions of transformational and transactional leadership styles (see Otterkiil, 2010 for details). In essence, we have developed our leadership measures based on the four core leadership practices presented by Leithwood and colleagues (2008) and adjusted them to fit with our adaptation of the Burke-Litwin model of organisational change with the aim of providing a set of representative items that describe leadership practices essential to schools that are in the process of implementing school-based interventions. When developing items for the Q-study, which later was applied to the current study, we aimed to address three subcategories of transformational leadership as described by Leithwood and Janzti (2006). We specifically drew on this work as their model of transformational leadership was related to interventions, although within a field different than behaviour. Additionally, we applied the perspectives of the Burke-Litwin model of organisational change, as previously discussed. In particular, we addressed their perspectives on leadership (transformational factor) and management (transactional factor). The 27 statements represented the following three transformational and one transactional category: Building Vision and Setting Directions (seven items), Understanding and Developing People (eight items), Redesigning the Organisation (six items) and Managing the Teaching and Learning Programme (six items). Acknowledging the complexity of aspects related to the three broad categories included in Leithwood and Jantzi’s (2006) study, we had no intention of covering all three categories in the same thorough way as did Leithwood and Jantzi. On the contrary, we intended to develop a shorter measurement that taps some of the key aspects addressed by Leithwood and colleagues, while at the same time acknowledging the need to address aspects of transactional leadership consistent with the Burke-Litwin model of organisational change. As discussed herein, one of the main reasons for the need for a shorter measurement is that existing measurements, such as the NSL, are fairly comprehensive. In an assessment system developed to address a wide range of key aspects of a school’s capacity to implement school-based interventions, 42 items for subcategories that addressed only leadership may be too extensive. In comparison, the Burke-Litwin questionnaire contains 150 items that address all 12 factors (see Burke and Litwin, 1992; Burke et al., 1996).

Although aware of potential problems connected to the identification of subscales for the three aspects of transformational leadership, we initially set out to validate the items aimed at investigating the possibility of developing subcategories of transformational leadership. To investigate content validity, a questionnaire was distributed to 250 teachers from three schools and to a panel of five experts within the field of educational leadership. We asked the participants to rate how well every item fitted within each of the four categories according to a given theoretical definition. The results echoed the findings previously discussed, thus suggesting that both teachers and members of the expert panel had trouble placing the items according to the theoretical model on which they were based. These findings indicated that both teachers and experts had difficulty separating the three different aspects of transformational leadership.

Innovation culture, staff freedom and teacher affiliation were measured on subscales using the School-Level Environment Questionnaire (SLEQ) (Fisher and Fraser, 1991). Principal leadership was measured on a nine-item scale that focused on key groups’ trust in the principal, as well as on whether the principal or informal leaders were the true leaders of the school. All items were scored on a scale ranging from zero (totally disagree) to five (totally agree).

Data analysis

Given the initial findings from the previous studies that elaborated on teachers’ understanding of the concepts, a major aim of this paper was to investigate transformational and transactional leadership as two distinct but interrelated concepts. Although the initial investigation of content validation indicated that teachers and experts had problems separating the three aspects of transformational leadership, a sample of schools that implemented a school-based intervention enabled us to further investigate whether the data supported subcategories of transformational leadership, and thus a four-factor solution based on the 27 items. Initially, the number of subscales in the 27 items was tested through exploratory factor analysis (EFA). Based on the findings of the EFA, a confirmatory factor analysis (CFA) was applied for further testing.

To perform the EFA and CFA in different samples, as suggested by Gerbing and Hamilton, 1996, and Hurley et al., 1997, the SPSS version 18.0 was used to draw two randomly selected samples of approximately 50% of the teachers for each. The first sample was used for the EFA and the second for the CFA. Teachers were nested within schools, thus enabling multi-level testing of the measurements. The model was fitted to the data by means of robust maximum likelihood procedures such as those implemented in Mplus 7.0 (Muthén and Muthén, 2012). The missing data method was applied, as this allowed for the use of all observations in the data set when estimating the parameters in the models. Conventional analyses were conducted using SPSS version 18.0.

In their recommendation for goodness of fit indices, Hu and Bentler (1999) suggested using a cut-off value close to .08 for the standardised root mean squared residual (SRMR) and supplementing it with indices such as the Tucker-Lewis index (TLI) and the comparative fit index (CFI) with cut-off values close to .95. They also recommended the inclusion of the root mean square error of approximation (RMSEA) with a cut-off value of approximately .06 or less. The RMSEA is supported by a 90% confidence interval (90% CI). Due to the relatively large number of subjects, traditional χ²tests may provide an inadequate assessment of model fit. Given this, the χ² test is included but not discussed. As recommendations for goodness of fit indices for multi-level modelling have not been developed, we apply the suggested cut-offs for single-level analyses.

Results

Descriptive statistics

Descriptive statistics for items from the transformational and transactional scales are presented in Table 1, including the means, standard deviations, values for skewness and for kurtosis. Muthén and Kaplan (1992) and Curran et al. (1996) suggested that there might be significant problems in estimation when univariate standardised skewness has an absolute value of 2.0 or greater, when standardised kurtosis has an absolute value of 7.0 or greater, or when both are true. Skewness and kurtosis for all items were well within these values. Descriptive statistics for the four items constituting transformational leadership and the four items constituting transactional leadership are provided for both the teacher level and the school level with respect to the EFA and the CFA subsamples (see Table 1). The school level results indicate widespread differences in mean school scores for the eight items. Interestingly, at both the teacher and the school levels, the scores for the transactional leadership items were lower than the scores for the transformational leadership items.

Table 1.

Item means (M), standard deviation (SD), Skwenes s (Ske), Kurtosis (Kurt) and Cronbach’s alpha for items for Transactional and Transformational Leadership.

	Sample EFA								Sample CFA
	Teacher Level (N = 585)				School Level (N = 45)				Teacher Level (N = 580)				School Level (N = 45)
Item	M	SD	Ske	Kurt	M	SD	Ske	Kurt	M	SD	Ske	Kurt	M	SD	Ske	Kurt
Transactional (α = .87–.87) Someone at our school exercises leadership by:
1. Making sure that the individual staff member is given clear instructions of what their responsibility is regarding a specific project/assignment.	2.90	1.23	–0.44	–0.26	2.89	0.61	–0.20	0.89	2.91	1.19	–0.48	–0.09	2.90	0.64	–0.66	0.34
2. Working out clearly written plans on how projects can be carried out.	2.69	1.26	–0.31	–0.43	2.68	0.56	–0.54	2.43	2.68	1.28	–0.31	–0.46	2.68	0.65	–0.68	0.95
3. Making sure that our work is progressing.	2.87	1.22	–0.51	–0.09	2.87	0.65	–0.56	1.52	2.92	1.20	–0.53	–0.06	2.91	0.66	–0.98	0.70
4. Making sure there are adequate resources to carry out planned tasks in a satisfactory way.	2.68	1.19	–0.35	–0.50	2.68	0.53	0.46	–0.03	2.68	1.18	–0.46	–0.30	2.68	0.55	–0.71	0.35
Transformational (α = .88–.88) Someone at our school exercises leadership by:
5. Giving me support through the appreciation of my contribution to the school.	3.26	1.27	–0.75	0.10	3.25	0.70	–0.43	0.32	3.36	1.25	–0.75	0.12	3.35	0.70	–0.25	–0.53
6. Encouraging a supportive atmosphere among the school’s staff.	3.01	1.17	–0.51	0.24	3.00	0.52	–0.52	0.55	3.04	1.18	–0.63	0.11	3.04	0.54	–0.92	0.63
7. Arranging so that challenges can be discussed constructively among colleagues.	3.18	1.34	–0.66	–0.21	3.18	0.59	–0.29	0.10	3.25	1.32	–0.66	–0.26	3.24	0.55	–0.26	–0.52
8. Involving staff in debates concerning the school’s goals and visions.	3.35	1.25	–0.75	0.16	3.34	0.66	–0.58	1.13	3.42	1.23	–0.73	0.16	3.41	0.67	–0.98	1.39

Note. M = item means; SD = standard deviation; Ske = Skweness; Kurt = Kurtosis; α = Cronbach’s alpha for scales measuring transformational and transactional leadership. Rating format is 0-5, where 0 = never and 5 = very often.

Exploratory factor analysis

To achieve psychometrically sound measurements that hold under cross-validation, EFA was used as a first empirical step. The primary goal of EFA is to reduce a larger set of measured variables to a smaller set. Our goal was to investigate whether the hypothesised subscales for transformational leadership are actually reflected in the collected data. We further aimed to reduce the set of 27 items so that those items that best represented transformational and transactional leadership without loading too high on one or more of the other dimensions would remain. It should be noted, however, that even with EFA, theory remains an important guideline when determining which items to keep and which items to remove (Henson and Roberts, 2006). To check the conditions for a stable factor structure, the sample size must be large enough (Ferguson and Cox, 1993). The two most frequently used criteria are the absolute minimum number of subjects and the relative number indicated by the subject-to-variables ratio (Van Dierendonck and Nuijten, 2011). While the minimum number of subjects ranges between 100 and 300, Guadagnoli and Velicer (1988) suggest that 100 subjects may be acceptable. Ferguson and Cox (1993) propose that the generally accepted minimum for reaching a stable factor structure is a subject-to-variable ratio of 5:1. Osborne and Costello (2004) argue, based on their results, that a combination of the two criteria, a high number of subjects and a high ratio, will provide the best outcome. With 585 teachers in our EFA sample and a subject-to-variable ratio of 21.67:1, both criteria are met.

Teachers were asked to rate the quality of leadership in their schools, according to the 27 items previously discussed. Design effects (2.5 to 3.7) were well above the 2.0 rule of thumb (Muthén, 1997) on all items, thus indicating that a two-level structure was appropriate. However, to further investigate the factor structure and to shed further light on the measurement, we also included one-level models.

Our initial EFA did not support a four-factor solution with three subcategories of transformational leadership and one factor representing transactional leadership. A one-level model provided a fair to good fit for most fit indices (χ² (249) = 907.72, p = 0.000, CFI = 0.95, TLI = 0.94, RMSEA = 0.067; 90% CI = 0.063 – 0.072). There was, however, no support for a multi-level model with four factors at either the within or between levels. A model containing two factors at the between and four factors at the within level provided a good fit (χ² (547) = 980.95, p = 0.000, CFI = 0.97, TLI = 0.96, RMSEA = 0.037; SRMR within = 0.024; SRMR between = 0.067). However, the items failed to load on the intended factors. Given this finding, there was no support for a four-factor model. That is, for the multi-level model, four of the transactional items loaded fairly strong to strong on one factor (0.94–0.50). Moreover, several of the items from all three subcategories of transformational leadership loaded strong on another factor. However, many of the 27 items had strong cross-loadings between two or more factors. Thus, the results support the content validity analysis, indicating that teachers failed to rate the items according to our theoretical model that consisted of three aspects representing transformational leadership and one aspect representing transactional leadership. Therefore, in our further investigation of the factor structure, we aimed at identifying one scale for transformational leadership that consisted of all three aspects of the concept, as well as one scale for transactional leadership. We performed further EFA on the four items that represented transactional leadership and the six items that represented transformational leadership. For transformational leadership, two items represented the subcategory ‘Building Vision and Setting Directions’, two items represented ‘Understanding and Developing People’, and two items represented ‘Redesigning the Organisation’. For all three subcategories of transformational leadership, we initially intended to choose the two items that loaded the strongest on one common factor. After careful consideration, however, we chose not to include one of the two items representing ‘Redesigning the Organisation’ that loaded strongest on the transformational leadership scale. One of the items was replaced with a slightly weaker one, as it did not fit nearly as well with our theoretical model as did the item with a slightly weaker factor loading. The EFA model composed of six transformational and four transactional items provided a good fit both for a one-level model (χ² (26) = 106.83, p = 0.000, CFI = 0.98, TLI = 0.97, RMSEA = 0.073; 90% CI = 0.059–0.088; SRMR = 0.018) and a multi-level model (χ² (61) = 146.84, p = 0.000, CFI = 0.98, TLI = 0.97, RMSEA = 0.049; SRMR within = 0.024; SRMR between 0.095) for most indices. One of the items for the subcategory ‘Building Vision and Setting Directions’ loaded 0.47 on the transformational factor and 0.37 on the transactional factor. The item was removed due to this cross-loading as the loading on one factor should be twice as large as the loading on another factor, according to a suggested strict rule for rejecting the existence of a cross-loading. None of the other items representing this category loaded strong enough on the transformational leadership scale to replace the rejected item. Furthermore, based on recommendations from the modification indices and careful theoretical consideration, one of the items representing ‘Understanding and Developing People’ was removed. This left us with a four-item scale for transformational leadership (see Table 1 for items).

Factor loadings for both within and between school levels for the final EFA model, consisting of four items on transformational and transactional leadership on within school level and one leadership factor on between school level, are presented in Table 2.

Table 2.

Exploratory Factor Analysis for the two within (teacher) and one between (school) level model for Transactional and Transformational Leadership. Loadings and cross loadings between the factors at teacher level. Standardized metric.

	Teacher Level		School Level
	Transactional	Transformational	Leadership
1. Making sure that the individual staff member is given clear instructions of what their responsibility is regarding a specific project/assignment.	0.77	0.01	0.99
2. Working out clearly written plans on how projects can be carried out.	0.90	–0.19	0.94
3. Making sure that our work is progressing.	0.74	0.12	1.00
4. Making sure there are adequate resources to carry out planned tasks in a satisfactory way.	0.47	0.23	0.81
5. Giving me support through the appreciation of my contribution to the school.	–0.07	0.87	0.94
6. Encouraging a supportive atmosphere among the school’s staff.	0.01	0.80	0.86
7. Arranging so that challenges can be discussed constructively among colleagues.	0.11	0.64	0.97
8. Involving staff in debates concerning the school’s goals and visions.	0.15	0.62	0.85

χ²(33) = 54.12, p =., CFI =.99, TLI = .99, RMSEA = .033; SRMR within = .016 and SRMR between = .103.

Notes. CFI = Comparative Fit Index; TLI = Tucker Lewis Index; RMSEA = Root Mean Square Error of Approximation; SRMR = Standardized Root Mean Squared Residual.

The model provided a good to acceptable fit for the different indices (χ² (33) = 54.12, p = 0.012, CFI = 0.99, TLI = 0.99, RMSEA = 0.033; SRMR within = 0.016; SRMR between 0.103). It is not uncommon for fewer factors to be found between levels than within levels. Moreover, a one-factor solution at both the teacher and the school levels was not supported by the data, as it failed to meet most of the suggested fit indices (χ²(40) = 358.54, p = 0.000, CFI = 0.89, TLI = 0.84, RMSEA = 0.117, SRMR within = 0.063, SRMR between = 0.103). The SRMRs for the between level, the two-factor at the within and the one-factor models were slightly higher than the recommended cut-off. Despite this effect, we retained the model. Table 3 provides information about the different EFA models tested.

Table 3.

Fit Indices for Models of Transformational and Transactional Leadership.

	χ2(df)	CFI	TLI	RMSEA	SRMR (within)	SRMR between
EFA models
One-factor/single-level	355.16 (20)	.89	.85	.017	.055
Two-factor/single- level	22.61 (13)	1.00	.99	.036	.010
One-factor/two-level	358.54 (40)	.89	.84	.117	.063	.103
Two-factor/two- level
Two within- one between	54.12 (33)	.99	.99	.033	.016	.103
CFA models
Transformational Leadership
One-factor/single-level	2.35 (2)	1.00	.99	.017	.006
One-factor/two-level	1336.85 (12)	.99	.99	.054	.015	.006
Transactional Leadership
One-factor/single-level	6.92 (2)	1.00	.99	.065	.010
One-factor/two-level	862.72 (18)	1.00	.99	.030	.011	.028
Combined models
One-factor/single –level	201.04 (20)	.91	88	.151	.049
Two-factor/single-level	43.12 (19)	.99	.99	.047	.022
Two-factor within/one factor between	76.95 (40)	.99	.98	.040	.027	.065

Note. χ2(df) = Chi square with degree of freedom (df); CFI = Comparative Fit Index; TLI = Tucker-Lewis Index; and RMSEA = Root Mean Squared Error of Approximation; SRMR = Standardized Root Mean Squared Residual.

Consequently, we reduced the number of items from 27 to eight, with four items used to measure the construct of transformational leadership and four items used to measure transactional leadership. Given this, transformational and transactional leadership were eventually measured by teacher reports using two scales, each of which consisted of four items. The eight items and the descriptive statistics for each are presented in Table 1.

Confirmatory factor analysis

After obtaining a two-factor solution on within and a one-factor solution on between levels with EFA, we aimed to replicate the factor structure with a new subsample of data and a more restrictive and powerful analytical tool – CFA. Similar to EFA, we also tested single-level models. Moreover, although not strictly necessary given the results of the EFA, we independently tested the measurements for transformational and transactional leadership, to ensure that they provided a good fit. The results for the independent goodness of fit tests for the transformational and transactional scales are provided in Table 3. First, a one-factor model where all items loaded on one leadership factor was tested. The one-factor model did not meet the suggested criteria for model fit (χ² (20) = 201.04, p = 0.000, CFI = 0.91, TLI = 0.88, RMSEA = 0.151; 90% CI = 0.136–0.167, SRMR = 0.049). The one-factor model was tested to rule out the possibility that this model fit the data better than the hypothesised two-factor model. A high factor correlation was identified (r = 0.81) in the two-factor model. The initial two-factor model showed acceptable fit to the data for subsample 2 (χ² (19) = 43.12, p = 0.001, CFI = 0.99, TLI = 0.99, RMSEA = 0.047, 90% CI = 0.028-0.065, SRMR = 0.022), suggesting that the hypothesised two-factor model was supported by the data. Given that the respective CIs associated with the RMSEA for the one- and two-factor structures did not overlap, a two-factor solution was indeed a better representation of the data.

The model’s standardised and unstandardised solutions for subsample two are presented in Table 2. As outlined herein, clustered sampling, i.e. teachers within schools, enabled the testing of a two-level model. This model was tested in subsample 2. Design effects greater than two on all items and intra-class correlations (ICCs) exceeding .10 for all items indicated a two-level structure (Muthén, 1997).

The indices in the estimated two-level model revealed a good fit for the observed data for subsample 2 (χ²(40) = 76.95, p = 0.000, CFI = 0.99, TLI = 0.98, RMSEA = 0.040; SRMR within = 0.027 and SRMR between = 0.065). The model’s standardised and unstandardised solutions for subsample two are provided in Table 4. Residual variances were checked for all models tested, and none of them were negative. Accordingly, they were not further elaborated on. Given the results presented above, we find support for transformational and transactional leadership as two interrelated but distinct dimensions of leadership. The results support both a single-level and a two-level model.

Table 4.

Confirmatory Factor Analysis with Latent Variables Transactional and Transformational Leadership. Factor Loadings in Unstandardized (Unst) and Standardized (St) Metric.

	Teacher level		School Level
	Unst/St	Unst/St	Unst/St
1. Making sure that the individual staff member is given clear instructions of what their responsibility is regarding a specific project/assignment.	1.00/.78		.98/.98*
2. Working out clearly written plans on how projects can be carried out	1.01/.73*		.98/.97*
3. Making sure that our work is progressing.	1.08/.84*		1.09/1.00*
4. Making sure there are adequate resources to carry out planned tasks in a satisfactory way.	.95/.71*		.67/.80*
5. Giving me support through the appreciation of my contribution to the school.		1.00/.82*	.80/.99*
6. Encouraging a supportive atmosphere among the school’s staff.		.92/.84*	1.03/.94*
7. Arranging so that challenges can be discussed constructively among colleagues.		.81/.75*	.85/1.00*
8. Involving staff in debates concerning the school’s goals and visions.		.81/.75*	1.01/.90*

χ2(40) = 76.95, p = .000, CFI = .99, TLI = .98, RMSEA = .040; SRMR within = .027 and SRMR between = .065

Notes. CFI = Comparative Fit Index; TLI = Tucker Lewis Index; RMSEA = Root Mean Square Error of Approximation; SRMR = Standardized Root Mean Squared Residual; *<.001. Note that the first path from each factor is fixed to 1.

External validity

As part of the validation, scores were related to other aspects relevant to organisational capacity. From the results presented in Table 5, it is evident that, as expected, the transformational and transactional leadership scales both correlated significantly negatively with staff freedom and positively with innovation climate, affiliation and the principal’s position at school, both at the teacher and the school levels. Staff freedom, innovation climate and affiliation are subscales of the school environment questionnaire (Fisher and Fraser, 1991).

Table 5.

Correlations (Spearmans’ ρ) Between Transformational leadership, Transactional leadership and Four Aspects relevant for Schools’ Capacity to improve in Sub-Sample 2.

	Teacher Level (N = 580)		School Level (N = 45)
	Transformational	Transactional	Leadership
Innovation climate	.38***	.39***	.47***
Affiliation	.45***	.27***	.44***
Staff Freedom	−.24***	−.28***	−.34***
Principals’ position	.56***	.50***	.76***

***p < .001

Discussion

The aim of this study was to investigate the concept of leadership through teachers’ perceptions regarding the two aspects of transformational and transactional leadership styles. Leadership is an imperative factor in any change effort within a school, and as discussed herein, it is a key aspect of a school’s capacity to improve. Given this fact, measuring transformational and transactional leadership may be an important aspect in identifying a school’s capacity to implement a school-based intervention.

The results provide support for transformational and transactional leadership at the school level. Using EFA and CFA, a measurement structure with two factors at the teacher level and one factor at the school level was found. Although there is some overlap between transactional and transformational leadership, the two concepts do measure distinct features. Accordingly, this structure reveals a better approximation than the rival one-factor model. There was, however, no support for a four-factor solution with three subcategories for transformational leadership and one for transactional leadership. Additionally, there was support for a two-factor solution as a single-level model. The two factors were also found to be highly reliable according to the Cronbach alpha coefficients.

The unit of analysis has been an ongoing area of concern in school-based research. From both a statistical and theoretical perspective, it seems that multi-level modelling is the optimal approach to handling the hierarchical data in the current study. The data are clearly hierarchical as teachers or students are nested within schools. Furthermore, intra-class correlation and design effects support the application of a multi-level analysis in the current study. The intra-class (or intra-unit) correlation is often used to indicate the proportion of variance accounted for by higher-level units. With teachers nested within schools, this coefficient is an intra-school correlation. However, as Muthén (1997) emphasises, it is not enough to consider only the ICC as this does not take into account the size of the group. From a statistical perspective, if the data are to be analysed using teachers as the unit of analysis, an effective sample size that takes into account the design effect of having teachers nested within schools can be employed (Snijders and Bosker, 2012). The higher the ICC, the higher the design effect, and thus, the greater the adjustment in the sample size if an analysis is conducted only at the individual level (Dorman, 2009).

However, researchers of school organisational factors often ignore the hierarchical nature of data (e.g. Ellett and Monsaas, 2011; Thoonen et al., 2011). Various reviews, e.g. of schools’ learning environments, over the past 20 years have recognised this issue as important because the inappropriate analysis of data can invalidate results and subsequent conclusions (Dorman, 2002, 2009; Fraser, 2007). Given the previous outline of transformational and transactional leadership styles, it seems clear that there may be a school effect due to the differences in ethos and identities within schools. In essence, the variance in teachers’ scores can be segmented at the teacher and school levels. Teachers influence and are influenced by other teachers within the school. The dominant methodology in school-based research has been, and remains, the collection of perceptual data from school personnel, including administrative staff, teachers and/or students, who are nested in a particular school setting. Traditional studies that have used teacher perceptual data, have used the teacher as the unit of analysis and have ignored school membership, or have used the school as the unit of analysis with aggregated data and thus ignored the individual teacher. A possible reason is the logistical difficulties in accessing large samples of schools. Proponents of multi-level modelling warn against ignoring the existence of groupings of hierarchical data due to the danger of statistical and interpretational errors (Rowe, 2007; Tabachnick and Fidell, 2007). Clearly, multi-level modelling is the optimal approach to handling hierarchical data. However, there are practical, logistical, and statistical issues associated with this approach.

Dorman (2009) argues that because of logistical constraints, researchers tend to gather school environment data from as many teachers as possible within the volunteer schools. That is, while teachers are the prime source of the data, the approval to collect the data is initially given at the school level. Given this practical reality, there is clear potential for inflated Type I error rates. An implication of such an error is that closer attention must be paid to research design and to the selection of samples when conducting research on leadership as well as on other organisational aspects of school settings. In the current study, the implementation of a school-based intervention in 45 schools provided us with an opportunity to test the measurement of transformational and transactional leadership. Accordingly, the study design for the research and the data collection were not designed for the current study. The 45 schools are, however, well within the 30 to 50 number of units typically suggested by textbooks as a minimum number for, e.g., school-level analysis. It should be noted that this number of schools allows for a less complex analysis, as the number of clusters (schools) limits the number of parameters that can be estimated. This observation illustrates that the number of schools in a study applying these measurements should, ideally, contain a significantly higher number of schools, thus allowing for more complex organisational models. To investigate a complex theoretical model, one needs a large sample of schools. For example, our model for capacity and readiness containing 12 factors (Otterkiil and Ertesvåg, 2012) will need more than 100 schools, depending on the number of items and scales, to estimate latent variable two-level models. These practical challenges are illustrated by the fact that we develop the measurement for use in studies of capacity and readiness for improvement. Implementing a complex school-wide intervention to improve student learning underscores methodological challenges related to school-based research, especially with respect to intervention research. As Dorman (2009) discusses, researchers must pay more attention to research design, in particular, to the selection of the sample.

The high substantive correlation between transformational and transactional leadership is an issue that needs attention as our study, as well as several other studies (e.g. Bycio et al., 1995; Judge and Piccolo, 2004), have found that the constructs of transformational and transactional leadership show substantive correlations. Although theoretically discrete, the empirical discriminant validity of the two leadership constructs is difficult to establish. The cause for this may be the substantive overlap that naturally occurs between the two constructs. Another reason could be the shared methods often used to measure the two constructs, such as leaders’ behaviours as rated by employees in one questionnaire, which result in same-source same-method bias (Kruger et al., 2011). One possible solution to this issue could be to make use of the multi-trait/multi-method (MTMM) approach (Campbell and Fiske, 1959) to separate and attribute the different sources to the shared variance. For instance, as a result of using this technique, Kruger and colleagues (Kruger et al., 2011) were able to empirically discriminate scales measuring aspects of the theoretically discrete constructs of transformational and transactional leadership styles. The results of their MTMM analyses indicated that when controlling for the method effects, the correlations between the leadership constructs decreased from a high to a small respectable medium level and the leadership scales could be discriminated. However, one of the reasons for the decreased correlations related to the use of this method may simply be due to the use of different data sources, e.g. leaders and followers, thus causing us to question whether this solves the problems, as the reduction in correlation seems to be the result of different opinions or experiences of different groups of people. Clearly, this question requires further investigation of both the theoretical and methodological aspects.

The strengths of this study lie in the statistical methodology used (EFA, CFA), in the use of different samples for the EFA and CFA tests and in the hierarchical nature of the data considered. However, the study has certain limitations that should be addressed. The limitations of the research design, especially the relatively low number of schools, have already been addressed. Furthermore, the measurement would benefit particularly from an investigation of whether it is invariant over time in a longitudinal sample. If the model is not invariant in such a sample, this may be problematic as one may be in danger of measuring concepts other than transactional and transformational leadership at different points in time rather than measuring change with respect to the two aspects of leadership (Ertesvåg, 2011). Future studies should address measurement invariance in longitudinal samples. In school improvement initiatives, it may be imperative to investigate whether the interventions have the expected outcomes. A call for more evidence based professional development also calls for measurements that have sound theoretical and empirical groundings to evaluate them.

Finally, we note that school leadership is a complex phenomenon and acknowledge that no matter how broad or extensive our measures are, we will never be able to capture every aspect of school leadership. It could, however, be questioned whether the developed measurement is extensive enough to capture such a complex phenomenon. When assessing a school’s readiness and capacity to implement an intervention, a mixed-methods approach for measuring leadership may prove to be effective. We suggest that the measurement developed herein could be used in combination with a Q-methodology study that includes the 27-item Q-sample discussed earlier (see Otterkiil, 2010). While this measure informs us how school staff rate their leaders, the Q-study provides information about what school staff expect and need from their leaders. Taken together, these two types of assessments may provide valuable contextual information about where efforts should be focused to improve leadership within a particular school.

Because our aim is to assess a number of organisational factors, not solely leadership, our challenge lies in identifying the very most essential set of items that represent the key leadership practices that will lead to a fragmentary but realistic indication of the leadership that exists within a particular school. Further testing is necessary before we can fully establish that the measurement is a reliable measurement for transformational and transactional leadership in school-based interventions. We further underscore that the purpose of this study is not to replace the more extensive measurements of transformational and transactional leadership but to develop a measurement suitable for assessment when implementing a school-based intervention. With respect to professional development, it may be imperative to investigate several aspects of a school’s capacity to improve (Oterkiil and Ertesvåg , 2012), and existing measurements are often too extensive for this purpose. The theory and empirical evidence presented in this paper thus shed light on the importance of leadership and how leadership may motivate teachers to engage in change efforts and to bring these changes into their classrooms. Being able to identify what teachers expect and need from their leaders may be one of the key factors in understanding what motivates teachers to engage in change initiatives and improve classroom practices. Furthermore, providing leaders of struggling schools with training and guidance on how to strengthen their ability to effectively balance transformational and transactional leadership styles (according to the specific context of their school) may provide these leaders with valuable information regarding where to focus their efforts to move their schools forward.

Footnotes

Conflict of interest

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Avolio

Bass

Jung

(1999) Re-examining the components of transformational and transactional leadership using the Multifactor Leadership Questionnaire. Journal of Occupational and Organizational Psychology 72: 441–462.

Bass

(1985) Leadership and Performance Beyond Expectations. New York, NY: Free Press.

Bass

(1998) Transformational Leadership: Industrial, Military, and Educational Impact. Mahwah, NJ: Erlbaum.

Bass

(1999) Current developments in transformational leadership. The Psychologist Manager Journal 3(1): 5–21.

Bass

Avolio

(1994) Improving Organizational Effectiveness Through Transformational Leadership. Thousand Oaks, CA: Sage.

Bass

Avolio

(1995) Multifactor Leadership Questionnaire: Duplicate Set: Rate Form, for MLQ for 5x-short. California: Mind Garden.

Bass

Avolio

(2000) Multifactor Leadership Questionnaire: Technical Report (2nd edn). California: Mind Garden.

Bülbül

(2012) Developing a scale for innovation management at schools: A study of validity and reliability. Educational Sciences: Theory and Practice 12(1): 168–175.

Burke

(2008) Organization Change. California, USA: Sage Publications.

10.

Burke

Litwin

(1992) A causal model of organizational performance and change. Journal of Management 18(3): 523–545.

11.

Burke

Coruzzi

Church

(Eds) (1996) The Organizational Survey as an Intervention for Change. San Francisco: Jossey-Bass Publishers.

12.

Burns

(1978) Leadership. New York: Harper & Row.

13.

Bycio

Hackett

Allen

(1995) Further assessments of Bass’s (1985) conceptualization of transactional and transformational leadership. Journal of Applied Psychology 80: 468– 478.

14.

Campbell

Fiske

(1959) Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin 56: 81–105.

15.

Carless

(1998) Assessing the discriminant validity of transformational leader behavior as measured by the MLQ. Journal of Occupational and Organizational Psychology 71: 353–358.

16.

Carter

Armenakis

Feild

Mossholder

(2013) Transformational leadership, relationship quality, and employee performance during continuous incremental organizational change. Journal of Organizational Behavior 34(7): 942–958.

17.

Conger

Kanungo

(1998) Charismatic Leadership in Organizations. Thousand Oaks, CA: Sage.

18.

Cormican

O’Sullivan

(2004) Auditing best practice for effective product innovation management. Technovation 24(10): 819–829.

19.

Curran

West

Finch

(1996) The robustness of test statistics to non normality and specification error in confirmatory factor analysis. Psychological Methods 1(1): 16–29.

20.

Dorman

(2002) Classroom environment research: Progress and possibilities. Queensland Journal of Educational Research 18: 112–140.

21.

Dorman

(2009) Statistical tests conducted with school environment data: The effect of teachers beeing clustered in schools. Learning Environment Research 12: 85–99.

22.

Ellett

Monsaas

(2011) Development and validation of a new measure of teacher perception of science and mathematics learning enviroment. Learning Enviroment Research 14: 91–107.

23.

Ertesvåg

(2011) Measuring authoritative teaching. Teaching and Teacher Education, 27,51–61.

24.

Ertesvåg

Roland

(2011) Leiing i implementering av skuleomfattande tiltak [Implementing school-wide interventions]. In: Midthassel

Bru

Ertesvåg

Roland

, (Eds.) Tidlig intervensjon og systemretta arbeid [Early intervention and systems work]. Oslo: Universitetsforlaget

25.

Ertesvåg

Roland

Vaaland

Størksen

Veland

(2010) The Challenge of continuation – Schools’ continuation of the Respect programme. Journal of Educational Change 11(4): 323–344.

26.

Ferguson

Cox

(1993) Exploratory factor analysis. A users’ guide. International Journal of Selection and Assessment 1: 84–94.

27.

Fisher

Fraser

(1991) Validity and use of school environment instruments. Journal of Classroom Interaction, 26(2): 13–18.

28.

Flaspohler

Duffy

Wandersman

. (2008) Unpacking prevention capacity: An intersection of research-to-practice models and community-centered models. American Journal of Community Psychology 41: 182–196.

29.

Fraser

(2007) Classroom learning environment. In: Abell

Lederman

(eds) Handbook of Research on Science Education. Mahwah, NJ: Erlbaum, pp.103–123.

30.

Gerbing

Hamilton

(1996) Viability of exploratory factor analysis as a precursor to confirmatory factor analysis. Structural Equation Modeling 3: 62–72.

31.

Gil

Rico

Alcover

. (2005) Change-oriented leadership, satisfaction and performance in work groups: Effects of team climate and group potency. Journal of Managerial Psychology 20(3): 312–328.

32.

Gingiss

Roberts-Gray

Boerm

(2006) Bridge-It: A system for predicting implementation fidelity for school-based tobacco prevention programs. Prevention Science 7(2): 197–207.

33.

Groves

(2005) Linking leader skills, follower attitudes and contextual variables via an integration of charismatic leadership. Journal of Management 31: 255–277.

34.

Guadagnoli

Velicer

(1988) Relation of sample size to the stability of component patterns. Psychological Bulletin 103: 265–275.

35.

Henson

Roberts

(2006) Use of exploratory analysis in published research. Common errors and some comment on improved practice. Educational and Psychological Measurement 66: 393–416.

36.

Bentler

(1999) Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling 6(1): 1–55.

37.

Hunt

Conger

(1999) From where we sit: An assessment of transformational and charismatic leadership research. The Leadership Quarterly 10(3): 335–343.

38.

Hurley

Scandura

Schricshcim

. (1997) Exploratory and confirmatory factor analysis: Guidelines, issues, and alternatives. Journal of Organizational Behavior 18: 667–668.

39.

Judge

Piccolo

(2004) Transformational and transactional leadership: A meta-analytic test of their relative validity. Journal of Applied Psychology 89: 755–768.

40.

Kouzes

Posner

(1995) The Leadership Challenge: How to Keep Getting Extraordinary Things Done in Organizations. San Francisco, CA: Jossey-Bass.

41.

Kruger

Rowold

Borgmann

. (2011) The discriminant validity of transformational and transactional leadership. A multitrait-multimethod analysis of and norms for the German Transformational Leadership Inventory (TLI). Journal of Personnel Psychology 10(2): 49–60.

42.

Larsen

Samdal

(2007) Implementing second step: Balancing fidelity and program adaptation. Journal of Educational and Psychological Consultation 17(1): 1–29.

43.

Leithwood

Jantzi

(2006) Transformational school leadership for large scale reform: Effects on students, teachers, and their classroom practices. School Effectiveness and School Improvement 17(2): 201–227.

44.

Leithwood

Sun

(2009) Transformational school leadership effects on schools, teachers and students. In: Hoy

Dipaola

(series eds) Research and Theory In Educational Administration (Vol. 8): Studies in School Improvement Programs. Charlotte, NC: Information Age, pp.1–22.

45.

Leithwood

Sun

(2012) The nature and effects of transformational school leadership: A meta-analytic review of unpublished research. Educational Administration Quarterly 48(3): 387–423.

46.

Leithwood

Aitken

Jantzi

(2001) Making School Smarter: Leading with Evidence (2nd edn). Thousand Oaks, CA: Corwin Press.

47.

Leithwood

Harris

Hopkins

(2008) Seven strong claims about successful school leadership. School Leadership and Management 28(1): 27–44.

48.

Leithwood

Menzies

Jantzi

(1994) Earning teachers’ commitment to curriculum reform. Peabody Journal of Education 69(4): 38–61.

49.

Leithwood

Tomlinson

Genge

(1996) Transformational school leadership. In Leithwood

Chapman

Corson

Hallinger

Hart

(eds) International Handbook of Educational Leadership and Administration. Dordrecht, Netherlands: Kluwer Academic, pp.785–840.

50.

Leithwood

Jantzi

Earl

. (2004) Leadership for large-scale reform. School Leadership and Management 24(1): 57–80.

51.

Leithwood

Jantzi

Silins

. (1993) Using the appraisal of school leaders as an instrument for school restructuring. Peabody Journal of Education 68(2): 85–109.

52.

McMurray

Islam

Sarros

. (2012) The impact of leadership on workgroup climate and performance in a non-profit organization. Leadership & Organization Development Journal 33(6): 522–549.

53.

McMurray

Pirola-Merlo

Sarros

. (2010) Leadership, climate, psychological capital, commitment, and wellbeing in a non-profit organization. Leadership & Organization Development Journal 31(5): 436–457.

54.

Midthassel

Ertesvåg

(2008) Schools implementing zero: The process of implementing an anti-bullying program in six Norwegian compulsory schools. Journal of Educational Change 9: 153–172.

55.

Moolenaar

Sleegers

Daly

(2011) Teaming up: Linking collaboration networks, collective efficacy, and student achievement. Teaching and Teacher Education 28: 251–262.

56.

Muthén

(1997) Latent variable modeling of longitudinal and multilevel data. Sociological Methodology 27: 453–480.

57.

Muthen

Kaplan

(1992) A comparison of some methodologies for the factor analysis of non-normal Likert variables: A note on the size of the model. British Journal of Mathematical and Statistical Psychology 45(1): 19–30.

58.

Muthén

(2012) Mplus User’s Guide (7th edn). Los Angeles, CA: Muthén & Muthén.

59.

Osborne

Costello

(2004) Sample size and subject to item ratio in principal components analysis. Practical Assessment, Research & Evaluation 9(11). Available at: http://PAREonline.net/getvn.asp?v=9&n=11 (accessed 5 April 2014).

60.

Oterkiil

(2010) Exploring Perceptions of Leadership for School Staff. Operant Subjectivity 33(1): 51–67.

61.

Oterkiil

Ertesvåg

(2012) Schools Readiness and Capacity to Improve Matter. Education Enquiry 3(1): 71–92.

62.

Podsakoff

MacKenzie

Moorman

. (1990) Transformational leadership behaviors and their effects on follower’s trust in leaders, satisfaction and organizational citizenship behavior. Leadership Quarterly 1: 107–142.

63.

Rowe

(2007) Practical Multilevel Analysis with MLwiN & LISREL: An Integrated Course (7th edn). Melbourne, Australia: Australian Council for Educational Research.

64.

Sashkin

(1990) The Visionary Leader: Leadership Behavior Questionnaire. Amherst, MA: Human Resource Development Press.

65.

Sergiovanni

(2007) Rethinking Leadership: A Collection of Articles (2nd edn). Thousand Oaks, CA: Corwin Press.

66.

Smith

Busi

Ball

. (2008) Factors influencing an organisation’s ability to manage innovation: A structured literature review and conceptual model. International Journal of Innovation Management 12(4): 655–676.

67.

Sniders

TAB

Bosker

(2012) Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modeling (2nd edn). London: Sage.

68.

Stigler

Hiebert

(1999) The Teaching Gap: Best Ideas from the World’s Teachers for Improving Education in The Classroom. New York: Free Press.

69.

Tabachnick

Fidell

(2007) Understanding Multivariate Statistics (5th edn). Boston: Pearson.

70.

Tepper

Percy

(1994) Structural validity of the multifactor leadership questionnaire. Educational and Psychological Measurement 54: 734–744.

71.

Thoonen

EEJ

Sleegers

PJC

Oort

(2011) How to improve teaching practices: The role of teacher motivation, organizational factors, and leadership practices. Educational Administration Quarterly 47(3): 496–536.

72.

Tidd

Bessant

Pavitt

(2005) Managing Innovation: Integrating Technological, Managerial Organizational Change (3rd edn). New York: Wiley.

73.

Van Dierendonck

Nuijten

(2011) The servant leadership survey: Development and validation of a multidimensional measure. Journal of Business Psychology 26: 249–265.

74.

Van Knippenberg

Sitkin

(2013) A critical assessment of charismatic-transformational leadership research: Back to the drawing board? The Academy of Management Annals 7(1): 1–60.

75.

Yukl

Lepsinger

(2005) Why integrating the leading and managing roles is essential for organizational effectiveness. Organizational Dynamics 34(4): 361–375.