Modeling the Influence of School Leaders on Student Achievement

Abstract

Purpose: The aim of this study was to examine the means by which principals achieve an impact on student achievement. Research Design: Through the application of structural equation modeling, a mediated-effects model for school leadership was tested, using data from 97 secondary schools in the Netherlands. Findings: The results showed a small positive effect of school leadership on the mean promotion rate in schools, mediated by a development-oriented school organization and favorable classroom practices. The promotion rate may be considered as a measure of efficiency. No indications of direct positive effects of school leader activities on student achievement were found. This might be the result of the relatively small differences in overall student achievement and school leader behavior between the schools studied. Conclusions: The results underline the important role school leaders play in school effectiveness and offer valuable insight in how school leaders actually can make a difference. School leaders were found to have a strong influence on development orientation in schools, which shows similarities with the idea of the “learning organization.” This study points to the importance of school context. The results show that contextual variables have considerable effects on several variables in the model.

Keywords

school leadership mediated effects structural equation modeling competing values model student achievement

Leadership is widely regarded as playing a significant role in school effectiveness and school improvement. In the past three decades, researchers have developed various models to understand the relationship between leadership and student achievement (Bossert, Dwyer, Rowan, & Lee, 1982; Hallinger & Heck, 1996, 1998; Leithwood & Levin, 2005; Pounder, Ogawa, & Adams, 1995). In the earlier studies, researchers were using models in which the relationship between leadership in schools and outcomes at the student level was measured as a direct causal link. More recently, researchers started to use mediated-effects models, which hypothesize that leaders achieve their effect on school outcomes through indirect paths. Throughout the years, various potential mediating variables have come to light, including the mission of the school, educational vision and goals, staff motivation, teacher classroom practice, and student engagement (e.g., Krüger, Witziers, & Sleegers, 2007; Hallinger, Bickman, & Davis, 1996; Hallinger & Heck, 1998; Leithwood, Day, Sammons, Harris, & Hopkins, 2006; Leithwood & Levin, 2005; Mulford & Silins, 2003; Pounder et al., 1995).

Although considerable conceptual and methodological progress has been made, little is known about the paths through which school leaders can enhance organizational and student outcomes and about the interplay with contextual factors (Krüger, Witziers, & Sleegers, 2007; Hallinger, 2003; Hallinger & Heck, 1996). Although different scholars have stressed the need to use more complex causal models, systematic empirical validation of mediated-effects models is scarce. Furthermore, researchers have used different models of school leadership, such as instructional, transformational, and strategic leadership, that focus on specific sets of leadership activities. Because of this conceptual diversity, research on the impact of school leadership has failed to give conclusive answers to one of the key questions in research on the role of school leaders in school effectiveness and school improvement: How can school principals become effective and “make a difference”? Recently, researchers have suggested using a more integrated model, which focuses on a broader set of leadership activities than those covered by the specific models applied in earlier research (Thoonen, Sleegers, Oort, Peetsma, & Geijsel, 2011; Hallinger, 2003; Leithwood & Levin, 2005; Robinson, Lloyd, & Rowe, 2008). Focusing on more generic functions of school leadership that might affect student and organizational outcomes is seen as promising to increase our understanding of the crucial role of school leadership for school effectiveness and school improvement.

This article makes a contribution to the empirical validation of complex causal models that will uncover some of the links in the chain of variables between the principal and the student outcomes. In this article, a causal model in which the influence of school leaders is mediated by school organizational and teacher work variables is presented based on the research models developed by Bossert et al. and Silins et al. Furthermore, an integrated leadership model is used to measure leadership practices. The model is tested using data from 97 Dutch secondary schools.

Theoretical Framework

Modeling the Impact of Leadership in Schools

The evidence on direct effects of school leaders is rather mixed; some meta-analyses, such as the one by Witziers, Bosker, and Krüger (2003), show negligible effect sizes, but others, such as the ones by Marzano, Waters, and McNulty (2005), Robinson et al. (2008), and Chin (2007), show small to medium effects. At the same time, it is quite plausible that successful leaders contribute to student learning through their influence on other people or features of their organizations (Hallinger & Heck, 1998).

Based on this idea, researchers have suggested using so-called mediated-effects models to better understand the impact of leadership on organizational and student outcomes. One of the first mediated-effects models was developed by Bossert and his colleagues in the late 1980s (Bossert et al., 1982; Dwyer, Barnett, & Lee, 1987; Dwyer et al., 1985). In this model, leadership is no longer proposed as having a direct influence on learning outcomes but as having an indirect influence through the way it has an impact on instructional organization and culture. One of the first studies testing the relationship among leadership (i.e., organizational leadership), mediating variables (i.e., four functions of effective organizations drawn from the work of Parsons, including goal achievement, integration, adaptation, and latency), and several measures of school effectiveness (i.e., perceived organizational effectiveness, student achievement, student absenteeism, and faculty and staff turnover rates) was conducted by Pounder et al. (1995). The findings of the path analysis showed that leadership has an indirect impact on the four measures of school effectiveness through two of the four organizational functions. The effect of organizational leadership on school effectiveness was mediated by the level of both goal achievement and latency (commitment). Drawing on the Bossert et al. (1982) model, Hallinger et al. (1996) examined the impact of leadership on student outcomes and found support for the model. The findings showed that the principal’s role in establishing a strong school climate and instructional organization appeared to be precisely the area that strongly predicts school effectiveness. Inspired by these studies, various conceptual and methodological approaches have been applied in research on the effects of school leadership (see Scheerens & Witziers, 2005; Heck & Hallinger, 2000; Krüger & Witziers, 2003; Leithwood & Levin, 2005).

These earlier studies have exposed a wide range of potential mediating variables at the level of the school through which the school leader could influence the academic performance that can be categorized in four functions of effective organizations (goal achievement, integration, adaptation, and latency) as distinguished by Parsons (1960) and Pounder and colleagues (1995) or four “domains” through which the school leader could influence the academic performance as formulated by Hallinger and Heck (1998) based on their literature review: “vision and goals,” “organizational structure,” “human capital,” and “organizational culture.”

Recently, Bryk, Sebring, Allensworth, Luppescu, and Easton (2010) presented a school improvement framework consisting of five core elements or organizational subsystems (leadership, instructional guidance, professional capacity, parent–community–school ties, and student-centered learning climate) to explain how the organization of a school interacts with classroom life to advance student learning. As in the earlier studies using mediated-effects models, in this framework of essential supports for school improvement several mediating variables at the school level (professional capacity, parent–community–school ties, and student-centered learning climate) are distinguished to explain the indirect influence of leadership on school outcomes. Next to this, variables at the teacher level (instructional guidance) are also considered to play an important role in mediating the impact of school leadership on student outcomes.

In addition to this line of research, Mulford and Silins (2003) developed a complex causal model, the so-called LOLSO (leadership for organizational learning and student outcomes) model (see Mulford & Silins, 2003; Mulford, Silins, & Leithwood, 2003; Silins & Mulford, 2002; Silins, Mulford, & Zarins, 2002). According to this model, transformational and distributive school leadership influences student engagement and student participation in school via organizational learning and teachers’ work. In the LOLSO model, more student engagement and student participation in school lead to higher retention rates (lower drop-out rates) and a better academic performance. Organizational learning is characterized by a trusting and collaborative climate, a shared and monitored mission, and taking initiatives and risks within the context of supportive, ongoing, and relevant professional development (Mulford et al., 2003). In accordance with earlier studies using mediated-effects models and the framework developed by Bryk et al. (2010), in the LOLSO model variables at both the school (organizational learning) and teacher levels (teachers’ work) are considered to mediate the impact of leadership on student learning. In contrast, the LOSLO model also distinguishes mediating variables at the student level (student engagement and student participation in school) to explain the indirect influence of school leadership on student academic achievement.

So far, only a limited number of mediated-effects models on the influence of school leaders on academic performance have been empirically validated (see Krüger, Witziers, & Sleegers, 2007; Hallinger & Heck, 1998; Heck & Hallinger, 2009; Leithwood & Jantzi, 2006; Louis, Dretzke, & Wahlstrom, 2010; Pounder et al., 1995; Sammons, Gu, Day, & Ko, 2011). There is still much to be learned about the paths through which school leaders can contribute to enhancing organizational and student outcomes. More fine-grained knowledge is also needed about how successful leaders respond to external policy initiatives and local needs. Different scholars suggest that the lack of attention to the possible impact of contextual variables could explain the contradictory results in studies on school leadership (Krüger, Witziers, & Sleegers, 2007; Hallinger, 2003; Leithwood & Levin, 2005).

In this study, the paths through which school leaders can make a difference and have an impact on student outcomes are examined. Drawing on the research models as developed by Bossert et al. and Mulford et al., a mediated-effects model is described. This causal model is tested by conducting structural equation modeling. In doing this, the study makes a contribution to a better understanding of the antecedents and effects of educational leadership and of the influence of the principal’s leadership on intervening and outcome variables.

General Versus Specific Leadership Models

Beside the need to validate more complex casual models of leadership effects, using a more integral conceptual model for effective leadership in schools is considered as a further challenge to future research on school leadership. Based on an extensive review, Leithwood and Levin (2005) recommend investigating a broader set of leadership activities than those covered by the specific models that have been applied earlier, such as “instructional” leadership, which is directly geared to the educational activities of the teachers, “transformational” leadership, which aims primarily to enhance the commitment and competencies of teachers, and “strategic” leadership, which focuses on coordination, planning, monitoring, and the allocation of resources. Marks and Printy (2003) explored the influence of “integrated” leadership on academic performance and concluded that transformational leadership is needed to promote change and distributive (shared) school leadership is needed to improve the learning performance of the students. Different scholars note that the generic functions of leadership are often obscured by the many labels that are used for different forms and styles of leadership (Hallinger, 2003; Leithwood, Louis, Anderson, & Wahlstrom, 2004; Robinson et al., 2008). An integrated leadership model would therefore provide more insight into the effectiveness of school leaders. According to Leithwood et al. (2006) successful school leaders make use of the same basic repertoire of activities, which can be split into four main categories.

Developing a vision and giving direction: Identifying and formulating a vision, creating a shared interest, demonstrating high expectations for performance, promoting the acceptance of group objectives, monitoring organizational performance, and communicating.

Understanding and developing people: Providing intellectual stimulation, giving individual guidance, and setting a good example. The school leader builds on the knowledge and skills of teachers and other personnel to achieve the school objectives.

Redesigning the organization: Building on cultures and cooperative processes, managing the environment and working conditions, building and maintaining productive relations with parents and the community, and connecting the school with the wider environment.

Managing the teaching and learning program: Creating a productive working environment for both teachers and students, promoting organizational stability, guaranteeing effective leadership with the focus on learning, appointing teachers and supporting staff to implement the curriculum, and monitoring school activities and performance.

These four categories of leadership activities show considerable overlap with the four models for organizational effectiveness as described in Quinn and Rohrbaugh’s (1983) competing values framework.

Quinn and Rohrbaugh (1983) developed the competing values framework as a way to invest the concept of effectiveness in organizational literature (see Figure 1). Their goal was to identify the structure among possible criteria used to evaluate organizational effectiveness. The model derived from the ordering of criteria that organizational researchers and researchers use to evaluate the performance of organizations. Underlying this spatial model are two, independent, bipolar dimensions: (a) an internal versus external focus dimension and (b) a flexibility versus control dimension.

Figure 1.

Competing values framework: Organization level

The first bipolar dimension, representing the opposing quadrants of “human relations” and “rational goals,” refers to the focus of the organization. It reflects the contrast between an internal, person-oriented emphasis and an external, organization-oriented emphasis. The underlying assumption is that some organizations are effective when there is a harmonious internal atmosphere, whereas other organizations benefit more from a strong orientation toward cooperation and competition outside the own organization.

The second bipolar dimension, representing the opposing quadrants “internal process” and “open systems,” relates to the structure of the organization. This dimension reflects differing organizational preferences for structure by representing the contrast between attention for stability and control versus attention for flexibility and change. Some organizations are effective when they are stable and predictable, and other organizations are effective when they are able to adjust to the external world. The quadrants refer to the four major models for organizational effectiveness:

The rational goals model: A clear vision leads to productive results. The emphasis is on the clarification of goals, rational analysis, and decisive action.

The internal process model: Stability is established via routines. The emphasis is on defining responsibilities, measurement, and documentation.

The human relations model: Engagement leads to effort. The emphasis is on participation, conflict management, and consensus building.

The open systems model: Ongoing adjustment and innovation lead to the acquisition and maintenance of external resources. The emphasis is on political adjustment, creative problem solving, innovation, and change management.

Although the competing values framework was originally developed to find a structure in criteria for organizational effectiveness, it has been applied in a broad range of organizational research, such as the investigation of organizational culture and strategy, organizational development, human resource development, the life cycle of organizations, and leadership (Bluedorn & Lundgren, 1993; Cameron & Freeman, 1991; Denison, Hooijberg, & Quinn, 1995; DiPadova & Faerman, 1993; Quinn & Cameron, 1983; Quinn & Kimberly, 1984; Quinn & McGrath, 1985; Zammutto & Krakower, 1991). Effective leadership is defined by Denison and his colleagues (1995) as “the ability to perform the multiple roles and behaviors that circumscribe the requisite variety implied by an organizational or environmental context” (p. 526). All leaders experience “paradoxical” demands or conflicting roles in their work, and the effective leader is able to meet these demands by displaying behaviors that are situated in at least three different quadrants. Findings from research on effective leadership behavior in profit organizations suggest that the competing values framework is a valid model for analyzing effective leadership (Kalliath, Bluedorn, & Gillespie, 1999; Quinn & Spreitzer, 1991; Thompson, 2000).

The four categories of the basic leadership repertoire defined by Leithwood et al. (2006, see above) correspond with the models in the competing values framework as follows: “developing a vision and giving direction” fits in the rational goal model, “managing the teaching and learning program” matches with the internal process model, “understanding and developing people” corresponds with the human relations model, and “redesigning the organization” is an important leadership behavior according to the open systems model. Studies in which the competing values framework as a general leadership model is used as an approach to measure effective leadership activities in schools are scarce, however. We believe that the competing values framework as a model for measuring generic leadership practices will provide more insights into effective leadership in schools.

A Mediated-Effects Model for School Leadership Effects

Based on the work of Bossert et al., our model (see Figure 2) assumes that the school leader indirectly influences academic performance via his or her influence on the school culture and school organization. It is expected, however, that the influence of school leader behavior will be greater on school organizational practices than on school culture (e.g., Author, 2007; Heck, Larsen, & Marcoulides, 1990). Following the LOLSO model (see above), it is assumed that the effect of school leader behavior runs through two mediating variables, namely, teachers’ work and student engagement. Leithwood and Jantzi (2000) see student engagement as a plausible predictor of academic performance (see Bredschneider, 1993; Dukelow, 1993; Finn & Cox, 1992; Mulford, 2003). In the LOLSO model, the effect of student engagement is reflected mainly in higher rates of retention (lower drop-out rates). The retention variable is represented in our research model by the promotion rate; in the Dutch context, this means the proportion of students from a cohort that passes the final central examination for secondary schools without delay (see the specification below).

Figure 2.

Research model for analyzing the influence of school leader behavior on the outcome variables

In our research model, school leader behavior is the main independent variable. As mentioned above, we used the four models of the competing values framework as developed by Quinn and Rohrbaugh (1983) to measure a broader set of school leader activities. As mentioned above, the competing values framework has also been applied in a broad range of organizational studies and has proved to be a valid instrument to measure all kinds of organizational concepts. We therefore used this model in our study to measure school culture and school organization. Using the same framework for measuring different (independent and mediating) variables offers possibilities to establish relationships between school leader behavior and different organizational characteristics (Quinn, 1998).

Our research model also incorporates school context and school composition variables, which were expected to have effects on the outcome variables as well as on other variables in the model. A relationship between two variables might be explained by a common cause. Therefore, these common causes cannot be left out of the model (e.g., Saris & Stronkhorst, 1984). The contextual variables included in our model refer to the composition of the student population, school size, denomination, and the degree of urbanization of the school neighborhood.

By conducting structural equation modeling, we attempt to validate our causal model and make a contribution to a better understanding of the way school leaders’ sets of activities affect the influence of mediating and outcome variables as well as the interplay with contextual forces that influence the exercise of school leadership. The next section describes the sample and instruments used to address the question of how school leaders can make a difference.

Method

Sample

Schools are the primary unit of analysis in this study. Although the data collection related to students, teachers, and school leaders, the analyses reported are exclusively based on school-level data. Data obtained from students and teachers were aggregated at the school level. The focus of the study is the influence of school leaders on student achievement. Information on these outcomes (academic performance and promotion rates) was available only via aggregate measures at the school level. Our analyses thus assess the impact of school leaders on mean school performance and promotion rates. The data provided by teachers were used in the analyses as indicators of school characteristics (school leader behavior, school culture, and school organization). This also applies to the data obtained from the students. The appendix reports the variability between schools (i.e., intraclass correlations) for the teacher and student variables that were included in the analyses.

All 485 secondary schools in the Netherlands with a HAVO stream (HAVO is senior secondary general education, which comprises five grade levels starting at the age of 12) were selected to participate in this study. The study focused on the HAVO stream because it has a homogeneous curriculum and student population (in terms of ability). We asked school principals to participate in our study. These school principals are members of the management team of the school and are formally responsible for the final 2 years (out of 5) of the HAVO stream (delegated authority). Next, we selected 15 teachers working with students from the last 2 years of the HAVO stream. We also asked all students from Year 5 to participate.

The school leaders were asked to fill out a questionnaire on school leader behavior (42 items) and contextual variables (e.g., school size, competition, etc.). Teachers were asked to fill out a questionnaire on school leader behavior (42 items), school culture (40 items), and organizational practices (40 items). We asked school leaders and teachers to fill out the same questionnaire on school leader behavior to test the validity of teachers and school leader perceptions for measuring school leader behavior. Despite the importance of this validity check, it is absent in most school leadership research. Finally, the students were asked to fill out a questionnaire measuring their engagement with school and their perceptions of different aspects of teachers’ work (42 items).

In total, 103 school leaders (response rate = 21%), 998 teachers (response rate = 65%), and 4,336 students from Grade 5 returned the questionnaires. A response rate of 21% for the school leaders may seem quite low; however, considering that an invitation was sent to every HAVO school in the country and every school was subsequently called, it can be considered a satisfying response rate. Of the 103 HAVO schools with a positive response, 97 schools were included in the analyses. Schools were included when both the staff (school leader and teachers) and the students of the school responded to the questionnaires.

The average school size was 1,148.15 students; the smallest school had 340 students, and the largest school had 2,251. The schools included in the sample were comparable with other schools in the country with regard to school size (four categories: 1 = fewer than 500 students, 2 = 500–1,000 students, 3 = 1,000–1,500 students, 4 = more than 1,500 students; χ² = 2.818, df = 3, p = .421). The sample was also representative as far as the distribution of the schools over the 12 provinces of the Netherlands is concerned (χ² = 19.334, df = 11, p = .0535). However, compared to the total population (four categories: 1 = public, 28.3%; 2 = Catholic, 26.0%; 3 = Protestant, 21.7%; and 4 = other, 24.05%), the sample included relatively more Catholic schools (38.1%) and fewer Protestant schools (14.4%) and schools with another denomination (19.6%). As a consequence, the schools in our sample were not comparable with other schools in the country with regard to denomination (χ² = 8.657, df = 3, p = .034).

With regard to the student population at the schools, we found that the HAVO departments have a mean of 21% of students coming from families with a low SES, 45% of students coming from families with a middle SES, and 34% of students coming from families with a high SES. Most of the schools in our sample have low percentages of students from cultural minorities; only five schools have more than 50% of students coming from cultural minority groups. Of the schools 75%, have fewer than 8% of their students from a cultural minority.

Most school leaders in our study are male; only 18.4% of our school leaders are female. This is in line with data collected from the Ministry of Education concerning the overall population of school leaders in the Netherlands. In the 2006 data collection of the Ministry of Education, 18% of managers in secondary education in the Netherlands were female. In terms of age, the youngest school leader in our sample was 35 years old and the oldest school leader was 63 (M = 52). Most of the teachers (71.6%) who participated in our study were also male. The youngest teacher was 27, and the oldest teacher was 68 (M = 48). School leaders have worked 9.28 years on average at their current school (Mdn = 4 years). Teachers work much longer at the same school; on average, teachers have worked for 17.7 years at their current school, with the median being 19 years.

Instruments

All concepts included in our research model were measured using items and (sub)scales from existing questionnaires as well as additional newly formulated items. We carefully translated and adapted English items for appropriateness in the Dutch secondary educational context. To verify the content validity of the items of the different instruments, experts (school leaders and expert teachers) reviewed the item formulations.

School Leader Behavior

Based on the competing values framework, four leadership practices were measured: rational goal behavior, internal process behavior, human relations behavior, and open systems behavior. We used existing scales and items to measure these four leadership practices (Krüger, Witziers, & Sleegers, 2007; Hallinger, 1994; Krüger, 1994). School leaders were asked to evaluate their own behavior and teachers to evaluate the behavior of their school leaders. In both versions of the questionnaire (teacher and school leader), similar aspects of school leadership behavior were addressed, although they were formulated slightly differently. For example, school leaders were asked to evaluate the statement “To what extent do you give positive feedback to teachers who have performed well?” whereas teachers were asked to evaluate the following statement: “To what extent does the principal of your school give positive feedback to teachers who have performed well?”

Sample items for each of the four models are the following:

“My school leader/I develop(s) goals that are easily translated into classroom objectives by teachers” (rational goal model, 8 items).

“My school leader/I ensure(s) that the rules of the school are being enforced” (internal process model, 11 items).

“My school leader/I support(s) and coach(es) teachers at the individual level” (human relations model, 14 items).

“My school leader/I ensure(s) that there are at least one or two educational experiments at this school” (open system model, 9 items).

Both instruments used a 4-point Likert-type scale (1 = hardly, 4 = very often).

School Culture

In the competing values framework, four different types of cultures (or core values) are classified: the rational goal model, internal process model, human relations model, and open systems model (Quinn, 1988). To measure these competing value models, the School Culture Questionnaire as developed by Houtveen, Voogt, Vegt, and van de Grift (1995) and Maslowski (2001) was used to examine and validate the different types of school cultures as classified by the competing values framework in Dutch schools. Sample items for each of the four models include the following:

“At our school, we have high expectations of the student outcomes of all students” (rational goal model, 10 items).

“At our school, we value clear procedures” (internal process model, 10 items).

“At our school, we consider collaboration as very important” (human relations model, 10 items).

“At our school, we are very much attached to having a clear profile” (open system model, 10 items).

Teachers were asked to indicate to what extent these values are part of their school culture using a 4-point Likert-type scale (1 = to a small extent, 4 = to a great extent).

School Organization

To measure the four different types of school organizational practices (rational goal, internal process, human relations, and open system), scales and items of existing questionnaires (Hallinger, 1994; Houtveen et al., 1995; Krüger, 1994) as well as additional newly formulated items were used. Sample items for each of the four practices include the following:

“At our school, student outcomes are systematically analyzed” (rational goal model, 10 items).

“At our school, we have clear procedures to ensure that every staff member will receive the right information” (internal process model, 10 items).

“At our school, teachers collaborate with each other to learn” (human relations model, 10 items).

“At our school, activities are organized to enhance the image of our school” (open system model, 10 items).

Teachers were asked to indicate to what extent these practices are actual practices at their school, using a 4-point Likert-type scale (1 = to a small extent, 4 = to a great extent).

Teachers’ Work and Student Engagement

To assess teachers’ work, students were asked to evaluate four different aspects of teachers’ classroom practice: interaction with teachers, the learning support of teachers, the working atmosphere in the classroom, and the organization of the subjects and the lessons. Student engagement was measured by asking students to evaluate the atmosphere at school and their relationships with fellow students using items adopted from the Engagement and Family Educational Culture Survey of Leithwood, Aitken, and Jantzi (2000). Several other student questionnaires (Engels, Aelterman, Schepens, & Van Petegem, 2003; Houtveen, Vermeulen, & van de Grift, 1993; Krüger, 1994; Organisation for Economic Co-operation and Development, 2003) were used for additional items. For example, students were asked to evaluate the following statements: “Most of the teachers put a lot effort in helping students” (teachers’ work) and “I like to go to school” (student engagement). Both instruments used a 4-point Likert-type scale (1 = completely disagree, 4 = completely agree).

Outcome Variables

Two outcome variables, academic performance and mean promotion rate, were measured using data from the Inspectorate of Education (made available by DANS).¹ Academic performance was based on the average final examination score for all subjects. This score is a weighted average, which takes into account the number of students per subject. The mean promotion rate is an estimate of the probability that students will be promoted to the next grade. This probability is computed by averaging the actual promotion rates for each school year. The average scores for the final examination and the promotion rates are calculated over a period of 3 years (2003–2005) to take into account annual variations (see Luyten, 1994; Maslowski, 2001).

Contextual Variables

School size (number of students) and school composition were included as important background variables because they are known to be directly related to school leader behavior, student engagement, and outcomes (Krüger, Witziers, & Sleegers, 2007; Mulford & Silins, 2003; Opdenakker & Van Damme, 2006). The composition of the student population has a substantial effect on the outcomes (see, e.g., Author et al., 2005). In the Netherlands, a negative relationship is often found between the percentage of students from cultural minorities and average academic performance (Driessen, 2002; Veenstra, 1999). Data on school composition (percentage of cultural minorities) came from a database for the league tables that are published for Dutch secondary schools by the Inspectorate of Education.

Denomination was included because studies have indicated that Catholic schools are more effective than public schools (Dronkers, 2004; Opdenakker & Van Damme, 2006). Data on denomination came from the same database we used for assessing school composition.

Although the results of PISA 2000 suggest that there are only a few countries where the school neighborhood has a significant effect on academic performance (Luyten, Scheerens, Visscher, Maslowski, Witziers, & Steen, 2005), Dutch studies have found a relation among urbanization, academic performance, and student well-being (Krüger, Witziers, & Sleegers, 2007; Lugthart, Roeders, Bosker, & Bos, 1989). Therefore, the degree of urbanization was also included in our study. Data on the urbanization of the environment of the schools were taken from demographic statistics.

Various authors maintain that home educational culture is a better predictor of academic performance than the socioeconomic status of the family (Leithwood et al., 2000; Marzano, 2000; Walberg, 1984). In the LOLSO project, Mulford and Silins (2003) found a negative correlation between the socioeconomic status of students and the perception of teachers’ work and a strong positive correlation between home educational culture and teachers’ work. Based on these findings, “valuing” of education was included as a background variable in our study. This variable is partly based on items that relate to family educational culture (e.g., “my parents are always willing to help me with my homework”) and partly on additional items that relate to the perceived value of education in general (e.g., “everybody should receive as much schooling as possible”). In contrast to the variables that relate to classroom practices (teacher work) and perceived school atmosphere (student engagement), this variable is less prone to manipulation by teachers and school leaders.

The last contextual variable included was competition from other schools, as reported by the school leaders. Schools that must compete with other schools to survive cannot afford to be selective and may therefore have a more diverse student population. Furthermore, research has indicated that schools that experience heavy competition are stronger in policy making than are other schools (Sleegers, 1991).

Analyses

To guide scale construction, confirmative factor analyses were conducted, resulting in the exclusion of some items because of a lack of correlation or stable factor structure.² The results showed that four leadership practices could be determined, representing the four models of the competing values framework for both school leaders and teacher data: rational goals behavior, internal process behavior, human relations behavior, and open system behavior. The reliability of these scales, as indicated by Cronbach’s alpha, was sufficient, ranging from .67 to .76 for the school leaders’ scales and from .73 to .91 for the teachers’ scales. With regard to school culture, findings from confirmative analyses showed that the four models of the competing values framework could also be determined. The reliability of the scales, as indicated by Cronbach’s alpha, was sufficient, ranging from .70 to .86. The confirmative factor analyses also showed that four organizational practices referring to the competing values framework could be determined. Cronbach’s alpha of these scales varied between .71 and .80. Finally, we conducted confirmative factor analyses on the student data. Findings showed that three variables could be determined: teachers’ work, student engagement, and valuing education. The reliability of the scales, as indicated by Cronbach’s alpha, was sufficient, ranging from .70 to .83.

Based on these findings, the variables as constructed at the individual level were aggregated to the school level using factor scores. By doing this, high correlations were found between the school culture and school organizational scales, raising the question of whether these scales measured different concepts at the school level. For this reason, we conducted additional principal component factor analyses on the eight variables of school culture and school organizational practices, resulting in two components that together explained 82% of the variance (at the school level). Based on these findings, two new school organizational variables, titled performance orientation and development orientation, were constructed. The performance orientation variable relates to an orientation on student performance with regard to both culture and practice. It consists of the sum of the aggregated scores for rational goals culture (e.g., “at our school, we have high expectations of the student outcomes of all students”), internal process culture (e.g., “at our school, we value clear procedures”), and rational goals practices (e.g., “at our school, student outcomes are systematically analyzed”). The development orientation variable indicates a focus on human relations and external responsiveness. It consists of the sum of the aggregated scores for human relations culture (e.g., “at our school, we consider collaboration as very important”), open systems culture (e.g., “at our school, we are very much attached to having a clear profile”), internal process practices (e.g., “at our school, we have clear procedures to secure that every staff member will receive the right information”), and human relations practices (e.g., “at our school, activities are organized to enhance the image of our school”).

Next, the reliabilities of the aggregated student variables were calculated, taking class size and intraclass correlation into account.³ The reliability of the aggregated variables was satisfying (ranging from .64 to .82) with the exception of the valuing education variable (α = .44).

To analyze the whole model, a two-step strategy was used. First, the influence of contextual and leadership variables on performance orientation and development orientation was analyzed. All nonsignificant relationships were removed. Next, an optimal model was determined for the influence of the contextual variables and the teachers’ work and student engagement variables on the outcomes. A relationship emerged between student engagement and teachers’ work, but student engagement had no direct influence on the outcomes. Finally, the two models were linked and a decision was taken on whether to add direct effects of school leader behavior and school organizational orientations on the outcomes.

Second, the final research model was tested twice, first with data measuring teachers’ perceptions of the behavior of their school leader, followed by data measuring school leader perceptions of their own actions. Although both tests showed a good fit, more significant relations were found using the teacher data. This finding matches findings from recent research on successful school principalship (see Mulford & Silins, 2011). Table 1 presents an overview of the correlations among all variables that were included in the analyses.

Table 1.

Correlations Among Variables Included in the Structural Equation Model

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
1. Average exam	1
2. Promotion rate	.364	1
3. Student engagement	.108	.370	1
4. Teachers’ work	.131	.246	.640	1
5. Performance orientation	.212	.359	.292	.265	1
6. Development orientation	−.077	.237	.230	.372	.411	1
7. RG behavior	−.132	.046	.089	.195	.353	.639	1
8. IP behavior	.095	.034	−.012	.116	.398	.258	.611	1
9. HR behavior	.043	−.042	−.004	.122	.227	.368	.769	.736	1
10. OS behavior	−.204	.015	−.063	.109	.241	.596	.570	.301	.233	1
11. School size	.074	.152	.236	.191	.219	.229	−.127	−.136	−.254	.146	1
12. Catholic denomination	.280	.311	−.077	−.086	.253	−.013	.059	.059	.130	−.083	.061	1
13. Percentage minorities	−.382	−.451	−.421	−.101	−.188	−.010	.011	.041	.099	.013	−.133	−.159	1
14. Degree of urbanization	−.304	−.307	−.229	−.106	−.037	.150	.189	−.046	−.007	.278	−.049	−.154	.281	1
15. Valuing education	−.116	.048	.382	.519	.127	.190	.137	.098	−.016	.050	.056	−.201	.000	.014	1
16. Competition other schools	−.243	−.462	−.419	−.171	−.303	−.213	−.057	−.054	−.082	.153	−.183	−.217	.231	.413	.044	1

RG = rational goals; IP = internal process; HR = human relations; OS = open systems. Correlations relate to variables aggregated at the school level (N = 97). Correlations significant at the .05 level (two-tailed) are in bold.

Results

The parameter estimates of the final model are presented in Figure 3. It appeared that the criteria for a good fit were met: the root mean square error of approximation (RMSEA) is less than .06, the standardized root mean square residual (SRMR) is less than .08, and the comparative fit index (CFI) is greater than .95. The model explains 28% of the variance in the average examination scores. The effects of the contextual variables on the other research variables are not shown because they would have made the figure intractable.

Figure 3.

The influence of school leader behavior on student outcomes

The results showed small but significant, positive, mediated effects from rational goals, internal process, human relations, and open systems school leader behavior on mean promotion rates via development-oriented school organization and teachers’ work. The actions of school leaders had no significant mediated effects on the average examination scores; only negative direct effects were found for both rational goals and open systems.

It appeared that a significant and positive relationship exists among rational goals, internal process, and open systems school leader behaviors and performance orientation. No relationship was found between human relations behavior and performance orientation. Furthermore, no significant relationship was found between performance orientation on the hand and organizational variables, classroom practices, and student outcome variables on the other hand. Next, the results showed that significant positive relationships exist between all four school leader behaviors and development orientation. Rational goals and open systems display a stronger relationship with development orientation than with performance orientation.

Looking at the right side of the model, no direct effect of student engagement on student outcomes was found. The findings also showed that, although not assumed (see Figure 2), teachers’ work affects the promotion rate of schools directly. Next to this effect of teacher work on student outcomes, we also found a reciprocal effect between teachers’ work and student engagement. Finally, the results showed that the effect of teachers’ work on the average final examination score is mediated by promotion rate.

The total effects (sum of all direct and indirect effects) of the leadership variables on the mediating and outcome variables are presented in Table 2. Significant positive total effects were found for school leader behavior on performance orientation and development orientation. Furthermore, positive total effects of leadership behavior on teachers’ work for rational goals and open systems leadership were found together with smaller effects for internal process and human relations behaviors. Finally, the results showed a negative significant total effect of rational goals and open systems leadership on the average final examination score.

Table 2.

Standardized Total Effects of School Leader Behavior on Mediating and Outcome Variables

	RG Behavior	IP Behavior	HR Behavior	OS Behavior
Performance orientation	.18*	.25*	—	.24*
Development orientation	.57*	.23*	.35*	.56*
Teachers’ work	.15*	.06*	.09*	.15*
Student engagement	.05	.02	.03	.05
Promotion rate	.03	.01	.02	.03
Average exam score	−.16*	.00	.00	−.18*

RG = rational goals; IP = internal process; HR = human relations; OS = open systems.

Significant at α < .05.

In Table 3, the standardized total effects of contextual variables on all the other variables are presented. The results showed that nearly all contextual variables were related both to the outcome variables (exam scores or promotion rates) and to either school leader behavior or school culture and organization characteristics. The only exception was the percentage of cultural minorities in the school. This variable was negatively related to exam scores, promotion rates, and student engagement but not any other variables; awareness of competition and percentage of cultural minorities in the school were negatively related to school characteristics and outcomes. Awareness of competition was negatively related to nearly all other variables except exam scores. A small negative effect was found for school size on human relations school leader behavior. In addition, school size also appeared to have a moderately positive influence on open systems school leader behavior and development orientation and a small positive influence on performance orientation and teachers’ work. The degree of urbanization of the school neighborhood had mainly positive effects on different leadership variables and organizational variables. Despite these positive effects, the results also showed that schools that are situated in a more urbanized environment have a lower average exam score than schools that are located in a less urbanized environment. With regard to denomination, the results showed that of all denominations, only the Roman Catholic denomination matters. Compared with other schools, Roman Catholic schools have school leaders who show more rational goal and human relations behavior and show a higher promotion rate. Finally, the findings showed that students’ valuing of education has a strong positive effect on teachers’ work and student engagement and a small positive effect on performance orientation, development orientation, and promotion rate.

Table 3.

Standardized Total Effects of Contextual Variables

	School Size	Catholic Denomination	Percentage of Minority Students	Urbanization	Valuing Education	Competition
RG behavior		.16*		.25*	.13*	−.13*
HR behavior	−.12*	.19*
OS behavior	.25*			.28*
Performance orientation	.06*	.03		.11*	.02*	−.35*
Development orientation	.26*	−.03		.30*	.08*	−.33*
Teachers’ work	.07*	−.01	−.10	.08*	.50*	−.18*
Student engagement	.02	−.00	−.35*	.02	.39*	−.36*
Promotion rate	.01	.20*	−.34*	.01	.09*	−.34*
Average exam score	−.04*	.03	−.37*	−.09*	.00	−.07

RG = rational goals; HR = human relations; OS = open systems.

Significant at α < .05.

Conclusions and Discussion

In this study, the relationships between school characteristics and student outcomes were analyzed using a mediated-effects model. Drawing on the research models developed by Bossert et al. and Mulford et al., our research model assumed that the effect of school leader behavior runs through mediating variables, such as school organization, school culture, teachers’ work, and student engagement. In addition, the competing values framework as developed by Quinn and Rohrbaugh (1983) was used to measure generic leadership practices in schools. By conducting structural equation modeling, the model was tested using data from 97 schools.

The findings showed that school leader behavior affected student outcomes both indirectly and directly. Rational goals and open systems behaviors had both significant, positive indirect effects on the average promotion rate and negative direct effects on the average final examination scores. Although direct effects were found in our causal model, these found effects need not always reflect “real” direct effects; they can also signify an “unexplained” effect if the model does not incorporate all relevant confounding variables.

Given the increased emphasis on student performance and tightened “output” controls, introduced by accountability policies, it may be that the negative effects found have to do with the response of school leaders to the past performance of their schools. Leaders of schools with a poor academic record are under more pressure to take action than leaders of schools with a good academic record. Leaders of schools with a good academic record are less inclined to develop and set goals, take innovative reform initiatives, and set up experiments to improve their school.

Our findings do suggest that school leaders have a strong influence on development orientation in schools. Rational goals and open systems behavior seem to have the greatest impact, followed by human relations and internal process. Performance orientation is less related with school leader behavior. There are modest relationships with rational goals, internal process, and open systems behaviors, whereas human relations leadership seems to have no impact on performance orientation. When a number of relevant contextual variables are taken into account, performance orientation does not appear to be related to the outcomes, whereas development orientation has an indirect effect on promotion rate. The absence of a relationship between performance orientation and student outcomes should give food for thought, given the fact that it is generally assumed that performance orientation enhances the effectiveness of the school (Scheerens & Bosker, 1997).

The development orientation variable measures the degree of importance attached to cooperation, professionalism, and innovation and shows similarities with the idea of the learning organization in the LOLSO model. According to Mulford and Silins (2003), the learning organization is characterized by mutual trust, risk taking, a shared mission, and ongoing professional development. As such, a development-oriented school culture may play an important role for the improvement of teachers’ work. Our findings confirm the importance of a development-oriented school culture as a mediating variable that can help to explain the impact of leadership on teacher work. An effective goal- and innovation-oriented school leader can thus promote a development-oriented culture and improve the professionalism of the teaching staff.

Our results also showed that teachers’ work is strongly related to student engagement. Furthermore, teachers’ work did have a small positive effect on the average promotion rate and, via the promotion rate, a small effect on the average final examination score. These findings indicate that the learning environment teachers create in their classroom can affect the degree students like to be at school and are engaged with school and their performance. Although student engagement is often considered to be an important variable for future academic performance and failure, our results do not show that student engagement mediates the influence of teachers’ work on promotion rate or academic performance. It may be that students who do not feel at home at school drop out before Year 5, resulting in a decrease of variation in student engagement scores. Inspection of the engagement scores of the Year 5 students seems to support this assumption: Relatively high scores were found with little variation. These results do not concur with the findings of the LOLSO research, in which an indirect effect of engagement on academic achievement through retention was found (Mulford, Silins, & Leithwood, 2004; Mulford & Silins, 2003). Furthermore, the LOLSO research also found a direct relationship between student participation in school, a variable we did not include in our model, and academic achievement. More research is needed into the mediating role of student engagement and student participation in school in explaining the relation among teachers’ work, academic performance, and student failure.

It was found that the average promotion rate has a moderately positive effect on the average final examination score. Although average final examination scores can be regarded as an indicator for school effectiveness and promotion rate as an indicator for efficiency (Maslowski, 2001), these results support the importance of the link between these different but related outcomes for measuring school performance. What is less clear, however, is how to explain the relation between flow through and examination results. It may be that a positive atmosphere at school and good teachers’ work could lead to a higher academic performance and fewer dropouts and hence increase the flow through. On the other hand, it is also possible that a higher promotion rate might actually lead to a lower average final examination score, if more students with mediocre scores were given the benefit of the doubt. More research is needed to explain the relation between flow through and the average final examination score. Future research should also include noncognitive outcomes of schooling as recent research has shown the importance of these school outcomes for a child’s social development for future life changes (Mulford & Silins, 2011). By including both school outcomes in future studies, the findings may help to increase our understanding of the multidimensional nature of school performance (see Scheerens, 1989).

In our causal model, different contextual variables that could influence academic performance and school leadership were included. Excluding variables that affect two or more variables in a structural model can lead to erroneous conclusions about the relationships in the model. Contextual variables that influence only one of the variables, such as gender, in the model can, however, safely be excluded (Verschuren, 1991). The most important covariables for academic performance in our model seem to be related to student background, especially valuing education. As in other studies (e.g., Krüger, Witziers, & Sleegers, 2007), our results confirm the key role contextual variables play in explaining the impact of leadership on outcomes.

The present study contributes to the development of mediated-effects models needed to understand how school leaders indirectly influence school performance, as has been requested by several scholars in different ways (Krüger, Witziers, & Sleegers, 2007; Bossert et al., 1982; Hallinger & Heck, 1998; Mulford & Silins, 2003). As mentioned earlier, different scholars have used all kinds of mediating variables to explain the paths through which school leaders have an impact on students outcomes. Mediated-effects models differ in the kind of mediating variables selected, including variables at one (school), two (school and teacher), or three levels (school, teacher, and student). The model used in our study included mediating variables referring to three different levels: school (school organization and school culture), teachers (teachers’ work), and student (student engagement). The findings do concur with findings from earlier studies and add to the existing literature the importance of selecting mediating variables at different levels in modeling the impact of school leaders on school outcomes. To develop a sufficient foundation on which to build theoretical and robust understandings of effective leadership in schools, researchers should select variables that have been shown to play an important role in mediating the relationship between leadership and school effectiveness.

To measure generic leadership practices the competing values framework as developed by Quinn and Rohrbaugh (1983) was used. According to this framework, effective leaders are able to perform the multiple roles and behaviors that are distinguished in this model (rational goal behavior, internal process behavior, human relations behavior, and open systems behavior). The findings indicate that school principals were able to display the multiple behaviors and that these behaviors have a differential influence on mediating and outcome variables. Although rational goals behavior and open systems behavior seem to have the greatest impact on school organizational functions, the other two leadership behaviors also have an influence on organizational functions and, in turn, teachers’ work, student engagement, and student outcomes. The findings suggest that using an integrated leadership model can provide more insights into effective leadership in schools. More research is needed to validate the findings of this study and explore the role and impact of different leadership behaviors on mediating variables and several measures of school outcomes.

Different scholars have emphasized the need to incorporate contingent characteristics of school leadership in future research (Krüger, Witziers, & Sleegers, 2007; Hallinger, 2003; Leithwood & Levin, 2005). In this study, several contextual variables were included in the model, and the results showed that these variables do matter. The findings indicate that a contingency model of leadership could be helpful to understand the path through which school leaders have an impact on outcomes. By using contingency models of school leadership, scholars can help to unravel the complex links among contextual factors, leadership variables, aspects of the school organization, and variables related to the school’s effectiveness.

Limitations and Future Directions

Three limitations need to be highlighted with respect to the research study. First, our study was limited by aggregating student and teacher data to the school level. Although this has led to a loss of variance, the reliability of the aggregated variables turned out to be acceptable. By applying multilevel techniques in the structural analysis, student characteristics can be more effectively discounted (see, e.g., De Maeyer, Rymenans, Van Petegem, Van den Berg, & Rijlaarsdam, 2007). A multilevel model can incorporate individual academic performance and background characteristics. The acquisition of individual student data, however, was outside the scope of this study. The introduction of a new student administration system in the Netherlands will make it possible, in the future, to build databases in which individual student performance is linked to background details. This certainly will open up interesting avenues for future research.

Another possible limitation of our study is its cross-sectional nature. This means that the causality of the observed relationships can only be inferred from theory. As the study shows a “snapshot image” of the situation, it may have exposed coincidental temporary relationships. Causality is not ensured because the measurements are obtained at a fixed point in time. Longitudinal and experimental research would be needed to obtain certainty regarding causal relationships. Although some scholars have recently started to use longitudinal designs to assess the impact of leadership on student outcomes (Hallinger & Heck, 2011; Heck & Hallinger, 2009), more longitudinal studies are needed (Shamir, 2011). Findings from these studies could shed more light on the causality of the assumed relationships among leadership behavior, organizational conditions, quality of teachers’ work, and school performance.

Though leaders of secondary schools in the Netherlands seem to have no (quantifiable) influence on academic performance, a relationship does exist between school leader behavior and school outcomes. Our findings discussed above have shown that the school leader can indirectly promote teachers’ work if the school culture is sufficiently development oriented—in other words, when careful decision making, teacher commitment, cooperation, professional development, and innovation are valued in the organization. Teachers’ work leads to better flow through and to an improvement (albeit modest) in academic performance. The latter effect is, however, so small that the contribution of school leadership can no longer be demonstrated, possible because of the homogeneity of the sample (HAVO stream). Further research could focus more on schools with a strong orientation toward development. These are probably not the schools with the best academic performance. What are the exact characteristics of these schools and their leaders, and how does academic performance change over time, under their influence? Findings from these questions could help us to find more conclusive answers to the question of whether school leaders do make a difference in the lives of their students.

Footnotes

Appendix

Intraclass Correlations of Aggregated Teacher- and Student-Level Variables

School leader behavior
Rational goal practices	.22
Internal process practices	.34
Human relations practices	.31
Open system practices	.22
School culture
Rational goal culture	.17
Internal process culture	.11
Human relations culture	.15
Open system culture	.22
School organization
Rational goal	.18
Internal process	.21
Human relations	.27
Open system	.17
Teachers’ work and student engagement
Teacher work	.07
Student engagement	.09

Note. The intraclass correlations denote the amount of variance situated at the school level for variables measured at the school or student level. In the analysis, these variables were aggregated at the school level.

Declaration of Conflicting Interests

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) disclosed receipt of the following financial support for the research and/or authorship of this article: This research was supported by a grant from the National Scientific Organization of the Netherlands (NWO).

Notes

Bios

Gerdy ten Bruggencate is currently working at the teacher training institute ELAN of the University of Twente in the Netherlands. Her research interests include educational effectiveness and the impact of school leaders on school performance.

Hans Luyten is an Associate Professor at the Faculty of Behavioral Sciences (Department of Educational Organization and Management) of the University of Twente in the Netherlands. His research interests include educational effectiveness, social inequality, educational policy, and international comparisons.

Jaap Scheerens is Professor Emeritus at the University of Twente, where he led the Department of Educational Organization and Management. He has been a project leader of numerous international research projects funded by the European Union, and a consultant for international organizations like OECD, UNESCO, and the World Bank. His publications are in the areas of school management, decentralization in education, school effectiveness, and educational evaluation and monitoring.

Peter Sleegers is Professor of Educational Organization and Management at the University of Twente in the Netherlands. He has published extensively on leadership, innovation, and educational policy in more than 40 scientific journal articles and several edited books. Current research projects are studies into the effects of educational leadership on student motivation for school, longitudinal research into sustainability of reforms, and design studies into professional learning communities.

References

Bluedorn

A. C.

Lundgren

E. F.

(1993). A culture-match perspective for strategic change. Research in Organizational Change and Development, 7, 137-179.

Bossert

S. T.

Dwyer

D. C.

Rowan

Lee

G. V.

(1982). The instructional management role of the principal. Educational Administration Quarterly, 18(3), 34-64.

Bredschneider

(1993). A study of student participation in and identification with one Ontario secondary school (Unpublished manuscript). Ontario Institute for Studies in Education, University of Toronto, Toronto, Canada.

Bryk

A. S.

Sebring

P. B.

Allensworth

Luppescu

Easton

J. Q.

(2010). Organizing schools for improvement: Lessons from Chicago. Chicago, IL: University of Chicago Press.

Cameron

K. S.

Freeman

S. J.

(1991). Cultural congruence, strength, and type: Relationships in effectiveness. Research in Organizational Change and Development, 5, 3-58.

Chin

J. M.-C.

(2007). Meta-analysis of transformational school leadership effects on school outcomes in Taiwan and the USA. Asia Pacific Education Review, 8, 166-177.

De Maeyer

Rymenans

Van Petegem

Van den Bergh

Rijlaarsdam

(2007). Educational Leadership and pupil achievement: The choice of a valid conceptual model to test effects in school effectiveness research. School Effectiveness and School Improvement, 18, 125-145.

Denison

D. R.

Hooijberg

Quinn

R. E.

(1995). Paradox and performance: Toward a theory of behavioural complexity in managerial leadership. Organization Science, 6, 524-541.

DiPadova

L. N.

Faerman

S. R.

(1993). Using the competing values framework to facilitate managerial understanding across levels of organizational hierarchy. Human Resource Management, 32, 143-174.

10.

Driessen

(2002). Sociaal-etnische schoolcompositie en onderwijsresultaten. Effecten van posities, concentratie en diversiteit [Socioethnic school composition and attainment: Position, concentration and diversity effects]. Pedagogische Studiën, 79, 212-230.

11.

Dronkers

(2004). Do public and religious schools really differ? Assessing the European evidence. In Wolf

P. J.

Macedo

(Eds.), Educating citizens: International perspectives on civic values and school choice (pp. 287-314). Washington, DC: Brookings Institution.

12.

Dukelow

G. A.

(1993). A statistical analysis of educational variables that influence high school students’ grades and participation (Unpublished master’s thesis). University of Victoria, Victoria, Canada.

13.

Dwyer

D. C.

Barnett

B. G.

Lee

G. V.

(1987). The school principal: Scapegoat or the last great hope? In ASCD Yearbook, Leadership: Examining the elusive (pp. 30-46). Alexandria, VA: Association for Supervision and Curriculum Development.

14.

Dwyer

D. C.

Lee

G. V.

Barnett

N. N.

Filby

Rowan

Alpert

Kojimoto

(1985). Understanding the principal’s contribution to instruction: Seven principals, seven stories (Vols. 1-7, methodological appendix). San Francisco, CA: Far West Laboratory for Educational Research and Development.

15.

Engels

Aelterman

Schepens

Van Petegem

(2003). Het welbevinden van leerlingen in het secundair onderwijs in Vlaanderen [The well-being of pupils in Flemish secondary schools]. Pedagogische Studiën, 80, 192-209.

16.

Finn

J. D.

Cox

(1992). Participation and withdrawal among fourth-grade pupils. American Educational Research Journal, 29, 141-162.

17.

Hallinger

(1994). A resource manual for the Principal Instructional Management Rating Scale. Nashville, TN: Vanderbilt University, Center for Advanced Study of Educational Leadership.

18.

Hallinger

(2003). Leading educational change: Reflections on the practice of instructional and transformational leadership. Cambridge Journal of Education, 33, 329-352.

19.

Hallinger

Bickman

Davis

(1996). School context, principal leadership and student reading achievement. Elementary School Journal, 96, 527-549.

20.

Hallinger

Heck

R. H.

(1996). Reassessing the principal’s role in school effectiveness: A review of empirical research, 1980–1995. Educational Administration Quarterly, 32, 5-44.

21.

Hallinger

Heck

R. H.

(1998). Exploring the principal’s contribution to school effectiveness: 1980–1995. School Effectiveness and School Improvement, 9, 157-191.

22.

Hallinger

Heck

(2011). Exploring the journey of school improvement: Classifying and analyzing patterns of change in school improvement processes and learning outcomes. School Effectiveness and School Improvement, 22, 149-173.

23.

Heck

R. H.

Hallinger

(2000). Next generation methods for the study of leadership and school improvement. Educational Administration Abstracts, 35, 141-162.

24.

Heck

R. H.

Hallinger

(2009). Assessing the contribution of distributed leadership to school improvement and growth in math achievement. American Educational Research Journal, 46, 626-658.

25.

Heck

R. H.

Larsen

T. J.

Marcoulides

G. A.

(1990). Instructional leadership and school achievement: Validation of a causal model. Educational Administration Quarterly, 26, 94-125.

26.

Houtveen

Vermeulen

van de Grift

(1993). Bouwstenen voor onderzoek naar de kwaliteit van scholen [Measuring the quality of schools]. Utrecht, Netherlands: ISOR.

27.

Houtveen

A. A. M.

Voogt

J. C.

Vegt

A. L.

van

Grift

W. J. C. M.

(1995). Zo zijn onze manieren. Onderzoek naar de organisatiecultuur van scholen [These are our manners: Research into the organizational culture of schools]. Utrecht, Netherlands: ISOR.

28.

Kalliath

T. J.

Bluedorn

A. C.

Gillespie

D. F.

(1999). A confirmatory factor analysis of the competing values framework. Educational and Psychological Measurement, 59, 143-158.

29.

Krüger

M. L.

(1994). Sekseverschillen in schoolleiderschap [Gender differences in school leadership] (Unpublished thesis). Samsom H. D. Tjeenk Willink, Alphen aan den Rijn, Netherlands.

30.

Krüger

M. L.

Witziers

(2003). Ontwikkelingen in het denken over leiderschap [Development in thoughts about leadership]. In Creemers

(Ed.), Handboek Schoolorganisatie en Onderwijsmanagement (pp. E3000-1-E3000-28). Alphen aan den Rijn, Netherlands: Samsom.

31.

Krüger

M. L.

Witziers

Sleegers

P. J. C.

(2007). The impact of school leadership on school level factors: Validation of a causal model. School Effectiveness and School Improvement, 18(1), 1-20.

32.

Leithwood

Aitken

Jantzi

(2000). Making schools smarter: A system for monitoring school and district progress (2nd ed.). Thousand Oaks, CA: Corwin Press.

33.

Leithwood

Day

Sammons

Harris

Hopkins

(2006). Successful school leadership: What it is and how it influences student learning (Research Report 800). London, UK: Department for Education.

34.

Leithwood

Jantzi

(2000). Principal and teacher leadership effects: A replication. School Leadership & Management, 20, 415-434.

35.

Leithwood

Jantzi

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

36.

Leithwood

Levin

(2005). Assessing school leader and leadership programme effects on pupil learning (Research Report 662). London, UK: Department for Education.

37.

Leithwood

Louis

K. S.

Anderson

Wahlstrom

(2004). Review of research: How leadership influences learning. New York, NY: Wallace Foundation.

38.

Louis

K. S.

Dretzke

Wahlstrom

(2010). How does leadership affect student achievement? Results from a national US survey. School Effectiveness and School Improvement, 21, 315-336.

39.

Lugthart

Roeders

Bosker

Bos

(1989). Effectieve schoolkenmerken in het voortgezet onderwijs: Deel I: een literatuuroverzicht [Effective school characteristics in secondary education: Part I: A literature review]. Groningen, Netherlands: Instituut voor Onderwijsonderzoek (RION).

40.

Luyten

(1994). School effects: stability and malleability (thesis). Enschede: University of Twente.

41.

Luyten

Scheerens

Visscher

A. J.

Maslowski

Witziers

Steen

(2005). School factors related to quality and equity. Results from Pisa 2000. Paris: OECD.

42.

Marks

H. M.

Printy

(2003). Principal leadership and school performance: An integration of transformational and instructional leadership. Educational Administration Quarterly, 29, 370-397.

43.

Marzano

R. J.

(2000). A new area of school reform: Going where the research takes us. Aurora, CO: Mid-continent Research for Education and Learning.

44.

Marzano

R. J.

Waters

McNulty

B. A.

(2005). School leadership that works: From research to results. Alexandria, VA: Association for Supervision and Curriculum Development.

45.

Maslowski

(2001). School culture and school performance: An explorative study into the organization culture of secondary schools and their effects (Unpublished doctoral dissertation). Twente University, Enschede, Netherlands.

46.

Mulford

(2003). School leaders: Changing roles and impact on teacher and school effectiveness. Paris, France: Organisation for Economic Co-operation and Development. Retrieved from http://www.oecd.org/dataoecd/61/61/2635399.pdf

47.

Mulford

Silins

(2003). Leadership for organisational learning and improved student outcomes—What do we know? Cambridge Journal of Education, 33, 175-195.

48.

Mulford

Silins

(2011). Revised models and conceptualisation of successful school principalship that improves student outcomes. International Journal of Educational Management, 25, 61-82.

49.

Mulford

Silins

Leithwood

(2004). Educational leadership for organisational learning and improved student outcomes. Dordrecht, the Netherlands: Kluwer Academic Publishers.

50.

Opdenakker

M.-C.

Van Damme

(2006). Differences between secondary schools: A study about school context, Group composition, school practice, and school effects with special attention to public and Catholic schools and types of schools. School Effectiveness and School Improvement, 17, 87-117.

51.

Organisation for Economic Co-operation and Development. (2003). Manual for the PISA 2000 database. Paris, France: Author.

52.

Parsons

(1960). Structure and process in modern society. New York, NY: Free Press.

53.

Pounder

D. G.

Ogawa

R. T.

Adams

E. A.

(1995). Leadership as an organization-wide phenomena: Its impact on school performance. Educational Administration Quarterly, 31, 564-588.

54.

Quinn

R. E.

(1988). Beyond rational management: Mastering paradoxes and competing demands of high performance. San Francisco, CA: Jossey-Bass.

55.

Quinn

R. E.

(1998). Persoonlijk meesterschap in management: Voorbij rationeel management [Personal mastery in management: Beyond rational management]. Schoonhoven, Netherlands: Academic Service.

56.

Quinn

R. E.

Cameron

(1983). Organizational life cycles and shifting criteria of effectiveness: Some preliminary evidence. Management Science, 29, 33-51.

57.

Quinn

R. E.

Kimberly

J. R.

(1984). Paradox, planning and perseverance: Guidelines for managerial practice. In Kimberly

J. R.

Quinn

R. E.

(Eds.), Managing organizational transitions (pp. 259-313). Homewood, IL: Dow Jones-Irwin.

58.

Quinn

R. E.

McGrath

M. R.

(1985). The transformation of organizational cultures: A competing values perspective. In Frost

P. J.

Moore

L. F.

Lundberg

C. C.

Martin

(Eds.), Organizational culture (pp. 315-334). Beverly Hills, CA: Sage.

59.

Quinn

R. E.

Rohrbaugh

(1983). A spatial model of effectiveness criteria: Towards a competing values approach to organizational analysis. Management Science, 29, 363-377.

60.

Quinn

R. E.

Spreitzer

G. M.

(1991). The psychometrics of the competing values culture instrument and an analysis of the impact of organizational culture on quality of life. Research in Organizational Change and Development, 5, 115-142.

61.

Robinson

V. M. J.

Lloyd

Rowe

K. J.

(2008). The impact of leadership on student outcomes: An analysis of the differential effects of leadership type. Educational Administration Quarterly, 44, 635-674.

62.

Sammons

Day

(2011). Exploring the impact of school leadership on pupil outcomes: Results from a study of academically improved and effective schools in England. International Journal of Educational Management, 25, 83-101.

63.

Saris

W. E.

Stronkhorst

L. H.

(1984). Causal modelling in nonexperimental research: An introduction to the LISREL approach. Amsterdam, Netherlands: Sociometric Research Foundation.

64.

Scheerens

(1989). Wat maakt scholen effectief? [What makes schools effective]. Gravenhage: Instituut voor Onderzoek van Onderwijs SVO.

65.

Scheerens

Bosker

(1997). The foundations of educational effectiveness. Oxford: Pergamon.

66.

Scheerens

Witziers

(2005). Educational leadership and student performance. Enschede: University of Twente.

67.

Shamir

(2011). Leadership takes time: Some implications of (not) taking time seriously in leadership research. Leadership Quarterly, 22, 307-315.

68.

Silins

Mulford

(2002). Leadership and school results. In Leithwood

Hallinger

(Eds.), Second international handbook of educational leadership and administration (pp. 561-613). Dordrecht, Netherlands: Kluwer.

69.

Silins

Mulford

Zarins

(2002). Organisational learning and school change. Educational Administration Quarterly, 38, 613-642.

70.

Sleegers

P. J. C.

(1991). School en beleidsvoering: Een onderzoek naar de relatie tussen het beleidsvoerend vermogen van scholen en het benutten van de beleidsruimte door scholen [Schools and policy: A study into the relation between policy making capacity and the use of policy space in schools] (thesis). Nijmegen: Nijmegen University Press.

71.

Thompson

M. D.

(2000). Gender, leadership orientation, and effectiveness: Testing the theoretical models of Bolman & Deal and Quinn. Sex Roles, 42, 969-992.

72.

Thoonen

E. E. J.

Sleegers

P. J. C.

Oort

F. J.

Peetsma

T. T. P.

Geijsel

F. P.

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

73.

Veenstra

(1999). Leerlingen, klassen, scholen. Prestaties en vorderingen van leerlingen in het voortgezet onderwijs [Pupils, classes, schools: Pupils’ attainment and achievement in secondary education]. Amsterdam, Netherlands: Thela.

74.

Verschuren

P. J. M.

(1991). Structurele modellen tussen theorie en praktijk [Structural models between theory and practice]. Utrecht, Netherlands: Het Spectrum.

75.

Walberg

H. J.

(1984). Improving the productivity of America’s schools. Educational leadership, 41(8), 19-27.

76.

Witziers

Bosker

R. J.

Krüger

M. L.

(2003). Educational leadership and student achievement: The elusive search for an association. Educational Administrative Quarterly, 39, 398-425.

77.

Zammutto

R. F.

Krakower

J. Y.

(1991). Quantitative and qualitative studies on organizational culture. Research in Organizational Change and Development, 5, 83-114.