I read a great piece by Dr Becky Allen in Schools Week this morning on inherent unreliability of school inspections. In it she makes the point that human beings are incapable of making reliable, high stakes judgements due to our adaptive reliance on heuristics and our inability to adequately introspect about our biases  and preferences. But despite the dangers, she says, “This is not to say that school inspection should not have a role in our system. It is possible that the threat of inspection, day-in-day-out, leads to better practice in schools that outweighs the obvious dysfunctional behaviours it creates.”

I think this is right. Human beings need to feel they are accountable if they are to be their best. In some cases this is straight forward: I am accountable to the readers of this blog. If I started turning out low quality posts, I would, I’m sure, rapidly lose readers. If I started churning out advertorials or paid for posts by third parties I would quickly lose all credibility. With schools, accountability is trickier to manage.

The first step might be to decide what school inspection is supposed to achieve. Are we trying to identify schools where students are routinely failed? Are we trying to help schools improve? Are we trying to reward schools who are going above and beyond minimum requirements? Are we checking to see how public money is being spent? Are we trying to identify and spread ‘best practice’? Are we trying to produce a ‘good schools guide’ for parents? Part of the problem is that at the moment we are trying to do all of these things and so, perhaps, we run the risk of doing none of them well. It’s ridiculously easy to create perverse incentives that encourage schools to look good rather than genuinely focus on something more substantial like making every single student cleverer.

I would suggest that we pick a single purpose and try to get good at doing that. The first thing to remember is that while we might value all sorts of things, few of the things we value can be measured with any degree of reliability. Flawed as it is, exam performance provides the best bet for establishing how well students are doing and exam performance can only hope to measure a narrow range of skills. This means we have to give up on holding schools to account on measures like ‘character’ or ‘creativity’. There’s a lot wrong with the way Progress 8, for instance, is calculated but that doesn’t mean we cannot correct some of the grosser errors whereby progress is averaged and establish a metric we all agree is good enough.

The next step would be to remove the pseudo certainty of grades. If inspectors believe they can reliably determine where a school fits on a 4 point scale in a two-day inspection then they are deluded. Even moving to a two-point scale of ‘good enough’ and ‘not good enough’ carries too great a risk of being wrongly applied to be adopted. All that can be done is report what may be genuine concerns and for these to be monitored and not published. As soon as inspector’s flawed and biased judgements are published they become true. Instead, schools should be put on notice that a second, more detailed inspection will be undertaken after, say, a year to put things right. The risk of allowing schools to founder for an additional year is vastly outweighed by the certainty of publishing wrong judgements which ruin careers and blight children’s life chances.

For accountability to be effective, two broad principles need to be observed. Firstly, the views of the observer must not be known in advance, and secondly, those being held to account must trust that the observer is well-informed and interested in accuracy. Only then will we create a climate in which schools are free to focus on doing what is most likely to help children get cleverer instead of trying to find ways to game the system because they know inspectors are looking for the wrong things.

Over to you.