A primer on the ‘what works’ debate, with key sources and a discussion of its pros and cons.
I recently joined and met with SURE - ‘School and University Research Enquiry’, a research group which has put several schools in the Glasgow area in contact with the University of Strathclyde in order to exchange knowledge and conduct new research. The ultimate aim is to promote a more evidence-informed approach to educational decision making and practice.
With this in mind, I thought it would be helpful to write a brief overview of the field of evidence-based education, including some of the main publications and debates.
What is it?
Firstly, evidence-based education is the idea that research of various kinds should be used to inform decisions about teaching and learning. It is conceived of as an alternative to teaching practice that is guided by intuition and/or experience.
An educator’s job includes a huge amount of decision making. For example, what should be taught today? What about tomorrow? What type of homework should be set, and when? How can a teacher maintain discipline effectively, and engage their pupils? Evidence-based education aims to tackle these questions pragmatically on the basis of past findings, and is sometimes referred to as a 'what works' approach.
Focusing on the example of homework, a traditional view might be that the teacher should allocate whatever they judge to be useful, or whatever is just ‘the way it’s done’ (or whatever is lying around the office, is quick to mark, or is in the textbook/revision guide!). An evidence-based alternative would be to look at this issue from the perspective of research which has shown that some strategies lead to more effective/durable learning than others - the cognitive psychology of memory tells us that learners remember more if there is a delay before they practice material that they have mastered in class, and that they remember more if they do a closed-book test rather than copying from notes. The teacher may therefore decide to set a practice test, and to do so after a one-week delay rather than on the same day as the material was done in class.
The above example relates to memory, and the what works approach as a whole usually refers to techniques or interventions that boost attainment (as measured by some form of test or exam), but evidence could inform many other types of decision too. For example, when considering an issue such as student motivation, evidence could be evaluated to help determine the most effective way to proceed.
As a model, this borrows from the philosophy behind evidence-based medicine. We would probably take it for granted that a doctor should select a treatment that has been shown by reliable (and replicated) research to be the most effective, rather than being guided by tradition (leeches, anyone?) or their individual gut feeling about what ought to work. In the same way, it is argued, teachers should look to the evidence rather than relying on their personal preferences or even on classroom experience. Insisting on evidence may have the incidental advantage of making educational practices less vulnerable to fads, such as the learning styles myth.
Sounds great! So everyone agrees with this…?
No! It has many critics, and their points are well worth taking on board. Firstly, the idea that education can derive a model of effective practice from medicine is open to doubt. Learning is not really like curing an illness - it’s cumulative, has no clearly defined end point, and there are important subtleties such as how well it can be transferred to new situations. The entire approach could therefore be seen as over-simplistic.
Secondly, what works for one group might not work for all. To take one example, Kalyuga (2007) has described the ‘expertise reversal effect’ whereby tasks that are effective with beginners become ineffective or at least inefficient when used with more advanced learners. Another example, much discussed in recent years, is that homework appears to be more effective for secondary students than for primary (Cooper et al, 2006). This is not a killer blow to the idea of evidence-based practice, but it does suggest that the use of evidence must be cautious and thoughtful - we can’t apply one-size-fits-all solutions.
Thirdly, there are concerns about the validity of some of the evidence used. Education is a notoriously tricky area to research - for ethical reasons it is often necessary to rely on correlations and secondary data, leaving some findings open to confounding variables. Meanwhile, a lot of the research evidence from cognitive psychology in areas such as working memory and learning is based on laboratory studies with university students. That doesn’t make it inherently bad research, but does mean that we should be cautious about generalising it to school pupils.
Finally - and linked to the previous point - some people argue that the evidence referred to in this approach is often positivist in its underlying scientific philosophy, whereas many educators and learning researchers subscribe to a social constructivist view of learning.
There is a lot of literature in this field, including both empirical research studies and reviews. For anyone who is new to this area, these are a few very useful publications to get you started. In the main they come from proponents of the idea, but I've also included some key critiques:
Broader than most, the APA’s guide includes such issues as creativity, classroom management, and growth mindset, as well as strategies that impact on learning more directly.
Gert Biesta here criticises evidence-based practice and also questions the broader assumption that closely-controlled lab work has ever contributed much to society (!). He argues that it tends to link to top-down approaches where administrators and governments say that strategies work on the basis of lab research, when they may not work in a specific context. Additionally, the notion of something working doesn’t address philosophical issues of who it works for, and to what social end.
Coe et al (2014)’s report ‘What makes Great Teaching?’ is useful in that it goes beyond the cognitive evidence and considers such issues as classroom climate, teacher knowledge levels, and how teachers can improve. Otherwise, it draws on a similar body of research to Dunlosky et al (2013; see below). The Sutton Trust also back the Education Endowment Foundation’s ‘Teaching and Learning Toolkit’, which provides a useful (if rather undiscriminating) visual guide to evidence-based strategies in terms of cost, lasting impact and the security of the supporting research.
The authors are psychologists and memory researchers, and this paper reviews a number of different findings from cognitive psychology. In particular, it endorses the use of retrieval practice (the ‘testing effect’) and distributed practice (the ‘spacing effect’), while noting that techniques such as re-reading and highlighting are generally ineffective as study strategies.
Australian researcher John Hattie is probably the biggest name in this field; he has synthesised numerous meta-analyses of educational research and built up a list of interventions together with their average statistical effect size. He takes an effect size of 0.4 as a 'hinge point' above which interventions fall into (roughly) the top half, i.e. they are among the more effective interventions - but the higher the effect size, the better. The work is also helpful in identifying some interventions that have tended not to make a large impact. It has its flaws, both conceptual and statistical, but it’s a useful starting place for finding out about several important strategies.
An excellent blog run by four cognitive psychologists who study learning and memory. It is aimed at students and teachers, and makes the science highly accessible without dumbing it down.
It’s useful to be aware of the work of Marzano et al (2001), one of the earlier evidence-based summaries of effective teaching interventions. The strategies they endorse include analogies and metaphors, student-generated study notes, and feedback/formative assessment. There have been important new findings and some of the key research questions have moved on a bit since it came out, however, so it is a bit dated.
The National Center for Education Evaluation and Regional Assistance (NCEE) in the USA offers the ’What Works Clearinghouse’. It usefully reviews studies of efficacy in terms of learning, but the focus tends to be on large-scale programmes, for example the “Great Explorations in Math and Science® (GEMS®) Space Science Sequence” curriculum, rather than on specific techniques that teachers could use in class. This makes their findings less immediately applicable.
In his paper ‘What works may hurt’, Zhao refers back to the analogy of evidence-based medicine and borrows a further concept - that of side effects. From this perspective, an intervention may ‘work’ from a learning point of view, but it could have any number of side effects. Just as with a drug, any benefits must be evaluated in that context. For example, an intervention that boosts learning over the short-term could also harm motivation over the longer term.
Is all of this a threat to teachers?
It is worth considering: does all of this amount to self-proclaimed experts telling us what to do (or what not to do)? At times that might be a valid concern, but the entire nature of making education more evidence based is that that evidence is (or can be) open to scrutiny. You may not agree with all of the conclusions from the sources above, but their arguments are probably backed up by a more thorough factual base than the opinion of a staffroom colleague. And if you are unsure, then you are free to scrutinise and evaluate the sources.
A problem, certainly, lies with teachers’ access to information. If teachers can’t or won’t access the evidence themselves, this puts a lot of power in the hands of central institutions who may try to push inappropriate programmes and interventions. Teachers (and schools more broadly) are in a stronger position to ward this off if they not only learn about the evidence but are also aware of its limitations.
For this to happen, practitioners require journal access, CPD time, and also the skills to critique the research methods and statistics used. How can that be achieved? This BERA report sets out a vision of schools and colleges as "research-rich environments in which to work" (p.5). It's a radical idea, and one that asks us to reconsider the very nature of what teacher professionalism involves.
Biesta, G. (2007). Why “what works’’ won’t work: Evidence-based practice and the democratic deficit in educational research. Educational Theory, 57(1), 1–22.
Coe, R., Aloisi, C., Higgins, S. and Major, L.E. (2014). What makes great teaching? Review of the underpinning research. Accessed 14 May 2017 at http://www.suttontrust.com/wp-content/uploads/2014/10/What-makes-great-teaching-FINAL-4.11.14.pdf
Cooper, H., Robinson, J. C., & Patall, E. A. (2006). Does homework improve academic achievement? A synthesis of research, 1987–2003. Review of Educational Research, 76(1), 1–62.
Hattie, J. (2013). Visible learning: A Synthesis of Over 800 Meta-analyses Relating to Achievement. London: Routledge.
Kalyuga, S. (2007). Expertise reversal effect and its implications for learner-tailored instruction. Educ Psychol Rev, 19, 509–539. doi: 10.1007/s10648-007-9054-3
Marzano, R. J., Pickering, D., & Pollock, J. E. (2001). Classroom Instruction That Works: Research-Based Strategies for Increasing Student Achievement. Alexandra, V.A.: ASCD.
Zhao, Y. (2017). What works may hurt: Side effects in education. Journal of Educational Change, 18(1), 1-19.