These are the slides from my talk at ResearchEd 2014.
The aim of the talk is to look at four different kinds of research and to consider the extent to which teachers might accept the findings and then allow them to influence their practice.
I’ve chosen four contrasting forms of research.
1. John Hattie’s meta-analysis of research into homework. I’ve written about the detail in this blog post. Here 160+ studies are compiled to generate a relative effect size but, if you engage with the detail, there is actually no neat conclusion. The effect depends on numerous variables; to make simple statements about homework in general is lazy.
John Hattie made the following comment on the blog:
2. Robert Bjork’s research into memory is fascinating but what kind of evidence does he have? Many of his ideas derive from experiments where people (often university students) are engaged in controlled trials where they are asked to learn and recall material in various formats. This paper ( Birnbaum_Kornell_EBjork_RBjork_inpress) sets out one example where information about birds and butterflies is presented in a blocked sequence and then in an interleaved sequence and the subjects’ capacity to use that information at a later time is assessed.
Here the findings are analysed for statistical significance and give grounds for suggesting that, on average, people learn more effectively when material is either interleaved or spaced, even if they perceive that they’ve learned better through learning material in a block. It provides evidence that the human brain functions in a particular way. Even though the sample sizes seem small – 100 or so and sometimes less – the experiments are repeated with similar results. There are grounds for considering the results as indicating some truths about how we learn. Should we take this on board in our pedagogy and planning? It seems sensible to look at interleaving and spacing in course planning if there are clear advantages in terms of longer term memory. However, are there issues around the transferability of these findings from learning specific sets of discrete information (as in these studies) to more complex synoptic learning tasks such as those students encounter in many curriculum areas.?
3. The third example is taken from a book by Mary James et al about Learning to Learning ( 2007). There is a whole section dedicated to the research evidence. In one study reported by Bethan Marshall (Kings) et al, 37 teachers were interviewed of whom 27 had lessons recorded on video and analysed. From this evidence, numerous conclusions were drawn including the idea that a few (20%) showed ‘the spirit of AfL’ in their lessons whilst the others modelled a more rigid ‘letter of AfL’ approach. This is linked to various other attitudes and beliefs; those showing ‘the spirit of AfL’ not only are judged to have delivered more effective lessons but, on interview, are seen to be more likely to accept their responsibility for overcoming external pressures; they see themselves as the source of the solutions – referred to as holding ‘incremental views of learning’.
Very significant conclusions are drawn from the research; there are some bold claims made based on a relatively small number of interviews and observations; these are discussed in the analysis as if they hold true for many more teachers than the data set allows. The values of the researchers are clear – their belief in the superiority of ‘the spirit of AfL’ is evident throughout. However, although there is a potential credibility gap, their conclusion that, in teacher development, “beliefs and practices need to be developed together” sounds sensible. It’s worth thinking about. The more specific analysis regarding AfL really depends on whether you belief that ‘the spirit of AfL’ is inherently a positive attribute. There’s a strong values component required to accept the findings.
4. The fourth example is an MA Thesis written by a former colleague at KEGS (a selective boys school). The subject of the research was the impact of extended dialogue as the precursor to writing. The process was to engage students in extended dialogues with the teacher regarding the text they were analysing and their plans for a piece of writing. The writing that was produced was considered to be superior than previous pieces. The teacher interviewed three students about the process seeking their insight into how the dialogues has helped them with their writing. The conclusions were insightful – especially to colleagues in the same school teaching similar students. Above all, the findings were useful to the teacher herself; the process told her something about her own practice and her students.
With a sample of three students in a specific context, is there any hope that this work could yield conclusions of general significance or are the insights only valid within the very specific situation from which they were obtained?
Each of these pieces of research has value and limitations and different people will absorb them in different ways. The large-scale study with averaged out findings is set against very small-scale studies with detailed findings relating to one context.
In order to engage with any research it seems that numerous questions need to be asked. There are no easy answers. The purpose of the talk is to encourage teachers to get behind any research findings to examine the details rather than taking headlines at face value:
- Is the initial research valid enough to base decisions on? How specific were the conditions? How well-defined were the parameters of the measures? Has the nuance been averaged out?
- Is there a specific values-system at work that informs of dominates the measured outcomes?
- Does the outcome have general significance suggesting specific actions that should be taken because of the universal insight into human brain function and behaviour?
- Does the outcome provide insights and/or raise questions that practitioners should ask about the learning in their context, even if the origin is from a small sample or specific context?