What next for Ofsted?

Ofsted isn't going to be abolished (nor should it be)

A few weeks ago Old Andrew wrote a typically incisive, and popular, post explaining why he felt Ofsted are now beyond redemption and should be abolished.

That isn't going to happen. Ofsted is absolutely essential to the regulatory model the coalition Government have constructed. Failing academies and free schools can only be shut down on the basis of a poor inspection - there's no other legal mechanism (unless they mess up their finances or do something illegal). Likewise the Teaching School / National Leader of Education processes rely on Ofsted identifying outstanding schools.

And as Andrew notes in this post there's no indication that Labour would behave differently. If anything they envisage a bigger role - in recent weeks Tristram Hunt has argued for separate inspection of academy chains.

But even if the politicians were arguing in favour of abolition - and some way was found to deal with all the structural problems that would create - I'd be opposed. One of the OECD's clearest findings from PISA is that the worst type of education systems are those with high autonomy and low accountability (conversely the best are high autonomy and high accountability). While it may not feel like it to classroom teachers we have one of the highest autonomy systems in the world. Most secondary schools now have control of their own curriculum; nearly all funding in the system is passed directly to schools; teacher training is increasingly run through schools too.

So while I think there are lots of ways we can make our accountability system smarter I don't think we should be removing a key mechanism for holding schools to account. If we did then we'd be left with exam data as the only basis for accountability. Yet we know that schools can game test-based accountability by choosing easier options; focusing on certain pupils; or simply spending a lot of time cramming. While I've never been to a great school getting bad exam results; I have been to bad schools that get decent results. We have to have a way to check the quality of education on offer that goes beyond gameable data. There's a reason that most countries have some form of school inspection (and many of those that didn't have introduced inspectorates over the past decade).

That doesn't mean there isn't a problem

Although I think Ofsted needs to be an important component of our regulatory system I recognise the validity of many of the criticisms (and wrote about them here). This is not just the standard whinging you'd expect from any profession about its regulator.

There are two separate but related issues.

First a percentage of inspectors are not following the framework. They are insisting on seeing certain types of teaching in lesson inspections. It's unclear how large this percentage is but there are enough "rogue" inspectors to leave school leaders uncertain about how to prepare for inspections. That means many schools are doing things that are not required by the framework - and go against the spirit of that framework's intentions - because they don't know what "type" of inspector they are going to get.

Secondly lesson observations have limited validity - even if being done properly. Professor Rob Coe gives the technical details in these slides. But it's actually pretty obvious that in most cases it won't be possible to accurately tell how good a teacher is from a single 20 minute observation - even if it's done well. It's also hard not to let prejudices slip into judgements.

Katie Ashford explained here how teachers should approach an inspection of their lesson - and a good inspector like Mary Myatt will respond well to that. But it's understandable that many feel this is a risky strategy and end up teaching to what they think Ofsted wants to see rather than what's in the best interest of their class.

So what needs to change?

The first, fairly obvious thing, is to stop grading individual lessons. It's not clear at all what the benefit of doing this is given the lack of validity.  It would be possible to build greater validity around observations but would require multiple observations by people rigorously trained in the same scoring system. And that's too expensive and complex for Ofsted to introduce.

Scrapping individual lesson judgements would significantly reduce the problems Ofsted causes. Teachers would feel under less pressure to deliver a specific type of "outstanding" lesson; they wouldn't feel unfairly judged when their lesson gets a grade 3 or 4; and schools wouldn't feel so bound to using Ofsted grades in their own lesson inspections (and could move away from scored observations all together).

It wouldn't reduce the validity of the overall inspection as the lesson judgements aren't particularly valid anyway.

Instead inspections should focus more on systems. Essentially Ofsted should be looking at what the school is doing to ensure consistent good teaching. They should be inspecting the school's quality assurance not trying to do the quality assurance themselves in the space of two days (good inspectors will see this as their role already - but it's nowhere near explicit enough).

In their observations of lessons they should be checking the leadership know their teachers and understand how best to support their future development. They should be checking that they have thought about professional development and about performance management (which shouldn't have to involve performance-related pay). They should be seeing if the behaviour policy is being enforced; and if the school curriculum is actually being used. This blog does a good job of showing how that might work in practice.

I'm not sure there even needs to be an overall grade for teaching. After all if the leadership's good (because there's consistent practice across the school) and the attainment is good then the teaching will, almost by default, be good too. Likewise if the leadership is poor then the school isn't good even if there are pockets of excellent teaching (as there nearly always are in bad schools).

In addition to this simple, if bold, change Ofsted should make use of their new regional structure to announce a full reset so that everyone in the system - inspectors and schools - have a shared understanding of the purpose of inspection. This could involve, for example, retraining all inspectors alongside representatives from local schools; licensing inspectors through published exams; banning lead inspectors from acting as consultants and, as Rob Coe suggests, introducing transparent and independent processes for quality assuring inspections. This would need to be communicated via a major campaign from Ofsted, DfE, and, ideally, Head's associations.

Perhaps this wouldn't be enough to restore everyone's confidence in inspection but I'd much rather try to improve what we have than push the system into total reliance on test data.