Feature

The Primacy of Superintendent Leadership

The authors' new research finds a strong connection between the work of the district CEO and student achievement by J. Timothy Waters and Robert J. Marzano

Imagine two superintendents, both viewed as strong leaders by their school boards, their communities and their staffs. Let's give these two superintendents names, Jane and David.

Both Jane and David serve in mid-sized school districts with communities comparable in terms of student, teacher and administrator populations. Per-pupil expenditures in each district are at the state average. Both are seen as strong superintendents who hold high expectations for their districts. However, average district-level achievement is approximately 10 percentile points higher in one of these districts than the other.

Our recent meta-analysis of the effects of district-level leadership on student achievement, summarized in the 2006 McREL report "School District Leadership That Works," explains this difference in student achievement. Before sharing our findings, though, let's take a quick comparative look at these superintendents and their districts to better understand how two superintendents, both considered strong leaders, can have very different effects on mean district-level student achievement.

Both superintendents believe in the importance of strong school-level leadership and expect their principals to provide it. They extend considerable autonomy to their principals. Yet there are differences in how much autonomy Jane and David allow in particular areas.

David's view is that meaningful change and improvement in education occur at the school level. Schools are small enough organizational units to initiate and sustain organizational change in a reasonable period of time.

Jane's "theory of action," on the other hand, is that meaningful change and improvement must occur at district and school levels simultaneously. Though the time trajectory of change at the district level may be extended, Jane believes that for change and improvement to be substantial and sustainable, it also must be systemic, which makes the school district and the responsibilities fulfilled by the district critical.

David's Approach
David is convinced instructional decisions are best left to each individual school, principal and teacher. He believes decisions about instruction should be made by those who are closest to students. After all, they were hired for their expertise and understand their students.

He takes seriously the guidance from the total quality management movement to move decision making about core institutional functions to appropriate levels of the system. In his view, instruction is a core institutional function and the appropriate level of decision making is the classroom.

David's approach to setting district goals for student achievement has been to "aggregate up" from individual school goals to establish districtwide goals for achievement. His district, like so many others in the United States, is focused on improved achievement in math and reading. The district goals for achievement in these two areas are that each school will improve sufficiently to meet or exceed state and federal standards for adequate yearly progress.

Because goals for achievement and instruction are set at the school level, and each school's instructional program reflects the knowledge, skills and experience of the principal and teachers, the district professional development program also is decentralized. The district budgets resources for professional development, but each principal, along with his or her teachers, decides how best to use these resources.

David spends a considerable percentage of his time and attention managing the interests and energy of his school board members. They, along with David and his central-office staff, field many questions from parents and other community members about schools, programs and district effectiveness.

The board has a difficult time responding to questions about achievement and instruction because the district's approach is so decentralized. The district office staff is challenged to find ways to support the variety of instructional and professional development programs being delivered in the district.

David is frustrated that districtwide achievement is lower than expected and, despite his efforts, it has not improved annually at an acceptable rate or to an acceptable level. Disappointing levels of student achievement lead to additional questions from board members and the community. David nonetheless remains confident that individual school performance will eventually be reflected in higher district-level achievement.

Jane's Approach
Given Jane's theory that sustainable improvement occurs simultaneously at district and school levels, she takes a different approach to her responsibilities as superintendent. Jane includes her school board members, principals and other key district stakeholders in a goal-setting process that produces broad, five-year district goals for achievement and instruction.

As in David's district, these goals are focused on math and reading. For each goal, Jane's district establishes measurable success/progress indicators and annual performance targets. Jane and her board members review school-level progress on these goals each quarter and consider revisions to annual performance targets based on evidence of progress (or the lack of it). This process helps as Jane, the school board and the principals closely monitor implementation of the districtÕs instructional program.

In Jane's district, the school board also adopts goals for a districtwide instructional program. Jane's instructional staff and her board decide what constitutes good instruction, especially where they have set achievement goals. They adopt a districtwide approach to instruction based on the best available research. It includes a framework for planning units and lessons and the use of research-based instructional strategies. It creates a common vocabulary about instruction for students, teachers, administrators and board members.

Principals in Jane's district closely monitor implementation of the districtÕs instructional program. They conduct routine "walkthroughs" of classrooms to monitor the quality, fidelity, consistency and intensity of implementation of the district's instructional program. They update Jane, who in turn reports to the board on a quarterly basis, on the status of implementation.

The professional development program in Jane's district is designed to build the knowledge and skills teachers and principals need to implement the district instructional program. It is budgeted and coordinated at the district level to ensure a districtwide approach to high-quality professional development that is research-based, ongoing and job embedded. It includes specific and immediate feedback to teachers and principals on the quality and fidelity of implementation of research-based practices.

Having a districtwide approach to instruction allows Jane's central-office staff to more effectively coordinate resources and services to support school-specific needs. Instead of spending time trying to figure out each school's instructional program, district staff devote its energies to helping principals and teachers implement the district's instructional program.

While Jane expects her principals to provide strong leadership in their schools and extends considerable autonomy to them, she makes it clear that she expects them to align their school-level efforts with the district's overall direction. In other words, she and the board set direction at the district level, then grant principals the latitude they need to guide implementation of the district's instructional program, organizational development and school-level change.

Like David, Jane is optimistic about her district and confident in its capacity for producing higher levels of achievement. Jane has reason to be optimistic. Teachers, students, parents, principals and central-office staff understand the districtÕs achievement goals and instructional program. Professional development resources are coordinated, aligned and used to develop research-based practices correlated with the district's goals.

Jane and the principals continually monitor the implementation of these practices and their effects on teaching, student learning, and the people implementing them. They use formative and observational data to make ongoing adjustments to implementation schedules and to professional development programming.

Based on demographics and economics, average district achievement in Jane's district should be identical to David's. However, mean achievement in Jane's district is 10 percentile points higher than David's. Using the results of our most recent analysis of the effects of superintendent leadership, we can explain this difference. Jane's theory of action about the meaningful and sustainable change occurring simultaneously at district and school levels, and her approach to fulfilling these responsibilities, are aligned with our findings.

A Research Grounding
In our study at McREL, we asked the following basic research question at the outset about the effects of superintendent leadership: What is the strength of relationship between leadership at the district level and average student academic achievement in the district?
In addition, we asked these related research questions:

What specific district-level leadership responsibilities are related to student academic achievement?

What specific leadership practices are used to fulfill these responsibilities?

Although not part of our initial set of questions, we are able to answer another question that we believe to be of interest to superintendents and local school board members, but is not specifically focused on superintendent responsibilities and practices: Is there a relationship between length of superintendent service and student achievement?

We think of the answer to this fourth question as a bonus finding that was not initially part of our inquiry.

We conducted our study using meta-analysis, a technique for scientifically synthesizing research findings from smaller studies into a single, large sample. In this case we targeted all available studies conducted in the United States from 1970 through 2005 that met the following criteria:

• Reported a correlation between district leadership or district leadership variables and student academic achievement or allow for the computing or estimating of a correlation, and

• Used a standardized measure of student achievement or some index based on a standardized measure of student achievement.

Of the 4,500 studies conducted during this period, 27 met these criteria. The demographics for these 27 reports were as follows:

Number of districts involved: 2,714

Number of ratings of superintendent leadership: 4,434

Estimated number of student achievement scores: 3.4 million

Key Correlation
The correlation between district leadership and student achievement was .24 (95 percent confidence interval). This correlation is significant at the .05 level.

One way to interpret this .24 correlation is to consider an average superintendent who is at the 50th percentile in terms of leadership abilities and leading a school district where average student achievement is also at the 50th percentile. Now assume the superintendent improves his or her leadership abilities by one standard deviation (in this case, rising to the 84th percentile of all district leaders). Given the correlation between district leadership and student achievement of .24, we would predict that average student achievement in the district would increase by 9.5 percentile points. In other words, average student achievement in the district would rise to the 60th percentile.

Imagine a normal bell-shaped curve to represent the range of achievement in David's district. Now imagine average achievement in David's district at exactly the 50th percentile. Finally, imagine on this same curve average achievement in Jane's district at approximately the 60th percentile, nearly 10 percentile points higher than in David's district. This difference represents the effect of superintendent leadership on student achievement when the superintendent effectively fulfills the responsibilities we have identified.

District leadership responsibilities correlate with student achievement. In addition, the general effect of superintendent leadership, our second research question, sought to identify the specific leadership responsibilities that produce gains in student achievement.

In the responses, we found five district-level leadership responsibilities with a statistically significant (p. 05) correlation with average student academic achievement. They are as follows:

• The goal-setting process;

• Non-negotiable goals for achievement and instruction;

• Board alignment with and support of district goals;

• Monitor progress on goals for achievement and instruction; and

• Use of resources to support the goals for achievement and instruction.

This table includes correlations (or effects) of these five responsibilities with mean district-level student achievement, brief descriptions of them and selected examples of the practices superintendents use to fulfill them.

Perplexing Finding
One set of findings from the meta-analysis that at first appears contradictory involves building-level autonomy within a district. One of the studies we examined reported that building autonomy has a positive correlation of .28 with average student achievement. However, this same study reported that site-based management had a negative correlation with student achievement of minus .16.

Other studies on site-based management reported slightly better results; yet the average correlation between site-based management and student achievement was (for all practical purposes) zero. This apparent contradiction begins to make sense, however, in light of the five district-level leadership responsibilities described above.

How can we find school autonomy positively correlated with student achievement and site-based management exhibiting a negligible or negative correlation with achievement? This question might be answered in at least two of the earlier findings.

The superintendent who implements inclusive goal-setting processes that result in board-adopted non-negotiable goals for achievement and instruction, who assures that schools align their use of district resources for professional development with district goals, and who monitors and evaluates progress toward goal achievement is fulfilling multiple responsibilities correlated with high levels of achievement.

When this superintendent also encourages strong school-level leadership and encourages principals and others to assume responsibility for school success, he or she has fulfilled another responsibility: to establish a relationship with schools. This relationship is characterized by defined autonomy, which is the expectation and support to lead within the boundaries defined by the district goals. The table below shows the correlation of defined autonomy with mean district-level achievement, a brief description of this responsibility and selected examples of practices superintendents use to fulfill this responsibility.

Defined Autonomy and Practices
Superintendent responsibilitiesSelected examples of practices used by superintendent and central office to fulfill superintendent responsibilities
Defined autonomy; superintendent relationship with schools

The superintendent provides autonomy to principals to lead their schools, but expects alignment on district goals and use of resources for professional development.
• Expecting principals to foster and carry out district achievement and instructional goals

• Developing a shared vision and understanding of defined autonomy

• Committing the district and schools to continuous improvement

• Hiring well-qualified teachers

• Establishing a teacher evaluation process that focuses on district instructional program as a priority for principals

• Establishing strong agreed-upon principles/values which direct actions of people

• Ensuring that schools have a clear mission focused on district goals

• Ensuring that all students have the opportunity to learn

• Maintaining high expectations for school performance

• Directing personnel operations to assure a stable yet improving and well-balanced work force

• Allowing for and promoting innovation at the school-level within the context of district goals

• Providing leadership for principals regarding how to implement district goals

A Bonus Result
Our meta-analysis produced an additional finding that initially was not a focus of our study. Two studies that we examined reported correlations between superintendent tenure and student academic achievement. Together, the weighted average correlation from these two studies was a statistically significant .19, which suggests the longevity of the superintendent has a positive effect on the average academic achievement of students in the district. These positive effects appear to manifest themselves as early as two years into a superintendent's tenure.

The positive correlation between the length of superintendent service and student achievement affirms the value of leadership stability and of a superintendent remaining in a district long enough to see the positive impact of his or her leadership on student learning and achievement. Of equal significance is the implication of this finding for school boards as they frequently determine the length of superintendent tenure in their districts.

In his 2005 book Crash Course, Chris Whittle contrasts CEO stability in major corporations with superintendent stability in large urban school districts. Over the last 20 years, Kansas City, Mo., has had 14 superintendents, yielding an average tenure of 1.4 years. Washington, D.C., has had nine superintendents over that time for an average tenure of 2.2 years. During the same time frame, General Electric was run by two CEOs. Federal Express, Microsoft and Dell had one chief executive each.

Whittle, who founded the Edison Schools, asserts that CEO stability at the corporations accounts for a large measure of their success. He argues that the instability of superintendent leadership accounts for much of the low student achievement found in too many school districts. If the stability of superintendents were to approximate the stability of CEO leadership, he claims, school districts likely would experience greater success, assuming superintendents focus on the right priorities and skillfully fulfill their responsibilities. The bonus finding in this truly supports Whittle's conclusion.

Measurable Impact
David and Jane, of course, are fictitious superintendents in fictitious school districts. Their experiences, however, are much closer to fact than fiction and play out in real time in school districts across the country.

Jane's theory of action and her practices are clearly grounded in research based on our findings. In her experience, Jane skillfully fulfilled key leadership responsibilities with statistically significant relationships to student achievement. She worked with her board of education to adopt and support district goals for achievement and instruction. The board supports district-level and school-level leadership in ways that enhance rather than diminish leadership stability.

It is important to note that superintendents cannot fulfill the responsibilities we identified in our research on their own. They need their school boards as well as central-office staff members to share their understanding of these responsibilities and to integrate them consistently into their practice. Along with district-level responsibilities and practices, they must support the school-level leadership responsibilities and practices. When they do, the primacy and impact of superintendent leadership is obvious and measurable.

Tim Waters, a former superintendent, is president and CEO of Mid-continent Research for Education and Learning, 4601 DTC Boulevard, Suite 500, Denver, CO 80237. E-mail: twaters@mcrel.org Robert Marzano is a senior fellow at McREL. They are co-authors of School Leadership That Works (ASCD).