(Half-written blogs haunt my google drive so I’m copying Toby Payne-Cook’s great housekeeping idea and publishing my 12 blog posts of Christmas. This is post 1 of 12 – I hope someone gets something out of reading these drafty ideas, but even if they don’t, I’ll feel better for being able to delete the drafts.)
Have you ever wondered why we let so many diverse educational approaches flourish in the English education system? Knowledge-rich approaches led by highly autonomous teachers… knowledge-rich direct instruction… collaborative and enquiry-led learning… integrated curriculum approaches… the list is infinitely long.
Surely we should just develop a mega-research experiment and measure the effectiveness of approaches against each other, rather like the famous American Project Follow Through? The Winner Takes It All.
In Chapter 2 of our new book, The Next Big Thing in School Improvement, we describe why inference from these types of randomised controlled trials is rather difficult. After finishing the book, I heard about the research of complexity scientists, such as Kristian Lindgren, who have studied optimal game-theoretic strategies within a complex system. I now think that inference is not only difficult; it might also be misleading if our goal is to create the best education system we can.
Most people know about the Prisoners’ Dilemma – the famous game theory problem about whether to betray your partner in crime. Many also know that in a repeated game (as life generally is), tit-for-tat (co-operating only if the other player did in their last move; punishing only immediately after another player has punished you) is generally a highly effective strategy. Kristian Lindgren and other complexity scientists working in the field asked the question: What do effective strategies look like if they are not fixed, but rather if they are able to evolve over a period of time.
Roughly speaking, these scientists developed 100 different game strategies and had computers play pairs of strategies against each other in repeated games with the aim of seeing what sort of strategies won out. If a strategy does well, winning a lot of games, it replicates. If it consistently does badly then it dies and is removed from the game. However, strategies are also able to mutate so that they can take on characteristics of other strategies that are also doing well. And strategies are also able to alter their perspectives or goals at random, with some adopting longer-term perspectives on how to respond to the behaviour of their opponent (in contrast to tit-for-tat which solely looks at the last move). Strategies look like they are able to get smarter and learn, even without human involvement. There is no end game because the system forms an ecology, constantly deepening perspectives, acquiring more smartness, with strategies learning from each other.
Now, imagine these game strategies were instead educationalists who come armed to the classroom with particular instructional approaches. Who wins and develops the most effective strategy? Well, at any point we could stop time, measure effectiveness and declare: ‘This is the best teaching strategy; a solution has been reached.’ However, the insight of the complexity scientists is that if the system is kept running, with teachers adopting new strategies that look promising, borrowing ideas from other strategies, developing mutations, chucking obviously unproductive ideas in the bin, then there is no end game. It might often look like the system has settled on the ‘best’ solution for years, but then every so often the system will go wild as innovation takes place and other strategy starts to dominate.
Not everything is destined to achieve success at some point, or under particular circumstances or iterations of the game. I can promise you that open-plan schools die in every version of the educational game!
Teachers and educational companies are agents in our complex, evolving system. They make moves in ill-defined systems. They try things out. They cook up different beliefs and theories about what is going on. They explore strategies and means of prediction or forecasting. They can be parasitic, only learning how to adapt their own instruction to make it effective if they are given the opportunity to encounter particular other strategies. They stretch and shrink the time horizons and goals. Their ideas are fluid.
In education, we are implored to avoid ‘lethal mutations’. To adhere to the fidelity of the programme as we implement it in our classroom. But mutations are the life-blood of complex, poorly understood, ecological systems. Mutations are the way we constantly search for winning strategies in the infinite game of improving education.
(Since I wrote this blog, I remembered about this excellent post on mutations…)