Centre for High Performance

School Improvement for those who don’t believe in School Improvement?

Research produced by the Centre for High Performance has grabbed some startling headlines recently: “unruly pupils ‘excluded by failing academies to boost standards’” and “Superheads – the true cost to schools”.

The actual report How to Turn Around a Failing School is more nuanced than these headlines, and Alex Hill’s talk in Dubai provides further clarity, but even the full report contains some pretty blunt findings on what works with school improvement.  Take step 4 of their 8-step ladder: “Student quality – exclude poor quality students, improve admissions and acquire a local primary school.”  This 8-step ladder isn’t buried in an appendix – it leaps out of page 1.

The report’s authors – Alex Hill and Ben Laker – are keen to point out that they are reporting on what has happened, not what should happen.  They studied 160 academies placed in special measures, then tracked progress over 7 years and plotted this progress against the actions taken by school leaders throughout the improvement process.

I’ve described to colleagues the model proposed in the report as ‘school improvement for people who don’t believe in school improvement’.  It reveals the effectiveness of supply-side reforms (‘better’ students, teachers, leaders and governors) rather than securing better outcomes with the same personnel.  It’s no surprise that one of the authors comes from an engineering background; his previous work includes titles such as “How do you stabilise your supply chains?” and “Essential Operations Management”.

Alex and Ben were kind enough to meet myself and a few colleagues this week.  Here’s what I took from our discussion:

  1. Changes to personnel (students, parents, teachers, governors) are easier in urban areas than rural and coastal areas due to the simple fact that urban schools can draw on a larger pool of people in close proximity to the school. This means that isolated schools simply don’t have access to some of the key drivers of school improvement mentioned in the report.
  2. Genuine school improvement takes time and money. If we want quick wins, canny heads can achieve them, but proper school improvement takes time.  The announcement that Ofsted will give new leadership teams of challenging schools an ‘improvement period’ in which they are safe from inspection is welcome.   The report also makes it clear that purse strings may need to be loosened until standards improve.
  3. Schools can’t do everything at once, but should start with the basics such as leadership, attendance and behaviour. It’s fashionable to talk about marginal gains in education, but failing schools need to do more than tinker at the edges (more on this HERE).
  4. We know very little about school improvement. As the authors point out in their criticism of super-heads, school improvement is often based on the hope that individuals with a track record elsewhere can work similar miracles in a new context.  This fixation with personnel suggests that our knowledge of school failure, and the support needed to revive failing schools, is limited.  I hope that reports such as this one might fuel dialogue which deepens our understanding of real school improvement, and I hope that the introduction of new performance measures this year (Progress 8 etc) might enable us to peak under the bonnet of transformed schools to see if the reality matches the rhetoric.

I think there are flaws in this report.  The authors come from a business background, and the outcomes of this report are jarring to those of us who know that schools can achieve better outcomes with the same students.  We know that it’s easier to get better outcomes with kids from richer families, but we go into education because we believe that we can make a difference in the lives of our students whatever their starting points.  I wish the authors had spent time researching schools that did make genuine gains with the same students and teachers.

The Centre for High Performance has studied success in a range of contexts and seeks to apply this to education, but all of the organisations mentioned (New Zealand’s All Blacks, British Boxing, NASA, Royal Academy of Music, Royal Marines, Royal College of Arts and Royal Shakespeare Company) can select elite performers from a wide pool of talent.  If we want to learn how to transform challenging schools we should focus not on selective organisations but on those which have delivered sustained improvement with the same resources as they had when they were failing.

This report has shone a light on the stark choices that schools face if they want to rapidly improve student outcomes.  It is deliberately devoid of value-judgements and social conscience.  I am sure I won’t be alone in seeking to fuse the findings of this report with the moral purpose that underpins our profession.


Studying Sweden

This paper follows a visit to Sweden in May 2015. Dame Sally Coates and I were invited to speak at a school leaders’ conference which assembled leaders from business, education and politics to discuss the state of education in Sweden.

Challenges in Sweden’s Schools.

There is so much to admire about Sweden: the innovative businesses (Ikea, Spotify, H&M, Skype); the generous maternity and paternity pay (16 months shared between father and mother at 80% of salary); subsidised childcare; the relatively small gap between its richest and poorest citizens.  Sweden has the highest proportion of working mothers of any developed nation, and there’s no other rich nation which accepts more asylum seekers per capita.  Only Luxembourg has a more generous aid budget.  It’s no surprise that Sweden is regularly ranked as one of the most liveable countries in the world.

Sweden has the ideal conditions for a first-class education system.  A highly skilled population, impressive levels of civic engagement and faith in state institutions, combined with low levels of poverty, make fertile ground for a successful education system.  For many of us in England, our first awareness of Swedish education was when it was lauded as a model of parental choice, autonomy and innovation – the inspiration behind the Conservative’s free schools programme.   Yet at the moment it seems that Sweden’s success as a country comes in spite of its education system, rather than because of it.

Following a fall in the Pisa rankings in 2013, a team from the OECD were invited to analyse its education system in 2015.  The OECD report confirms that “Student performance on the Programme for International Student Assessment (Pisa) has declined dramatically, from near the OECD average in 2000 to significantly below average in 2012.  No other country participating in Pisa saw a steeper decline than Sweden over that period.”

The report identifies a range of issues in the Swedish system, most of them captured in this paragraph: “[the] Swedish school system faces a number of challenges. These include the low and decreasing performance of Swedish students, with large numbers of low performers and few high performers in all PISA domains. Learning environments are not always conducive to learning and can be insufficiently challenging, with high student truancy and lack of perseverance in learning. Conditions are not adequate to nurture an excellent teaching profession. Teaching is considered a low-status and relatively unattractive profession, due to heavy workloads, relatively low salaries for experienced teachers and limited opportunities for appraisal, feedback and professional development. School leaders and their employers (municipalities and independent schools) do not accord sufficient priority to pedagogical leadership. In addition, their heavy workload and the unclear relationships and distrust between principals and their employers contribute to high turnover.”

Since my trip to Sweden I’ve reflected on what can be done to address the decline in standards.  Here I will explore three areas in which improvements can be made: behaviour, leadership and assessment.

Behaviour: Democracy and Discipline

When Sally Coates and I addressed a group of Swedish principals we told them about the rigorous behavioural expectations that we’ve become familiar with in London academies, such as same-day detentions for minor infringements and a rigid school uniform policy.  Our audience was unconvinced.  A question came from the floor “how do you reconcile this discipline you talk about with democratic principles?”

After reading the full OECD report, I think this question captures part of the challenge that Sweden is facing.   Take these stats from the report:

  • Sweden had the highest proportion of students who arrived late for school among OECD countries; more than one in two 15-year-olds reported that they had arrived late for school at least once in the two weeks prior to the PISA test (Figure 1.10), and this likelihood was higher among immigrant students than non-immigrant students.
  • Although Swedish students are only slightly more likely to skip classes than on average across OECD countries (20% had skipped classes in the two weeks prior to the PISA test), students from socio-economically disadvantaged backgrounds (the most disadvantaged quarter of students) are significantly more likely to skip classes (26%) than disadvantaged students in other countries (19%).
  • 34% of Swedish students reported that the teacher had to wait a long time for students to quiet down, compared to 28% of students on average across OECD countries.
  • 38% of Swedish students reported that there was noise and disorder during lessons, compared to the OECD average of 32%. Such disturbances or distractions during lessons equate to less time for quality teaching and learning. In Sweden, as in most other countries, schools whose student population is predominantly disadvantaged tend to have a more negative disciplinary climate.

One thing we know about learning is that it requires students to be attentive.  The tardiness, truancy and low-level disruption described above suggest that Swedish classrooms often lack the conditions required for students to focus on their learning.

Swedes don’t need outsiders to tell them to sharpen up on behaviour, but surely an honest discussion about the costs of their rather relaxed attitude to discipline would be helpful?  As much as democracy is founded on the rights of individuals, it also relies on civic responsibility: we serve on juries and pay our taxes, or face reasonable consequences.

JS Mill’s ‘harm principle’ provides a clear limit to an individual’s liberty and justifies intervention when one person’s actions harm others.  Classroom disruption and tardiness would appear to pass this test, and provides justification for schools to intervene when the actions of a few students impinge the learning of the majority.

It’s fair enough that Swedes are turned off by school uniforms and silent corridors, but Sweden must surely reaffirm the sanctity of the classroom as a place of orderly calm so that students can maintain the attentiveness required for learning to take place.

In an individual school, I firmly believe that until behaviour is sorted, there’s not much point focusing on anything else.  I would say the same for the whole education system in Sweden.

p.s. see this brilliant article by Tom Bennett for more on the lack of sanctions in Sweden’s school system

Leadership: Autonomy and Alignment: “The stronger alignment we have, the more autonomy we can afford to grant.” 

One of the school leaders we met in Stockholm suggested that some long-serving teachers get paid more than their principals.  It’s a point supported in the OECD report: “In 2013, the average salary for a principal was approximately 40 000 SEK per month [£3,200], the lowest among the Nordic countries”.

Raising the pay of headteachers would do little in itself to raise standards in Swedish schools, but now could be an opportune moment to re-assess the role of principals in the Swedish system, and ensure that the pay awarded to principals reflects the critical role that they play in school improvement.  Successful Swedish businesses could support a school leadership academy which would be open to new and existing school leaders.  These new leaders would be well paid and would be supported by leadership coaching and mentoring.  In exchange, they would be placed in some of Sweden’s most challenging schools.  Low pay for school leaders can be justified when standards are fine, but not when change is needed.

A national school leadership academy would solve a second issue raised in the OECD report: “the heavy workloads, unclear relationships and distrust between principals and their employers have contributed to high turnover”.  These new school leaders would have a direct mandate to drive improvements in student outcomes, cutting through the fragmentation of the current system, where local municipalities enjoy significant autonomy over schools in their region.

In an article in The Guardian Andreas Schleicher, the OECD’s director for education and skills, was asked what lessons the UK might learn from the Swedish experience. His response: “When establishing school autonomy and choice, the Swedish system did not strengthen oversight and intervention functions (which were traditionally weak in Sweden). So the system had no means to learn from successful schools and scale their success, and it had no means to detect and address educational underperformance.”

This problem of autonomy without oversight reminded me of an article on one of Sweden’s success stories, the music-sharing site Spotify.  Spotify typifies the innovative, quirky, high-tech, underdog, efficiency that I associate with the Swedes, and this line from Spotify’s Henrik Kniberg suggests how the firm strikes a balance between freedom and compliance: “We’re trying hard to be aligned and autonomous and we keep experimenting with ways of doing that.  The stronger alignment we have, the more autonomy we can afford to grant.”  The ideal state that Kniberg and his colleagues strive towards is the top-right quadrant of the grid below (diagram and quote from this excellent article)


autonomy alignment.jpg

There appears to be no lack of autonomy in Swedish schools, but perhaps it’s now time to work on the alignment.  The Spotify model appears to involve a shared understanding of the big priorities (the alignment bit), with a large degree of autonomy about how to meet these.  In the context of schools, it might not take too much to align the current providers towards a common set of priorities.  


Once a balance between autonomy and alignment is established, it will need to be clear which schools are performing well, and which are not.  From reading the OECD report, and from my visit to Sweden, I get the impression that student performance data is currently pretty murky.  The OECD report suggests that a reliance on teacher assessment, over standardised tests, might be behind this: “An area of particular concern noted in the 2011 OECD report – one that remains a primary concern in our view – is the equivalence of student grades (i.e. reliability) across schools. Current reporting of student outcomes in Year 9, at the end of compulsory school, heavily relies on reliability of the grades awarded by teachers. Various sources have pointed towards uneven scoring of students’ assessments, as the weights of the test results in students’ grades are determined locally.”

The Ministry of Education in Stockholm would appear to agree, and in recent years has introduced national tests in Years 3, 6 and 9 and a new grading scale.  Yet the OECD report suggests that this new approach to testing is yet to bear fruit: “Sweden’s evaluation and assessment arrangements are underdeveloped, with unreliable national student achievement data, and hence cannot adequately support accountability and improvement.”

Now that the Swedish education system includes large numbers of immigrants from nations such as Syria, Afghanistan, Somalia and Iraq, it is more important than ever that the progress of students can be reliably tracked.  These students are currently lagging behind their Swedish peers: “Students with immigrant background scored 58 points less in mathematics performance in PISA 2012. This is among the largest differences in OECD countries (average difference of 34 points)”.  The yawning gap in performance between immigrant and non-immigrant students provides a clear case for an assessment system which provides accurate and transparent data about student performance.

Accurate assessment will allow overseers to bridge that gap between autonomy and alignment mentioned above.  Students who are failing to make expected progress, particularly new arrivals with limited Swedish, need accurate assessment data more than anyone so that schools can quickly intervene with additional support when it’s needed.

It’s not very trendy, and not very Swedish, to suggest a more robust assessment framework, but it’s difficult to see standards improving quickly unless these standards can be accurately tracked.  I’m sure that the brains behind Spotify and Skype could think of some clever ways of devising challenging online tests which could provide instant formative feedback for schools and students, while also holding schools and teachers to account.

How might this new assessment framework affect teachers?  There is some evidence from the OECD report that it’s currently difficult to identify the effectiveness of teachers, or to do anything to expose excellence or inadequacy: “there are still no national standards for teachers and school leaders regulated by law and no formal procedures exist, and teacher appraisal has not been a central topic in the Swedish school policy debate” (MoER, 2015).  This might help explain the fact that “only 51% of students report that their teachers checks their exercise books regularly” (OECD average 72%).

Higher levels of accountability might well nudge teachers towards marking their books more frequently, but I’m sure that the vast majority of Swedish teachers have the very best of intentions for their students, and that many of them would welcome an assessment framework that would reveal the fruits of their labour in an honest, open and transparent manner.

Skype has its issues, but at least after a crackly call we get the chance to provide instant feedback via its ‘call quality feedback’ tool.  If only Sweden’s teachers and school leaders were provided with similarly clear performance data.


My slightly pessimistic view of education is that most schools don’t make much of a difference.  Children who are dealt a decent hand in life’s lottery (educated, loving parents; stable home life; lots of verbal stimulation in early years etc) tend to do well at school, while students with a sub-optimal home-life tend to do badly.  The best schools are those that actively transform the trajectory of their students, and the best school systems are those that do this at scale.

Perhaps the biggest change in the Swedish school system in recent years is that it now contains more children who require transformative education, rather than laissez faire education.  I take the view that transformative education requires a particular approach to schooling: one based on impeccable behaviour, robust leadership and transparent assessment.