Contents

Learning From Leadership: Investigating the Links to Improved Student Learning

Click here to download the full report:
 Learning From Leadership: Investigating the Links to Improved Student Learning

Key Findings

  • On average, schools experience fairly rapid principal turnover: about one new principal every three to four years.
  • Rapid principal turnover has moderately negative effects on school culture.
  • Rapid principal turnover seems not to have much effect on classroom content or instruction.
  • Rapid principal turnover explains a modest but significant amount of variation in student achievement across schools.
  • Coordinated forms of leadership distribution have the potential to mitigate at least some of the negative consequences of rapid principal turnover.
  • Principals newly assigned to schools who initially work within the existing culture of their schools, rather than attempting to quickly, substantially change it, are more likely to avoid negative turnover effects.

Introduction

Our analysis of principal turnover and its effects appears in Part Two of our final report because principal turnover is fostered in part by district policies. Some districts, for example, still have policies requiring regular principal rotation.193 Many districts now have increased accountability requirements for schools and principals to the point where potential candidates may be deterred from applying for leadership positions.194 Also, it is typically the district‘s responsibility to find replacements for departing principals, whatever the reasons for departure. Principal turnover is a problem districts help to create, and so must help to resolve.

While principal turnover is inevitable in every school, too rapid turnover—or succession—is widely thought to present significant challenges to districts and schools. Many districts, for example, struggle to find suitably skilled and experienced principals, partly because of the above-average replacement rates required by a bulge in the proportion of incumbents currently becoming eligible for retirement. It is far from a trivial problem. Schools experiencing exceptionally rapid principal turnover, for example, are often reported to suffer from lack of shared purpose, cynicism among staff about principal commitment, and an inability to maintain a school-improvement focus long enough to actually accomplish any meaningful change.195

Our efforts to learn more about the nature and consequences of rapid principal turnover have been guided by five questions:

  • How frequently does principal turnover occur in the average school?
  • Does principal turnover significantly affect conditions across the school and in classrooms?
  • Does principal turnover significantly affect student achievement?
  • Do coordinated forms of distributed leadership, as some evidence suggests, have the potential to reduce negative influences arising from frequent principal turnover?
  • What, if anything, can incoming principals do to minimize the negative effects of rapid principal turnover?

Prior Evidence

School and Classroom Conditions Influenced by Rapid Turnover

For the most part, school leaders influence students indirectly. Efforts to increase leaders‘ influence on students will therefore depend on identification of factors that mediate what leaders do. Rowan‘s (1996) framework identifies one promising set of mediators. This framework suggests that the performance of teachers—clearly the most powerful mediator of leaders‘ influence on students196--is a function of their abilities, motivation, and the nature of the settings (or conditions) in which they work. It follows that leaders‘ influence on students will depend on their success in improving teachers‘ abilities, motivations, and working conditions. In light of this background, we focus here on teachers‘ school and classroom working conditions, exploring the degree to which variations in the rapidity of principal turnover may influence school culture, as well as curriculum and classroom instruction.

We know from prior research that the impact of school leadership on student achievement is mediated by school culture: shared values, norms, and contexts.197 Healthy school cultures correlate strongly with increased student achievement and motivation.198 School leaders who build productive "cultures of change"199 can enhance teacher motivation, build teacher capacity, promote teacher efficacy,200 and create the professional unity and cohesion required for effective instruction201 and student success.202 Principals have a strong effect on school culture and on classroom conditions—which, in turn, affect student success.203

Principal Turnover Effects

Evidence about principal turnover often associates it with negative consequences. Grusky (1963) and Bruggink (2001) report that changing principals disrupts staff members‘ focus on improving student achievement. Others argue that principal turnover disrupts school change processes when a leader who supports a project leaves and is replaced by a leader with different priorities;204 when a "charismatic principal departs who had 'radically transformed' the school in four or five years";205 or when there is a poor "fit" between the leader and school.206 207

While principal turnover often has negative consequences, the outcome is not consistently negative. Partlow (2004), for example, argues that student achievement operates independently of changes in school leadership. Miskel and Owens‘ (1983) study of 89 schools in the midwest region of the U.S. found that principal succession had no significant effects on staff members‘ job satisfaction, communication, instruction, school discipline, or school climate. But there is considerable evidence to the contrary.

Leadership turnover does not have to occur every year or two to be problematic. Even in cases where a principal‘s tenure extends over a period of several years, teachers may remain alienated when principal turnover is the result of a district leadership rotation policy.208 Teachers may become cynical and resistant to change because of the "revolving door syndrome"—the uncertainty and instability turnover causes, and the perception of the new leader as a "servant to the system."209

Some teachers develop a deep distrust of the new leader‘s loyalty, suspecting that he or she is more committed to career advancement than the long-term welfare of the school and community. Under conditions of regular principal turnover, teachers learn to "wait them out."210 That is, teachers maintain barriers between themselves and new leaders, ensuring that their school‘s culture becomes self-sustaining, "immunized," and impervious to change instigated by those in formal leadership positions.211

Frequency of Principal Turnover

Nevertheless, principal turnover is nevertheless inevitable in all schools. It is therefore important to ask about the optimum frequency of turnover: how frequent is too frequent? How long is too long for a principal to stay in one school? We have been guided by two theoretical perspectives, stage theory and change theory, in our efforts to answer these questions.

Stage theory conceptualizes leadership succession as a process with distinct phases and demands, rather than a singular event.212 Patterns in the process have been identified, and the ways in which each phase of the succession process shapes and influences the outcome of subsequent phases have been described.213 Most stage models predict that it takes at least five to seven years to build relationships of trust that can serve as a foundation for movement to later stages of the succession process—"consolidation and refinement," in Gabarro‘s (1987) terms. According to this view, principals need to be in their schools for about five years in order to have a positive impact. After five years, the principal‘s work may continue, but continuity from then on does not seem to be related to continued improvement.

Change theory includes a concept of change as a process of initiation or adoption, implementation, and institutionalization or continuation.214 According to Fullan (1991), all successful schools experience an "implementation dip," a drop in performance and confidence when people are faced with innovations that demand new knowledge, skills, strategies, and relationships. People who are experiencing fear and anxiety about their capacity to manage change require leaders they can trust, as well as leaders who are empathetic and socially skilled.

Fullan asserts that, while there is no standard formula for changing the culture of an organization, sustainable improvement requires several years of effort to work through complex cultural issues such as resistance to change and acculturation of the new leader.215 Turnover that occurs every two or three years makes it unlikely that a principal will get beyond the stages of initiation and early implementation. Like stage theory, then, change theory also argues that leader-tenure much beyond three years is necessary if significant improvements are to occur in response to a principal‘s initiatives.

This leaves us with questions about the upper limit of a principal‘s tenure in a school: is there a "best by" date for principals, beyond which they should move on, or be moved on? Does a principal become stale or stagnant if he or she remains in the position for too long? We have little hard evidence bearing on this question, but that fact has not prevented some districts from creating policies reflecting the professional experiences of their staffs. District superintendents, for example, often justify their principal rotation policies as a means of reinvigorating school administrators who seem to reach their peak effectiveness after five to seven years.216 Realistically, there is bound to be enormous variation among individual principals, suggesting that districts should avoid a one size fits-all approach to principal succession.

Distributed Leadership

Evidence about the effects of principal turnover assumes that a considerable proportion of the leadership in schools is delivered by the principal. But suppose school leadership was more dispersed or distributed. Would more leadership distribution within a school moderate the effects of rapid principal turnover, as some are now suggesting?217 Part One of this report reviewed research and theory about distributed leadership in some depth, as well as reporting new evidence on the concept. For present purposes, then, we describe only the conceptual choices we have made for this sub-study of principal turnover.

Among the many different conceptions of leadership distribution in the literature,218 we have chosen to view it through a lens developed by Leithwood, Mascall, and Strauss (2009). Leithwood et al. describe four patterns of leadership distribution observed in schools:

  • Planful Alignment. In this pattern, leaders‘ tasks and functions result from prior, planful thought by organizational members, and functions are rationally distributed in ways comparable to Gronn‘s (2009) holistic notion of "institutionalized practice."
  • Spontaneous Alignment. In this pattern, leadership tasks and functions are distributed with little or no planning, and tacit or intuitive decisions determine who should perform which leadership functions. Fortuitous, positive, short-term working alliances evolve.
  • Spontaneous Misalignment. Here there are disjunctions among leadership functions, causing unpredictable outcomes and negative effects on short- and long-term organizational effectiveness and productivity.
  • Anarchic Misalignment. This pattern is similar to the condition Hargreaves and Fink (2006) describe as anarchy: members of the organization reject or compete with one another in making claims of leadership regarding decisions, priorities, and activities.

Recent scholarship suggests that leadership distribution may moderate the effects of principal turnover on school culture. Hargreaves and Fink (2006) conclude that the post-succession process is best managed when the departing leader leaves a legacy of distributed leadership marked by shared vision, investment, and capacity that ensures the sustainability of school improvement initiatives. This leads us to hypothesize that in times of frequent principal turnover (leader changes every one, two, or three years)— involving leaders shaped by different experiences, priorities, and leadership styles— teachers are encouraged (or forced) to take leadership into their own hands, and to develop some stability by means of a self-sustaining professional culture that operates independently of the principal. The result then will be distributed leadership in one form or another.

Where teacher leadership evolves strategically (planned and aligned with school goals), a self-sustaining culture can become both collaborative and productive. When leadership distribution is neither planned nor aligned, then the self-sustaining culture drifts, gradually loses its collective sense of vision and purpose, and becomes increasingly balkanized; each teacher focuses on his or her classroom, works in relative isolation from colleagues, and takes responsibility only for his or her own work. The result is an ineffective organization of "neglect" and "anarchy," where student achievement may remain unchanged, or even deteriorate.

New Evidence

Method

We used quantitative and qualitative methods to answer the five questions described in the Introduction to this section. Data from quantitative studies derive from responses to questions we posed about average principal turnover rates, effects on school culture, curriculum, and instruction, and student achievement. Data from qualitative studies derive from responses to questions we posed about the potential for some patterns of distributed leadership to mitigate the negative effects of rapid principal turnover, and what, if anything, incoming principals might do, to minimize negative turnover effects.

Quantitative evidence. For this evidence we examined responses to 36 of the 104 items included in the first teacher survey. The construct for school culture comprises the following seven items, ranked on a 6-point scale, using the stem To what extent do you agree or disagree with the following statements:

  • Disruptions of instructional time are minimized.
  • Most teachers in our school share a similar set of values, beliefs, and attitudes related to teaching and learning.
  • Students feel safe in our schools.
  • In our school, we have well defined learning expectations for all students.
  • Students in our school meet or exceed clearly defined expectations.
  • We provide opportunities for students to discuss the effects of intolerance on their lives.
  • Our student assessment practices reflect our curriculum standards.

The construct for classroom, curriculum and instruction comprises the following five items, ranked on a six-point scale, using the same stem:

  • I have sufficient written curricula on which to base my lessons.
  • My instructional strategies enable students to construct their own knowledge.
  • I maintain a rapid pace of instruction in my classes.
  • I feel adequately equipped to handle student behavior in my class.
  • Our school/district provides a rigorous core curriculum for most of our students.

The achieved sample for this sub-study was 2,570 teachers (a 78% response rate) from a total of 80 schools in which four or more teachers completed usable surveys and for which usable student achievement data were available. The principal survey provided data on the number of principals in the school over the past 10 years for those same 80 schools.

To measure student achievement across schools, we collected data from state websites. These data were school-wide results on state-mandated tests of language and mathematics at several grade levels over three years (2003 to 2005). For purposes of this study, a school‘s student achievement level is represented by the percentages of students meeting or exceeding the proficiency level (usually established by the state) on language and mathematics tests. We averaged these percentages across grades and subjects in order to increase the stability of scores, producing in a single achievement score for each school for each of three years.

Our data on student achievement for these schools covers only the most recent three years, yet the turnover of principals is measured over the past 10 years. The premise is that there would be a cumulative effect of principal turnover during this time, which would appear as an overall low level of achievement in the schools in the most recent three years.

Qualitative evidence. From the 40 schools included in the first round of site visits, we selected the four with the highest principal-turnover rates as case study schools, based on the principal survey question about the number of principals that those 40 schools had had over the past 10 years. Each of these schools was located in a different state, and the states were widely distributed geographically. We then conducted NVivo coding searches within the transcripts of the interviews with the principal and five teachers in each of the four schools.

Principal Turnover: Frequency and Effects on Schools, Classrooms, and Students

Table 2.4.1 reports the means, standard deviations, and scale reliabilities (Cronbach‘s alpha) of variables for this sub-study. As the first row in this table indicates, the average number of principals in the school over the past 10 years was 2.78, for an average length of tenure of 3.6 years per principal. The standard deviation for this measure is a relatively large (1.34).

 

We calculated Pearson‘s correlation coefficients to assess the relationships between meditating variables, the independent variable (the number of principals in the school in the past 10 years), and the dependent variable (student achievement). Table 2.4.2 summarizes these relationships. Relationships among principal turnover and measures of school and classroom conditions are negative.

 

Principal turnover is moderately and negatively correlated with school culture and with classroom curriculum and instruction; it has a weak negative relationship with student achievement. School culture is strongly related to both curriculum and instruction and student achievement; curriculum and instruction is moderately related to student achievement.

Figure 8 summarizes the results of a path model (using LISREL) we used to explore the relationships among these variables more precisely.

 

Figure 8: Testing the mediated effects of principal turnover on student achievement

This model is a good fit with the data (RMSEA = .00; RMR = .02; AGFI = .96; NFI = .99), and it explains 41% of the variation in student achievement. The total effects of principal turnover explain 24% of the variation in student achievement. Principal turnover has significant and moderately negative effects on school culture (-.37), although school culture has moderately strong, significant, effects on student achievement (.68). The effects of turnover on curriculum and instruction are insignificant, and the measure of classroom curriculum and instruction is negatively, but very weakly, related to student achievement. It is interesting to see that the partial correlations between these mediating variables and student achievement are strong and positive, but the addition of principal turnover to the model reduces the effect of curriculum and instruction on student achievement to a very low level (-.06).

In sum, results suggest that principal turnover has significant negative effects on student achievement. These effects are mediated more by school-level than classroomlevel conditions. The weaker impact of principal turnover on classroom variables might suggest that teacher classroom practice is in some way buffered from direct effects of changes in principal leadership. We speculate that teachers may continue to feel secure in their classrooms, regardless of the school culture around them. While buffering of this sort limits the negative effects of principal turnover, it may also limit positive effects of a principal‘s improvement efforts.

Leadership Distribution and Leader Turnover Illustrated

Given the significant influence of principal turnover on student achievement, mediated primarily by school culture, we developed four case studies to examine this dynamic in greater detail and to learn what part patterns of distributed leadership play in the relationships. The four schools are profiled below.

Culbertson Elementary School

Culbertson is an urban elementary school with an enrollment of just over 600 students, almost all of whom meet state achievement expectations on the grades 3-5 standardized tests in reading, science, and mathematics. At the time of our study, three principals had been at the school in the last three years, and the current principal was promoted to the post from a district intern position. High principal turnover had become a challenge for the district, in part because a new state retirement policy had induced 20% of the district‘s principals to retire in the year that a new option was announced. To deal with the challenges of principal succession, district leaders established a number of support mechanisms to help new principals acclimatize themselves in their new jobs; these included monthly meetings and a mentoring program with retired principals.

Principal turnover in Culbertson had no measurable impact on student performance, positively or negatively. From 2003 through 2006, the percentage of students meeting or exceeding state norms held consistently to a range in the high nineties across all grades and subjects.

The principal three years earlier had explicitly encouraged teachers to assume leadership roles in the school, in accordance with district policies that supported the designation and implementation of formal teacher-leader positions. The principal also saw to it that this leadership distribution was both planful and well aligned with the school‘s goals. By the time of our study, leadership had become distributed to a considerable extent, and teacher-leaders were able to help introduce incoming principals to the school culture. Since student achievement was not a source of concern in the school, there was little pressure to bring about any radical changes in teaching and learning. Consequently, new principals did not feel compelled to innovate either rapidly or radically.

A planful alignment pattern of leadership distribution had stood the staff in good stead through two succeeding principals. The teachers were able to work together, share the leadership for that work, and sustain the learning of their students, despite changes in principals. The current principal seemed to be in tune with this approach to distributed leadership.

Molina Elementary School

Molina is a small elementary school in an urban community. At the time of our study, 31% of the students in the district qualified for free and reduced-price lunches, and the school had a 35% non-white (mostly Hispanic) population. Student achievement scores were uneven across grades and subjects: strong in grade 3, but weak in grade 4; strong in reading but not in writing. In the three years for which we had data, however, overall levels of achievement had been improving.

State policy on principal retirement was in flux at the time of our study. This was situation was encouraging some principals who were facing an uncertain future to get out "while the getting is good." Over the five years prior to our study there had been a high level of retirements across the district, and Molina had not been immune to this trend, having had four principals in that period of time. District office staff remarked on early retirement as an ongoing problem and a significant source of stress on the system‘s capacity to train and replace its district and school leaders. The pressures of early retirement—as many as 20% of the total number of principals in the district changing in any one year—had spawned district initiatives to address the turnover problem. As a result of a District Literacy Initiative, there had been a structural shift to create teacherleader Literacy Coaches in each school. Molina had five of these Literacy Coaches, with an additional Literacy Coach position scheduled for the next year.

Cultural and emotional turmoil was apparent in Molina because principal turnover had been accompanied by fundamental changes in philosophy and leadership style. The four principals in five years at Molina had had different personalities and insufficient time to establish trust and rapport. Long-serving support-staff members—familiar to teachers, parents, and students—were able to take on certain leadership roles in light of the annual change of principals. This case provides, accordingly, some evidence for our expectation that greater distribution of leadership would ameliorate some negative effects of rapid principal turnover. But life in schools is not shaped by a single variable. In the case of Molina, a high rate of teacher turnover exacerbated the effects of rapid principal turnover, thereby muting the potential values associated with more teacher leadership.

Molina‘s pattern of distributed leadership could best be described as spontaneous misalignment. There was no planned effort to share leadership, nor was there a sense that leadership as it evolved was being aligned with school goals. Despite the best efforts of the teachers to provide leadership for their school, along with efforts by the district to establish formal teacher-leadership positions, the combined effects of frequent principal turnover and frequent teacher turnover made it impossible for this school to sustain any momentum in its improvement efforts.

Blake Elementary School

Blake is a small elementary school in an inner-city district. At the time of our study, a high proportion of its student population was black, and a significant proportion of the community lived below the poverty line. Student achievement was not high; achievement levels in grade 3 and 4 Communication Arts and grade 3 Math tests were at or above state averages, but results for grade 4 and 5 Math and grade 5 Communication Arts remained below state averages. The number of children achieving at the state standard in literacy, however, had been increasing steadily over the past three years.

Three administrators had been appointed to Blake in seven years. There had also been a significant number of new senior administrators in the district in the past two years: a new superintendent and three new directors at the district level, and three new administrators at the school level, across a total of seven schools.

Blake‘s story has much to do with a charismatic principal whose vision for a Professional Learning Community (PLC) had shaped the school‘s identity, structure, and culture. While the principal in position at the time of our study had not initiated the PLC concept, she had chosen to carry on with it as the central feature of the school‘s shared vision. Thus, the PLC provided the foundation for cultural and structural continuity from the previous principal to the current principal.

Principal turnover did not result in cultural chaos or teacher alienation at Blake, because there was a clear and planned focus for school culture and instruction. This school-wide focus survived rapid principal turnover, partly because collaborative structures were well established and accepted and partly because the new principal‘s philosophy and practices supported the existing school culture. Blake therefore provides another case of planful alignment in the distribution of leadership. Teachers at Blake had developed a shared vision for the school and were able to sustain it despite the change in leadership. Indeed, the new principal‘s support for the existing vision became a key element in further developing a positive culture in this school.

Rhodes Middle School

Rhodes Middle School is located in a low-income community; at the time of our study, 13% of the population fell below the poverty line, and 60% of the Rhodes students qualified for free or reduced-price lunch. Four different principals had served at Rhodes in four years, and the student and teacher populations were highly transient. The first of the four principals believed strongly in site-based management and fostered a culture in which teachers learned to rely on their own leadership to get things done. Theirs was an autonomous teacher culture in which each staff member was encouraged to take personal responsibility for her or his own classroom practice, but not much else. Collaboration was not encouraged. Student achievement, however, had been consistently high over the previous three years.

While many teachers at Rhodes seemed satisfied with their autonomous culture and its contribution to sustaining their efforts through frequent principal turnover, the principal current at the time of our study saw professional entrenchment and barriers to administrator influence. This new principal set about changing the culture of the school, without going so far as to dismantle its existing decision-making structures. She aimed for a balance of authority between herself and the staff, given the instability caused by frequent principal succession. She set out to establish a collective focus on instructional practice and data-driven decision-making.

The school seemed to be poised on the cusp of moving from traditional forms of teacher autonomy to a more planful pattern of leadership distribution. The approach of the new principal was more directive than collaborative. But her intention was to create a more collaborative culture, with teachers exercising more leadership across the school as they learned to work together.

Across the Cases

All four schools experienced high rates of principal turnover in the time in question—from a new principal every year, for three or four years, to one every two years, for seven years. In all four schools there had been some attempt at distributing leadership, but each school approached distribution differently, as the culture varied from school to school. While the four schools seem to have little in common beyond rapid principal turnover, two schools found ways to deal productively with changing leadership, while two did not.

Culbertson took a deliberate approach to the distribution of leadership, driven by a principal and district leaders committed to collaborative work and planfully aligned leadership distribution. Blake built a strong professional community, also producing planfully aligned patterns of leadership distribution capable of surviving changes in leadership. In both cases, leadership was distributed among a number of teachers. Despite frequent changes in principals, the supportive cultures developed in these schools continued to thrive.

In the other two schools, there was less success with leadership distribution. In Molina, the district‘s attempts to foster teacher leadership as one response to frequent principal turnover ran afoul of frequent teacher turnover. In Rhodes, the efforts of an earlier principal to foster a high degree of individual teacher autonomy had been sufficiently successful that the principal in place at the time of our study was experiencing considerable difficulty in her efforts to promote collaboration and more leadership distribution. Teachers still remained independent, in a strong culture of individual isolation.

In sum, these cases suggest the following:

  • Leadership distribution has the potential to moderate the negative consequences of rapid principal turnover.
  • Principals have significant leverage in the distribution of leadership across their schools.
  • Planfully aligned patterns of distributed leadership seem likely to contribute most to school improvement efforts once they are established.
  • The challenge of fostering leadership distribution is greatly influenced by the existing culture of the school; autonomous teacher cultures are strong sources of resistance to leadership distribution efforts.
  • While rapid principal turnover has negative effects on student achievement "on average," some individual schools are able to manage rapid turnover in ways that prevent achievement decline. It seems very unlikely, however, that student achievement will improve under most conditions associated with rapid principal turnover.

Implications for Policy and Practice

Three implications for policy and practice emerged from this section of our study.

  1. Districts should aim to keep most principals in their schools for a minimum of four years, and preferably five to seven years. Assuming the principal is working productively with staff and other stakeholders on improving the school, more frequent changes in principals typically results in wasted energy, dissipation of scarce resources and considerable skepticism on the part of teachers that they will receive the support they need when the change process begins to confront the most difficult challenges.
  2. Under conditions of rapid principal turnover, districts need to encourage incoming principals to understand and respect the school-improvement work in which staff members have previously been engaged. Incoming principals will likely have a smoother transition if they see their job as continuing and refining that work. Principals assigned to schools identified as being in need of being "turned around" are clearly exempted from this recommendation.
  3. Incoming principals should not have the sole responsibility to encourage distributed leadership in schools that have previously experienced rapid principal turnover. Under such conditions, districts need to directly encourage and support planfully aligned forms of leadership distribution, providing training and support to staff members in carrying out shared leadership functions. District leaders have a responsibility to help ensure a smooth transition from one principal to the next. This can be done by clarifying the district‘s expectations for the job to be done by the incoming principals, and by participating with teachers and the new principal in initial discussions about expectations for the new principal‘s work. On their own, teachers are in a weak and sometimes risky position with the incoming principal, to argue for continuing attention to the initiatives they have been working on with the outgoing principal and that are showing signs of progress.

< < Previous | Next > >

References

193. Macmillan (2000).

194. Blackmore (1996).

195. Fink & Brayman (2006).

196. Heck (2007).

197. Deal (1993); Nanavati & McCulloch (2003); Senge (1990); and Stoll (1999).

198. Macneil, Prater & Busch (2007); Stolp (1994).

199. Patterson & Rolheiser (2004).

200. Tschannen-Moran, Woolfolk-Hoy, & Hoy (1998).

201. Stewart (2000).

202. Sarason (1982); Schein (1993).

203. (E.g., Ross & Gray (2006); Waters, Marzano & McNulty (2003).

204. Corbett, Dawson, & Firestone (1984).

205. Fullan (1992).

206. Davidson & Taylor (1998, 1999); Ogawa (1995).

207. The notion of "fit" between leader and school is central to district administrators‘ decision-making concerning principal placement.

208. Macmillan (2000).

209. Reynolds et al. (2008).

210. Hargreaves et al. (2003).

211. Macmillan (2000); Macmillan, Meyer & Northfield (2005).

212. Hargreaves et al. (2003).

213. Miskel & Cosgrove (1984); Miskel & Owen (1983); and Ogawa (1991).

214. Fullan (1991, 1993).

215. Fullan (1991); Hargreaves & Fink (2006).

216. Boesse (1991); Rebhun (1995).

217. Harris (2009).

218. E.g., Gronn (2002); MacBeath (2009); and Spillane (2006).