Measuring inclusion? Beware the unintended consequences

As part of a wider ‘DEEP dive’ into the effectiveness of schools groups, the Education Policy Institute has this week published a research paper on how to measure inclusion in different school groups, accompanied by a consultation open until March 14. I know I speak for many when I say it is very welcome. Many schools work incredibly hard to ensure they are reflective and inclusive, but some simply do not share their moral purpose. Accountability is an important lever to recognise the former and incentivise the latter.

In some areas, it is an open secret: Children’s movements from one school to another are not the result of positive choice but of children and their parents feeling they would be more welcome elsewhere. The groundhog day scenario of phone calls home for poor behaviour. The quiet conversation suggesting another school will better cater for your child’s needs. These moments are never captured in exclusion data. They only ever see the light of day because of a proactive local authority admissions team or vocal groups of parents.

So the EPI consultation is an important one, but it also raises concerns. For a start, what do we mean by ‘inclusion’? Ask 20 people to define it and you are likely to get 20 different answers, some contradictory. The EPI paper offers a broad definition: an intake that is reflective of the community and support for all pupils to achieve their best in an appropriate, quality education once they are admitted.

That ‘once they are admitted’ is already problematic, because we know some schools put parents off before admission. The problem is attempting to measure inclusion through quantitative metrics. The EPI paper does note that other qualitative data can help to build a more holistic picture, but their own focus narrows down to two things: attendance and exclusions, and pupil achievement.

The next problem is that pupil achievement is a blunt measure of access to the curriculum, and we simply can’t talk about inclusion without that. Sitting in a classroom but only interacting with a teaching assistant is not inclusion. Spending most of your time working outside the head’s office isn’t either. Good achievement can mask many an exclusionary practice.

Inclusion and high attendance are also not synonymous. Attendance is important, but we must be wary of the unintended consequences of using it as some sort of gold standard. While high attendance may be the sign of a happy, motivated school community, we must also consider that fear of castigation and school attendance orders forces some parents to put themselves in sometimes dangerous situations to get their children into school. Is asking children to be brought into school in their pyjamas to get their attendance mark inclusion? And what about the child with significant medical needs who has a number of absences for treatment? Is non-attendance really a marker of the school’s failure to include?

And finally, exclusions data is a poor proxy measure of inclusion. In terms of accountability, what might be more helpful is looking at data about what happens after an exclusion. How many children go on to be assessed for an education, health and care plan? How long were they in school prior to exclusion without one? How many exclusions are successfully appealed at governors’ discipline committees and independent review panels?  

In the end, if what we want is for more schools to adhere to a moral code on inclusion, then quantitative measurements are only likely to create more perverse incentives. In the meantime, exclusions are required to sit within the remit of the law (proportionate and fair), and there is definitely scope to improve on that.

After all, this is a deep dive about school groups, and we still have a situation whereby children can be moved within trusts without any choice about where they are going. Those children aren’t excluded. No pupil exit is recorded for them. And they have no recourse to complain because they can perfectly legally be directed off-site to improve behaviour.

We could possibly measure that, but are we sure it tells us anything about inclusion?