Welcome to the Curriculum & Leadership Journal website.
To receive our fortnightly Email Alert,
please click on the blue menu item below.
Curriculum & Leadership Journal
An electronic journal for leaders in education
ISSN: 1448-0743
Follow us on twitter

School improvement: what does it take to make school systems better?


International benchmarking of school performance continues to attract strong interest at governmental level as well as among Australian education systems. This interest is evident, for example, in the Grattan Institute report Catching up: Learning from the Best School Systems in East Asia, and in the Australian Government's Review of Funding for Schooling, which states that 'funding arrangements should be aimed at achieving an internationally competitive high standard of schooling' (pxiv). In this issue Curriculum Leadership publishes an article covering an earlier report of continued relevance: How the World's Best-performing School Systems Come Out on Top, from McKinsey and Company. The article originally appeared in ISQ Briefings Volume 15 No 2, 2011.

In 2007 McKinsey and Company published a paper which collated research carried out by them between May 2006 and  March 2007 which looked at the qualities of the best-performing schools and systems internationally. This paper clearly identified the vital role of teachers in ensuring best performance in schools and systems and has influenced systems across the world to move their focus on teacher development and teaching standards.

In 2010 McKinsey and Company released a follow-up paper identifying reforms undertaken in 20 systems from a variety of regions, all with improving levels of performance, that might be replicable for school systems everywhere; and outlining what it takes to 'achieve significant, sustained, and widespread gains in student outcomes' (Mourshed, M., Chikioke, C. & Barber, M., 2010). Findings from the study were evaluated on both quantitative and qualitative measures using the results on national and international assessment across multiple subjects, predominantly during the period 1995 to 2010 as well as 200 interviews with system leaders, staff and educators across 20 education systems.

The systems studied were Armenia, Aspire (a US charter school system), Boston (Massachusetts), Chile, England, Ghana, Hong Kong, Jordan, Latvia, Lithuania, Long Beach (California), Madhya Pradesh (India), Minas Gerais (Brazil), Ontario (Canada), Poland, Saxony (Germany), Singapore, Slovenia, South Korea and Western Cape (South Africa). Almost 575 reform interventions carried out by these systems were analysed to ascertain how these interventions 'interacted with each other and with the system's broader context to deliver better outcomes for students' (ibid).

At the heart of the analysis was the development of an extensive database cataloguing the interventions undertaken by schools during the reform period. The interventions were then further categorised into ten areas of impact, such as professional development, accountability and learning models, and then disaggregated into 60 sub-areas. For example, the area of accountability included: performance assessment, school inspections and self-evaluation.

Findings from the research

Some of the major findings from the research were that:

  • despite their different contexts, all improving school systems appear to adopt a similar set of interventions;
  • while context is important, it is secondary to getting the fundamentals right;
  • significant improvement can be achieved within as little as six years;
  • despite spending on education in most OECD countries doubling, and even tripling in some cases, in real terms between 1970 and 1994, student outcomes in many systems either stagnated or regressed; and
  • systems with similar education spending have widely varying levels of performance.

The report showed that no matter where schools start they can all improve and in order to do so system leaders and/or school leaders must integrate three aspects in their improvement journey: the performance stage, which identifies the current position of the system in terms of student outcomes, including making a judgment about where the system sits in the continuum from poor to excellent; the intervention cluster, which identifies the set of interventions necessary to make the desired improvement; and the contextualising stage which is where the system adapts to the prevailing context. The latter includes such elements as the history, culture, politics and structure of the system and the country. The McKinsey researchers found that each of the three stages is associated with a dominant cluster of interventions and, although these interventions change in how they are implemented and in emphasis, depending on where the school is in its improvement journey, 'the intervention pattern is strikingly consistent for systems pursuing similar outcomes', regardless of context.

Two other elements are discussed in the report in relation to system improvement: sustaining and ignition. The first of these refers to how a system puts in place processes to sustain improvement over the longer term and the second describes the conditions necessary to 'ignite' a system into reform.

How systems and schools begin and sustain improvement

A more detailed look at the 20 school systems studied by McKinsey provides rich information about how systems and schools begin and sustain their improvement journeys. Two types of interventions implemented by improving schools were identified: those appropriate to a particular stage of development and those applying equally during all stages. The research emphasised that while each stage of improvement involves dominant intervention clusters with regard to what systems do, there was wide variation in how they did it.

In relation to common interventions no matter what the stage of development six interventions were identified as occurring with equal frequency across all improvement journeys; that is, in schools moving from 'poor to fair', 'fair to good', 'good to great' and 'great to excellent'. These were an examination and revision of curriculum and standards; ensuring an appropriate reward and remuneration structure for staff; assessing students; establishing a data system; publication of policy documents; and implementation of education laws.

In addition to the above, each system implemented a further set of interventions depending on their stage of development.

Systems moving from 'poor' to 'fair'

In the stage where schools moved from 'poor' to 'fair' the interventions focused on providing the basic needs of students; supporting students in achieving basic numeracy and literacy; supporting low-skilled teachers; and lifting the minimum quality threshold of all schools in the system. Ghana, Chile, Minas Gerais (Brazil), Madhya Pradesh (India) and Western Cape (South Africa) were the systems studied in this category. The systems faced a variety of challenges including less experienced and motivated teachers and principals, governing bodies with little capacity for managing schools, widely varied performance across the system, limited resources, and low levels of literacy and numeracy of students, exacerbated by high levels of absenteeism.

While Ghana and Chile focused on improving the systems environment in terms of resources and giving students more time on task, the other three systems implemented programs to raise basic literacy and numeracy, particularly in primary school, across the whole system. Three major interventions enabled this to occur:

1.   Providing scaffolding and motivation for low-skill teachers and principals – The system created instructional objectives, lesson plans and learning materials for teachers so that teachers were executing lessons rather than devising them. Coaches were employed to visit schools and work with teachers in class to help them effectively deliver the curriculum. Instructional time was increased. The systems' central leaders visited schools to observe, meet and motivate staff and discuss performance and the system rewarded schools and teachers who achieved improvement against set targets.

2.   Getting all schools to minimum quality standard – The system set benchmarks for schools/students, mandated frequent learning assessments and implemented data processes to monitor progress. School facilities were improved and resources provided that constituted the minimum threshold for student learning and attendance. Funds were targeted at low-performing schools.

3.   Getting students in seats – The system provided extra places so that more students had access to education and provided students' basic needs (including such things as free school meals, free uniforms and bicycles) to raise attendance.

Systems moving from fair to good

In the second group of schools, those identified as moving from fair to good, the emphasis was on getting the system foundations in place. These systems had completed the stage to achieve basic literacy and numeracy standards and now sought to raise student skills. Their challenge was how to 'configure the foundations of their system, including the creation of systems for data tracking, teacher accountability, finance, organisation and accountability' (Mourshed et al, 2010). At this stage too there was the beginning of the loosening of central control of the system, a phenomenon noted by the McKinsey researchers who observed a 'strong correlation between a school's systems improvement journey stage and the tightness of central control over the individual school's activities and performance' (ibid).

Once again, the research identified three major interventions appropriate to the particular stage of development. These were:

1.   Data and accountability foundation – The system established student assessments and inspections in order to create reliable data, monitor performance and hold schools accountable for student performance. Focus was placed on analysing the data to lift the performance in specific areas where it was falling behind (eg subjects, gender, year levels, indigenous students).

2.   Financial and organisational foundation – The system established a fair funding mechanism and governance processes, including the delineation of decision rights.

3.   Pedagogical foundation – The system established a school learning model which included the number of years students spend at each educational stage, set standards and curriculum, identified resources for each level and made decisions about such things as streaming, tracking and the language of instruction.

In Poland, for example, where prior to 1999 half of Polish students were placed on a vocational track after eight years of primary education, the system leadership increased general education by one year in order to give secondary students a wider range of opportunities. Post 1999, therefore, the school structure became six years of generalist primary, three of generalist lower secondary, and three years of secondary school with academic, general and vocational tracks. At the same time the government's administrative and financial powers were decentralised to districts and municipalities and school principals were given the power to hire teachers. Teachers were also given some power over the curriculum with the government allowing them to choose which curriculum to use from a pre-approved list of over a hundred providers.

An important focus in the fair to good stage of system development in Poland was the introduction of a system-wide assessment system for year 10 students. The rationale for this was that it enabled schools to identify students who were falling behind and needed extra support and it held educators accountable for raising student outcomes. Data from the results of the first assessment in 1998 were also used as the basis for a professional development program for principals in 2001. Evidence suggests that system-wide assessment systems do lift student outcomes in the fair to good stage. In Massachusetts, for example, where the Massachusetts Comprehensive Assessment System (MCAS) was introduced in 1998, the percentage of students passing the state exams in mathematics increased from 33 percent in 1998 to 84 percent in 2008, and in reading from 43 percent to 91 percent.

From 'good to great'

The next stage of the improvement journey, from 'good to great', was identified in the McKinsey report as the move from getting the system and structures in place to professionalisation of educators. In this stage the emphasis is on teacher development and on relying on 'the values and behaviour of (system) educators to propel continuing improvement' (op cit). Systems in this stage employed a cluster of interventions designed to mirror in education the mentorship that occurs in law and medicine.

Three major interventions characterised this stage:

1.   Raising the calibre of entering teachers and principals – The systems raised the entry level for new teachers and lifted the quality of pre-service training and certification requirements.

2.   Raising the calibre of existing teachers and principals – This involved lifting professional development requirements and changing the nature of professional development so there were more opportunities for self-, peer- and centre-led learning and development; providing instructional coaches to assist teachers to improve their skills in lesson planning, student data analysis and pedagogy; and devising career pathways through teacher and leadership specialisations that raised both expectations and pay with each successive pathway rung.

3.   School-based decision making – The systems introduced school self-evaluation, made performance data available and gave schools the flexibility to introduce specialised programs.

Long Beach Union School District (LBUSD) California is cited as an example of a system in the good to great category. Leaders in this system described their aspiration as wanting all educators 'to speak a common language about the craft of teaching, and to have the same calibration of what quality teaching and learning looks like…our litmus is would you put your child in this school?' (Mourshed et al, 2010). The major starting point for the professionalism the system wanted to embed was an ethos based on honest evaluation of objective student performance data, which was made available to parents as well as educators. Data transparency was accompanied by 'walk-throughs' which involved system superintendents walking through schools and classrooms with school principals each month to look at the data and give feedback. On the basis of the outcomes of walk-throughs, the district allocated coaching resources to support struggling schools by providing expert teachers in priority areas such as maths and literacy.

The coaches were generally assigned to four or five schools and worked with three teachers in a school on any given day. The coach first ran a demonstration class, then co-taught a class with the teacher and then observed the teacher taking the class alone. In this phase of the improvement continuum, the aim of coaching was to transmit effective teaching strategies, as opposed to the 'tips and tricks' strategies employed at the poor to fair stage.

An interesting addition to the interventions employed by LBUSD was the connection made at the pre-service stage with the School of Education at California State University.

LBUSD provided teachers from the curriculum department to teach classes on teaching method at the University, recruited 80 percent of its teachers from the University and provided all new teachers with three to five hours of coaching a week in their first year of teaching and seven days of PD training in the second and third years. The aim of the PD program was to offer coherent training that concentrated on good teaching and learning as opposed to one-off 'silo' presentations. The system was also proactive in noting and understanding good practice and collecting and disseminating the good practice it observed in much the same way as occurs in 'evidence-based medicine'. In the five years from 2004–2009 LBUSD improved student performance on the California STAR examination by 20–75 percent in grades two to five.

From 'great to excellent'

Finally, the report looked at those schools which moved from 'great to excellent'. Systems included in this category were Hong Kong, South Korea, Singapore, Saxony and Ontario. At this stage of the improvement journey, the report found educators within the system were highly skilled so that much of what they did was innate. Thus the emphasis moved to creativity and innovation. Once again three major interventions characterised this stage, although it must be remembered that, as with all other stages, the six interventions common to all stages still applied.

In the great to excellent stage, the following interventions were common to the sample schools:

1.   Raising the calibre of entering teachers and principals – The system encouraged school-based professional learning communities, peer-led support and accountability, collaborative practice, and educator autonomy in making decisions about pedagogy.

2.   Creating additional support mechanisms for professionals – The system provided higher levels of administrative and support staff so teachers and principals could focus on teaching and learning.

3.   System sponsored innovation across schools – The system sponsored and identified innovative practices and developed mechanisms to share these practices across schools.

Some of the examples of interventions in the great to excellent stage included the strong focus on action research that applied to the systems. For example, both Hong Kong and South Korea created funds for teacher research and encouraged schools to lead their own research topics. In South Korea teachers were also encouraged to open up their classrooms so that other teachers could visit and observe lessons. Overall, teachers in this stage took responsibility for good teaching and learning, held each other accountable, and set high standards; professional development shifted away from a focus on technical training to collaboration and development.

As the researchers describe:

As the level where system performance is above the global average, there is no longer a single example of a system prescribing scripted instruction in our sample. As system performance rises, accountability also expands, moving from centre-led standarised student assessments to also include school and teacher self-evaluation. In order to achieve teacher improvement in student outcomes, lower-performing systems focus on raising the floor, while higher performing ones focus on opening up the ceiling (ibid).

Further observations

Four other observations can be made before leaving the performance, intervention and contextualising stages of the improvement journeys outlined above. The first of these is that throughout the journey the overwhelming majority of systems addressed the question of appropriate reward and remuneration for principals and teachers, but only after the system had made significant progress in achieving the goals of that stage's intervention cluster. Second, as systems progressed, they shifted from a small number of standardised training programs to a larger number of personalised training sessions that gave teachers much more autonomy in choosing the professional development they needed. Third, the nature of assessment changed so that in the early stages of improvement assessment was focused on achieving specific and narrow targets in literacy and numeracy, while at the later stage the scope of tests tended to expand to cover problem solving and more qualitative aspects.

Finally, the researchers found that while interventions could be categorised into process, structure or resource-based, the vast majority of interventions at all stages were process based. That is, they dealt with delivery of instruction. Only 15 percent of interventions dealt with the content of system instruction (eg standards, curriculum); and 10 percent to resources. 75 percent dealt with those processes that improved teaching and learning practices in the classroom and of these 75 percent, 25 percent related to professional development and 15 percent to accountability.

As outlined at the beginning of this paper, an important element of the McKinsey research was to determine how it was that improving systems were able to sustain improvement. What the researchers found, perhaps unsurprisingly, was that sustained improvement was about changing teacher culture so that not only was there emphasis on how and what teachers teach but on how they think about teaching. This change was supported by what the researchers called collaborative practice, a mediating layer and architecting tomorrow's leaders.

Collaborative practice ensures the routines of instructional and leadership excellence are embedded in practice, classrooms are public, teachers coach each other and a professional career pathway is articulated that not only enables teachers to chart their own professional development but also allows the sharing of skills throughout the system. Collaborative practice pays attention to detail and expects teachers to commit not only to improving their own practice but that of others by sharing what they have learned. An interesting corollary of collaborative practice is that it 'serves as a mechanism of peer accountability, substituting for other formal accountability measures such as teacher appraisals or requalification' (op cit). The researchers found that while teachers were the greatest recipients of support in improved school systems, with 56 percent of all professional development and coaching initiatives aimed at teachers, they were the recipients of only 3 percent of the accountability measures. The reason for this appears to be that teachers are held accountable through the learning of their students (via student assessment) and by their peers through collaborative practices. Direct teacher appraisal represented only 3 percent of teacher accountability measures.

The mediating layer is described in the report as similar to the 'operating system' of a computer. In relation to computers, the operating system performs basic tasks. It is tied closely to both the user interface and the central processer and mediates between the two; similarly, the mediating layer is that space between the system (the central processor) and the school. The mediating layer provides targeted support for schools, shares and integrates improvements across schools and acts as a communications buffer between the school and the system. The report describes the importance of the mediating layer in acting as a buffer between schools and the 'centre' as follows:

The mediating layer can amplify constructive communication by ensuring each school receives and understands guidance from the centre, and the centre hears feedback, requests and ideas from schools. It can also buffer resistance to change, resolving the issues that can be tackled locally and highlighting those that the centre needs to understand and deal with, while filtering out much of the unconstructive noise that always accompanies challenging changes (op cit).

Continuing with the computer analogy, the central processor is what ensures that what happens next time is what happened the last time. In a school system, the continuity of the system's leadership is what ensures what happens next time is what happened last time. Thus, sustained improvement means schools need to traverse smoothly from one leader to the next so that change becomes evolutionary in nature. Systems that successfully sustain improvement pay attention to leadership succession and are always 'architecting' tomorrow's leaders.

Getting started, what the report calls 'ignition', is the final element of the school improvement journey. One of the barriers to school improvement at both the individual school and the system level is it simply seems too hard. Even when the reform process is put in place it often peters out before sustained change occurs because the effort required to implement long-term change is too challenging in systems where educators see themselves with little power and few resources. In the research into the 20 systems identified in this paper, the researchers identified three circumstances that ignited change.

In some systems a political or economic crisis was the catalyst for change. In others a high profile and critical report about school performance was the impetus for large-scale change. The final change agent was the advent of a new political or strategic leader who, because he or she was new to the system, was able to reset relations with critical education stakeholders while possessing plausible deniability for past shortcomings in the system.

In every one of the 20 systems studied a new leader was involved in igniting change and in 15 of the 20 systems the conjunction of two catalysts was present. By far the most important agent for reform was the impact of a new political or strategic leader who had staying power. In the study, the median tenure of the leaders was six years for strategic leaders and seven years for political leaders.

It seems that the appointment of a new leader creates a disjuncture in the system that provides opportunities for determined leaders to implement change. Where this change is successful and 'takes', Mourshed and colleagues found five steps occurred in the improvement 'playbook'.

Five steps to improvement

First, the leaders decide on the non-negotiables. For example, in the Long Beach reforms referred to earlier, the non-negotiables were student achievement standards and professional development; in Slovenia the non-negotiables were adequate resources, curriculum and professional development; in Poland the introduction of 4000 lower secondary schools in one year was the non-negotiable, and so on. These non-negotiables became the anchor points for the reforms and were not compromised even when there was political pressure from outside the system.

Second, the reform leaders installed capable and like-minded people in key positions. Sometimes these people were from outside the system, sometimes they were highly skilled and trusted insiders, but in all cases they believed in the reform and had the skills in implementation.

Third, the reform leaders engaged with stakeholders, including parents, teachers, principals and community leaders in order to manage the direction and pace of change. This involved continuous communication with those affected by the change, allowing stakeholders to air their grievances and concerns and ensuring that positive relationships developed between stakeholders and the reform leaders.

Fourth, the reform leaders secured resources for the non-negotiables. For example, in England where the priority was to raise literacy and numeracy, financial resources from the existing school budgets were reallocated to create a supply of literacy and numeracy coaches; in Hong Kong new funds were injected into the school system by allowing private operators to run schools. Under this system, the government provides 80 percent of capital expenditure and 100 percent of operational expenditure, while sponsoring bodies such as charities and Church missionary societies operate the overwhelming majority of schools.

Finally, successful reforms leaders got 'early wins' on the board quickly. In Long Beach, for example, new leaders, realising they had to gain stakeholder trust quickly to meet their commitment to act against escalating gang violence, introduced compulsory school uniforms so that gang affiliations would not be visible from clothing. This very visible change reassured parents that change was blowing through their system.

Overall, where successful continuous improvement has occurred, leaders have characteristically taken advantage of the 'clean slate' they brought to the process, have followed a common 'playbook' of practices and have brought longevity to their roles.

The McKinsey report does not suggest that continuous, sustained change in school systems is either quick or easy. On the contrary it requires dedicated leadership, commitment from educators and implementation over time of a suite of research-based, proven interventions. It is a journey which, once begun needs to be sustained, and one which is never over. The report does, however, offer reassurance that school systems can go from poor to fair, from fair to good, from good to great and from great to excellent and offers a roadmap to begin the journey.

Recommended Reading

Mourshed, M., Chikioke, C. & Barber, M. (2010), How the World's Most Improved School Systems Keep Getting Better, McKinsey and Company, http://www.mckinsey.com/clientservice/ Social_Sector/our_practices/Education/ Knowledge_Highlights/How%20School%20Systems%20Get%20Better. aspx

Interesting site: http://clients.mediaondemand.net/ MCKINSEY/2010/SCHOOLS/player

In this Webinar presentation, researchers and educators discuss and answer questions at the launch of the McKinsey and Company research paper, How the World's Most Improved School Systems Keep Getting Better, and summarise their conclusions about what they have discovered drives school improvement.



McKinsey & Company (2007), How the World's Best-performing School Systems Come Out on Top, http://mckinseyonsociety.com/how-the-worlds-best-performing-schools-come-out-on-top/

Mourshed, M., Chikioke, C. & Barber, M. (2010), How the World's Most Improved School Systems Keep Getting Better, McKinsey and Company, http://mckinseyonsociety.com/how-the-worlds-most-improved-school-systems-keep-getting-better/


Subject Headings

Education finance
Education policy
Educational planning
Educational evaluation