James Croft is Executive Director of The Centre for the Study of Market Reform of Education and author of Collaborative overreach: why school collaboration probably isn’t key to the next phase of school reform

Recently, a spate of policy papers have appeared from think tanks and research organisations across the spectrum suggesting that collaboration between schools to improve performance is likely to be key to the next stage of school reform. It’s not difficult to see why. As politicians, policymakers, and schools try to figure out how to get more for less, propositions involving ‘sharing’, or resource re-distribution, seem worth exploring.

The problem is that we’ve little to go on, either in terms of theory or evidence, to suggest this strategy makes any difference for pupil outcomes. So the return on investment isn’t at all clear.

The overwhelming majority of studies are qualitative. These essentially reflect the pre-existing consensus view of collaboration and what features are important to support school improvement. Given that much of the theory on this subject is a massive distraction from what schools’ primary purpose is or ought to be however, this consensus view isn’t worth a great deal.

This type of analysis focuses entirely on ‘outstanding’ schools that collaborate for professional development and teacher support. Collaboration is a good in and of itself, irrespective of the educational merit of the specific practices disseminated – because that’s not really the point. Only a handful of studies escape this charge.

The lack of interest from the research community in quantifying the effects of collaboration for pupils means that there’s been little engagement either with the challenges associated with accounting for the relative influence of multiple simultaneous initiatives that may be underway in a given schools context.

A plethora of qualitative studies have, for example, set out to show the positive effects of the ‘London Challenge’ collaborative school improvement programme launched by the Labour government in 2003. There were many elements of this programme and we don’t know anything about their relative influence – all the more so since these studies do not account for other factors at work, such as cultural, economic, and other factors.

Recent quantitative research has found that the improvement in London Secondary school results wasn’t anything to do with what schools were doing, but was instead because of a more favourable ethnic composition, likely to be related to immigration.

Simply put, schools benefited from an influx of motivated and hard-working Asian students. It wasn’t anything to do with policy. This is confirmed by another recent study supplying statistics indicating the beginnings of an improvement trajectory across both primary and secondary schools as far back as the mid-1990s, well ahead of the launch of the London Challenge programme.

There’s no shortage of research reports that will try to tell you otherwise, but these are essentially fishing about for plausible explanations that researchers are not in a position to substantiate. London is an ‘interesting city’ for educationists, because it ‘seem[s] to have a promising story to tell about policy leading to improved quality outcomes for schools’ (as one report put it recently), but establishing the links and drawing applications for policymaking is much harder than it suggests.

So if there’s no evidence for positive effects for pupils from collaboration, what about the alternatives – federation for example? Federation is something different from the less binding, more informal arrangements that characterise school collaboration.

Recent quantitative research in this area, though also unable to draw causal inferences, suggests that those types of federation most expressly purposed to improving pupil attainment, and which have organised themselves to deliver, are likely to be most impactful. In the chain, and more specifically, ‘hard-federation’ context, where individual schools are merged and integrated into a single structure, there are resources to manage variable school performance within the group. This research suggests that school leaders ought to give more serious consideration to the merits of joining Multi-Academy Trusts.

But it isn’t clear cut enough to support government intervention to drive such consolidation (and at the same time stifle competing propositions). Studies of chain and federation effects have set much store by what are called ‘matching methods’ which seek to compare similar schools based on observable features of the way they are set up, resourced and composed, but these cannot speak to unobservable factors such as prior capacity for improvement or quality of leadership. This makes their findings of limited use for those set on one-size-fits-all solutions.

Rather, in that the decision to collaborate, or indeed to federate, is predicated on having the autonomy to do so, the best way for the government to help schools work together effectively is via extending their freedoms, and working on the incentive structure to sharpen the focus on attainment and ensure optimal operational efficiency.

Under more competitive market conditions, schools would be likely to move expeditiously in the direction of merger of their own volition, without need for the instruments of the current academisation process. Theory suggests this would give rise to more focused, efficient and effective collaborations also – obviating the need to foster it for its own sake.