Category Archives: Research perspectives

More on Hammersley, critique, and a framework for research design


I’ve done a couple of prezi’s, elaborating my adaptation and interpretation of Hammersley’s framework for critical reading of ethnography, which as you know by now, I think has currency as a framework for critique more generally.

Hopwood’s interpretation of Hammersley framework

The first prezi explains, with a visual accompaniment (partly prompted by a nudge in the comments to one of my posts about a lack of visuals!), how I think the whole thing works. It goes alongside the podcast I’ve published previously. See below for some brief comments on research design.

Hammersley-Hopwood framework for researchers to complete

The second prezi is very simple, and is essentially a ‘fill in the blanks’ version aimed at helping social science researchers (maybe others, too! I’d be fascinated to know if it works in other areas), to think about their research. It has deliberate (but for now unexplained) ‘fit’ with Kamler’s approach to writing abstracts (Locate, focus, report, argue). If you haven’t finished your research yet, it’s still very possible to complete the whole thing. Imagining or projecting what your claims and conclusions might be is really important. Normally we have some sense of what kind of things we might find and why they might be important. Of course we still want to leave space for the empirical world to surprise us!

Hopwood research design 5-part  framework

I just want to explain some references in the first prezi to a 4-part design framework. This is my  system for clarifying different components of research design. I’m not claiming it’s totally unique or original. It’s just the way of working / use of terms I’ve come to find useful.

1. Research Strategy – this is the big picture. Are you doing a case study? What kind of cases?  Ethnography? What of? What kind of ethnography? Longitudinal design? How long? What sequence?

2. Theoretical / Conceptual Framework – this is about the set of ideas or sets of ideas that help you frame your research in more abstract terms. Such a framework is

  • part of the location of your work in scholarly debates, fields, contexts
  • something that enables you to frame a concrete problem in more scholarly terms (and in doing so make a local or specific issue something of wider relevance)
  • operationalise something tricky to pin down (e.g if you want to research something ‘invisible’ like ‘learning’ or empirically slippery like ‘class’, or ‘sexism’, then a theory or conceptual framework will enable you ‘see’ it
  • often something that you speak to in your contribution to knowledge (advancing the framework, adding to it, applying it in new areas, challenging aspects of it).

3. Sampling and selection – this goes down a level to think about sampling. Samples include who your participants are (if any) and how they relate to a wider population. But let me be clear this isn’t a kow-tow to positivism. It’s about the fact that we nearly always study something smaller than the phenomenon of interest. This involves selections or samples in space, time, people, documents etc.

4. Methods for data generation – what it sounds like it is! Are you using a survey? interviews? observation? document collection? visual methods? why?

4. Techniques for data generation – this goes into a bit more detail, and refers to the craft and artistry involved. Within a survey, what kinds of items are you using? how have they been piloted? what kind of interview are you doing (semi-structured? life history?), and how good an interviewer are you? (probing questions, noting body language etc). If you’re observing, what are you looking for, what are you noticing? how structured is your observation? what kind of notes are being produced? when are they being written up?

You can see a video explaining this framework here.


Follow-up on limits to RCT, and also role of blogging in critique of academic papers

The post by @DeevyBee on 21st March is really worth a read.

A nice example of how a scholar reads a paper critically, identifies some limitations / problems, and invites the authors to respond (the original critique is here). In this case the original authors do respond, and this post is the blogger’s response to their response! The authors defended their methodology (on a point around the validity of a small sample or whether their study was ‘underpowered’ or not), and questioned whether a blog was an appropriate forum for criticising a peer-reviewed paper. 

Interestingly the paper authors draw on the idea of the impact factor of the journal being a direct guarantee of quality, and almost like a force field against critique. I’ve written elsewhere on the meaningfulness, or lack thereof, of impact factors.

The blog author, @DeevyBee, raises a number of interesting points that are instructive and illustrative of many important issues in scholarly research:

Small samples in RCTs

1.  “The authors reply with an argument ad populum, i.e. many other studies have used equally small samples. This is undoubtedly true, but it doesn’t make it right”. I agree – precedent (no matter how prestigious a journal) doesn’t make for sufficient argument. Prestigious journals have published studies in the past that would now be deemed unethical – does precedent override contemporary ethical considerations? How far back in history can we go in relying on such precedent? Statistics moves very fast, as computing power open up modelling possibilities.

2. “In the field of clinical trials, the non-replicability of large initial effects from small trials has been demonstrated on numerous occasions, using empirical data – see in particular the work of Ioannidis, referenced below. The reasons for this ‘winner’s curse’ have been much discussed, but its reality is not in doubt. This is why I maintain that the paper would not have been published if it had been reviewed by scientists who had expertise in clinical trials methodology.”. Now I’m not sure I would be as bold as the blogger in assuming what the expertise of the reviewers was, but from my editorial experience I do know that we editors have to make lots of compromises in finding reviewers who are available and willing to do a review, which is not necessarily the same as the ideal reviewers with the closest areas of expertise. I think the point that large effects from small trials are hard to replicate is really important. It’s not only a question of empirical generalisabilty in terms of sample:population relationships, but that the degree to which something appears to ‘work’ can be inflated.

The blogosphere as home to scholarly critique

The authors of the original paper took issue with the blogger’s decision to air criticisms in public. The blogger replies “I don’t enjoy criticising colleagues, but I feel that it is entirely proper for me to put my opinion out in the public domain, so that this broader readership can hear a different perspective from those put out in the press releases. And the value of blogging is that it does allow for immediate reaction, both positive and negative. I don’t censor comments, provided they are polite and on-topic, so my readers have the opportunity to read the reaction of Facoetti and Gori.”

In my experience it is relatively rare for journals to publish critical responses and rejoinders (this does happen, and when it does it is really good. Educational Researcher, a journal published by the AERA, does this quite a bit and it’s fantastic. First of all it shows that peer review does not end all matters of dispute. Also it shows how criticism, whether in the shadows of peer review, or more plain daylight, is part of parcel of scholarship.

Of course there are critiques levied within subsequent papers all the time, but these are not the same as more explicitly focused and engaged critiques of the sort that the @DeevyBee offered, and that you see in published responses/rejoinders. So given the rarity of the latter, I’d encourage blog-based critiques. @DeevyBee published the author’s responses, suggesting to me the point was not to have the final word, but to open up a debate, and to enable scholars and practitioners to engage with the findings (which were promising and potentially seductive) in a critical way. Indeed @DeevyBee says she is ready to revise her opinion if persuasive arguments come forward.

The responses to both the original critique and the blogger’s response are well worth reading. Perhaps one of the more direct comments, says “If you don’t want people to critique your work, then don’t publish your work”. Fair enough – being subject to critique is part of scholarship, and we cannot confine this to the privacy (and I would say often cloak-and-dagger) realm of blind peer review. We stand up in conferences and have to be ready for strong criticism. I may come to regret writing this, but I cannot for now think of good reasons why scholars should refrain from offering critiques on blogs, provided they are specific, give justifications, and done in the spirit of advancing scholarly knowledge and conversation (all of which apply to @DeevyBee’s critique in my view). 

These kind of things are not, perhaps, about resolution (particularly in qualitative social sciences where hard and fast rules are rarer), and more about enriching our understanding and building collective capacity to engage with research and not accept peer review as a guarantor of truth or quality. Peer review is important, but messy and not free of political, personal, and practical influences. A topic for a post in the future methinks…



Some comments on Ben Goldacre’s paper about the role of RCTs and evidence-based practice in education, and the response to it

Ben Goldacre has recently authored a paper Building Evidence into Education, published through the UK’s Department for Education (which invited him to do). It has been extensively (mis)reported in the press and social media. A video of a related speech is available on youtube. I am intrigued by Goldacre’s arguments, and fascinated by the response his paper has provoked, see

1. The Guardian website, following an article that bears a curious subtitle, which Goldacre corrects);

2. Goldacre’s own website, with the paper and responses.

3. Geoff Whitty’s (Director Emeritus of the Institute of Education) ‘guarded welcome‘ to the paper, on the Centre for Education Research & Policy website, which asserts the need for a degree of realism

4. Prof Mary James’ (President of British Educational Research Association, and University of Cambridge) response on the BERA website, which raises crucial critical questions and doubts

5. Twitter, where, for example @MarkRPriestly refers Goldacre to literature about the need to treat RCTs with caution in social sciences.

A bit of context

For those who may be less familiar with science policy and media in the UK, Ben Goldacre is a key public figure who, among other things, holds scientists to account in terms of quality research, and the media to account for questionable reporting of research findings. He has recently moved out of the medicine / pharmaceutical area and written on educational research. An easy reaction is perhaps to position Goldacre as an outsider – neither teacher nor educational researcher. ‘Get off our lawns’ is generally not a very helpful response in my experience. I think educational researchers have benefitted enormously from insights from outside – historians, anthropologists, sociologists, economists: why not medics, too?

I wish to lay out clearly what I understand Goldacre is and is not arguing.

What Ben Goldacre is arguing

1. That we can improve outcomes for children and increase professional independence of school teachers by collecting better evidence about what works best, and establishing a culture where this evidence is used as a matter of routine.

2. That education can reap these benefits by replicating the successes that have been enjoyed in medicine, including the emphasis on randomised controlled trial (RCT) research, information architectures, and cultures of being research literate and research engaged.

3. That RCTs are the best process we have for comparing one treatment or intervention against another, for answering questions of whether something works better than something else.

I do not disagree with any of the aims Ben Goldacre is trying to achieve – who would argue for worse outcomes for children or reduced professional independence? (NB. I am not ignoring widespread views that teacher professional autonomy has been undermined in recent years; this is an important, but tangential issue for now).

Neither do I disagree with any of the claims I’ve identified above (a selection from many arguments Goldacre makes). I do have some caveats that I would append to his arguments, and some questions and comments that might explain aspects of the response that Goldacre says surprised him.

What Goldacre is not arguing

My reading of the paper, and of Goldacre’s rejoinders to comments, leads me to the following understanding:

1. He does not suggest that RCTs should replace qualitative or other quantitative approaches to research in education. He does suggest the balance needs to swing to extend the number of RCTs.

2. He does not suggest that RCTs are a kind of gold standard for all educational research. He is generally very careful in his wording, arguing that RCTs are the best we have for answering questions of what works. He explicitly acknowledges the limits of RCTs (for example they can’t tell us why something works), and acknowledges a valuable role for qualitative research (and presumably other quantitative methods, such as quasi-experiments where interventions are applied to groups not themselves created through random allocation).

Why I agree that RCTs are important and there should be more of them in (British) educational research

RCTs are an incredibly powerful research tool. Through random allocation of participants to either an intervention group or a control group, RCTs can take care of a whole bunch of complex factors that make research involving human beings (whether medical, educational, development-based etc) difficult. As Goldacre argues, they really are the fairest test for whether something works better than something else. Note, they only ever give a comparative answer, not an absolute one: they don’t provide a complete recipe for ‘what works’. They tell us that something works differently (and we interpret that difference as better/worse) than something else. Goldacre is clear on this.

They are amazing, and on top of the examples given by Goldacre, I’d add the High/Scope Perry project, that continues to build an incredible evidence base relating to the difference that a particular approach to nursery education can make to people throughout their lives. RCTs really are like Heineken: they can reach parts of understanding that other methods cannot. As long as your question is ‘Does X work better than Y?’.

The limits to RCTs

Ask any other kind of question, and the value of an RCT quickly diminishes. Goldacre acknowledges this, too. Ask ‘what should we do?’, and until you have a new idea or intervention to try against something else, an RCT is useless. Often challenges in practice and problems in research start from a more open question – the luxury of having things to compare comes later. And other research approaches, as well as pioneering practitioners, ideological commitments to social justice, technological advances, social changes etc. all have a role to play in getting us to the point of being able to try something new out and exploit the power of RCTs to give us causally robust comparisons of X and Y. We don’t always need an RCT to make important changes, either. We no longer allow corporal punishment in schools. It didn’t (nor should it) take an RCT to show that schools are better places without the cane.

RCT’s are not perfect, nor are they the best method in all circumstances. I am in no dispute with Goldacre on this point.

There are lots of important questions where RCTs might never figure. RCTs can not tell us what it is (morally) right to do, what is just. Other approaches are often better place to identify social inequalities that might remain hidden (for example, around gender or racial differences in educational outcomes, where results from national tests are very useful). RCTs are not necessarily blind to questions of social justice (depending on the outcome measures involved). I’m reinforcing the simple point here that there are lots of questions where RCTs are not the best approach.

Goldacre is not arguing that we should ignore these questions, but there is a risk in the way he presents his arguments, that questions of ‘what works’ are heralded as the most important. Paying less attention to other kinds of questions serves the argument for more and better RCTs, but could (unintentionally) lead to a narrowing of the conceived role of research in education.

RCTs as a valuable contributor to evidence that can contribute to evidence-informed practice

It might be possible to read Goldacre’s paper and to (wrongly in my view) equate the evidence used in evidence-based practice with outcomes from RCTs. Goldacre doesn’t say this, but he doesn’t talk in any detail about other kinds of evidence, or other sorts of relationship between evidence and practice. We might, hypothetically, start seriously looking at literacy teaching practices based on evidence that current curricular and testing regimes disadvantage certain students and reproduce social inequalities. If the only evidence that could influence practice was that coming from RCTs, we would have to ignore the other evidence that the status quo is grossly unfair.

Evidence is valuable to practitioners in helping point to ‘what works’ but is also valuable in other ways, and these alternatives are played down in Goldacre’s paper, in his construction of an argument that seeks to redress a perceived imbalance and neglect of RCTs. Such diverse evidence-practice connections apply in medicine, too – there are lots of things that doctors advise (eg suggesting you give up smoking) or do (eg heart transplants) based on other kinds of evidence, not just RCT outcomes.

So I agree with Goldacre – we do need more, better, and more joined-up RCTs, because they are so powerful and the best tools for comparing two or more approaches against each other. But to avoid possible over-extensions of this argument, it is important to be very clear about the important role of other kinds of evidence. It’s not that other research is useful for other things, and only RCTs can be used in evidence-based practice. Evidence-informed practice (a phrase I prefer because it points better to the requirement for professional judgement in interpreting evidence, which Goldacre mentions), can be enriched by all kinds of evidence.

What is educational research for?

I think Peter Mortimore (2000, p. 18) captured some of this in his writing on the role of educational research:

“Who else but independent researchers would risk making themselves unpopular by questioning the wisdom of hasty or incoherent policy? Who else could challenge inspection evidence and offer a reasoned argument as to how empirical faws had led to erroneous conclusions? Who else would dare say ‘the King has no clothes’? Who else would work with teachers and others in the system in order to look below the surface:

  • to notice the unfairness suffered by those who are young for their school year yet for whom no adjustment is made to their assessment scores;
  • to count, and to identify variations in, the numbers of minority pupils excluded from school;
  • to point out that many of the supermarket shelf-fillers are our further education students trying to get by financially;
  • to investigate whether adult learners need the same or a different pedagogy from pupils;
  • to make fair comparisons of schools, as opposed to the travesty of league tables;
  • to tease out why poverty is associated with failure in a competitive system, in which only so many can succeed, rather than just being an excuse for low expectations or poor teaching;
  • to monitor trends and changes in educational aspirations, attitudes and attainments.

On the relevance of the medical model

Goldacre argues that education has much to learn from medicine in terms of the way research is conducted, and in particular the way practitioners are involved in research and the way doctors are trained and expected to be research literate.

As an educational researcher who has in the past done considerable work based in schools, there is no greater reward than thinking what you have found out has made a difference to the lives of teachers and/or pupils. There is no greater insult, to me at least, than to find one’s research accrues value only through citations by other researchers and makes no connection to the ground. As Goldacre rightly suggests, this connection is not only a property of the quality of evidence. It is also a question of the (perceived) relevance of that evidence, and on the user/reader’s ability to make good sense of that evidence. The relationship between evidence and practice isn’t as simple as finding out X is better than Y and then making sure all teachers read the relevant paper.

So I would be quick to welcome and celebrate a shift in school teaching cultures that brought teachers into a different relationship, both more routine and more critical (as Goldacre advocates), with research. I would add this should be with the full richness of the evidence base that educational research has to offer, not just outcomes from RCTs (and I don’t think I’m contradicting Goldacre here either).

Medicine and education may not be that different in some respects

Goldacre’s paper, and the responses and rejoinders online, do raise questions about how valid or relevant comparisons with medical research, forms of evidence, and practice are. Many (including Goldacre) have rightly shot down crude arguments that medicine is about ‘physical stuff like cells’ and education is about ‘people’ or that all diseases/patients are treated the same, that medicine is devoid of the kinds of social complexity that pervade education. One only has to look at the complex deliberations underpinning the development of NICE guidelines to understand that what happens in hospitals and doctors’ surgeries is not simply a question of knowing ‘what works’. Questions of what can be afforded, what is politically acceptable (remember headlines about postcode lotteries in breast cancer treatment?), what is practical – these all have a bearing too.

It’s easy to police boundaries and protect education from medical experts by screwing our eyes shut and shouting ‘but classrooms are different!’ as loud as we can. There are presumably some very important things to consider about the nature of medical and educational research and practices, and whether elements from one system might inform those of another (and shock, horror, this might even involve medicine learning from education!). I don’t think Goldacre offers an adequate account of these issues, but at least he acknowledges them. My critique is not of Goldacre’s oversight (there’s only so much one can say in a paper or a 20 minute talk), but of the risks that others simply elevate medicine as an ideal type and naively expect education (and other systems) to follow.

I am, however, curious to learn more about the sorts of trials Goldacre argues can be done cheaply, efficiently, and effectively in education. As he points out, medical trials have developed into complex designs involving treatments that are not just based on one pill or quantifiable medication regimen versus another – trials of psychotherapeutic interventions, for example. My understanding is that steps must be taken in RCTs in education to ensure compliance – that what is being done in classrooms is actually what the trial is supposed to be testing (as happens in some medical trials). This often requires teams of researchers to check and observe what is being done, which is very expensive, or places a burden of documentation on teachers. I’m not saying it’s not possible. I’m saying that I’d like to see Goldacre’s vision of cheap, robust RCTs (which involves all sorts of considerations about levels of randomisation and their relationship to levels of outcome measurement) explained in more detail.

Why medicine over other models?

What Goldacre doesn’t address, except perhaps indirectly in his introduction where he attributes the leaps forward made in medicine to RCTs, is why medicine should be held up as a model to replicate rather than other systems. Sure, medicine has come a long way in recent decades. So have other aspects of life, too. Why are educational practices and evidence best approached in a medicine-like fashion? Why not sociological? Anthropological? Why not arts-based? Why are medical models better than approaches that might acknowledge things that RCTs can’t, like morality, or justice? Answers like ‘because those aren’t objective’ or ‘can’t be tested fairly’ miss the point (it should be obvious why). I’d love to see Goldacre develop more sophisticated arguments as to why medicine should trump other conceptions of evidence and other notions of evidence-practice relationships. I am guessing there is more to this than the mere fact that Goldacre knows medicine best, but we would benefit from further explanation on such issues. I’m not defending educational research against outside influences. On the contrary, I strongly believe we are and will be better off for these. But we should understand this as a complex choice, with significant implications depending which perspectives get left out from arguments.

On the defensive reaction by some qualitative educational researchers

Goldacre replies to some of the responses to the Guardian article as follows:

“It is very odd, I think we’ve seen some rather peculiar protectionism here from qualitative researchers working in education. I’ve not seen this attitude among the very good multidisciplinary teams working on mixed methods approaches to medical research, where quantitative and qualitative research is done harmoniously with mutual respect, in my experience at any rate. It may be a peculiarity of the qualitative research community in education, or it may be that we are seeing only bad apples in this thread. I don’t think they do their profession any favours”.

What might lie beneath ‘protectionism’? Why might qualitative educational researchers react differently from their medical colleagues in mixed-methods teams? Why would we expect them to react in the same way?

Notwithstanding the histories of marginalisation that many qualitative researchers would argue they have suffered at the hands of pseudo-scientific dominance in educational research, I think part of the explanation lies in some of the ways in which Goldacre’s language might be interpreted, and the genuine sense of threat that such interpretations could pose to some scholar’s values, ethical commitments, and livelihoods.

Why might qualitative educational researchers (of which I am one), react differently from medical researchers in mixed-methods teams? Maybe because many of us are not in mixed-methods teams (for better or worse), but instead collaborate in other ways, for example working with teachers and schools in solely qualitative paradigms. Arguments that the pendulum should swing back to re-emphasise RCTs can be interpreted as a move that will diminish the place of other approaches. This was not Goldacre’s intention, but this is what many perceive has happened in the US as a result of the way federal funding for educational research has been allocated. Protectionism seems quite understandable, as part of a professional ethos that preserves mutual respect and place for different kinds of research (an ethos that Goldacre himself subscribes to). What surprises me is that Goldacre was surprised by this reaction.

On the representation of qualitative research

Goldacre writes:

“Qualitative” research – such as asking people open questions about their experiences – can help give a better understanding of how and why things worked, or failed, on the ground. This kind of research can also be useful for generating new questions about what works best, to be answered with trials. But qualitative research is very bad for finding out whether an intervention has worked… The trick is to ensure that the right method is used to answer the right questions.” (p.13)

I agree wholly with the point that the right method is used to answer the right questions. In my view Goldacre’s paper does not adequately capture the range or value of qualitative approaches, and risks them being positioned as subservient to trials. I do not follow qualitative researchers who campaign against RCTs. I think we should have more of them. But this should not be at the expense of other approaches, and certainly not based on accounts of qualitative research that convey a potentially misleading and diminished view of what the alternatives are and what they offer. Goldacre does not clarify the extent to which he thinks ‘what works’ questions should trump other questions. Protectionism may reflect concerns that others may take Goldacre’s arguments as a basis for a narrowing of the kind of question (and by implication the kind of research that is valued) in educational research.

Indeed, Goldacre makes the very good point, I think, that educational research (or at least that which focuses on teaching and learning in schools), could be enhanced by pursuing agendas and questions from the ground up – ie. those identified as priorities by teachers. This would be very welcome, although I would always seek to preserve space for outsiders to pose questions, too, for they can often challenge assumptions and see possibilities that are difficult to imagine from the inside. But the bigger problem here, is if RCTs become a blanket preferential mode of enquiry (which is not what Goldacre advocates, but is not implausible). Rather than opening up the possibility for teachers to lead the direction of research, this would close it down by limiting the kind of questions that teachers can ask to those of a ‘what works’ variety. There are myriad other important kinds of question that teachers want to ask, too.

The overlooked value of locally-based, locally-relevant research

There’s something curious about Goldacre’s critique of piecemeal individual projects that are oriented to figuring out what works locally, and his open admission that RCTs don’t often generalise: there is rarely going to be a ‘what works’ solution that applies to all schools, age groups, subjects etc. I agree that isolated pockets of poorly supported research that never leaves the boundaries of a particular institution isn’t a great set-up. So yes, we need more joined up infrastructure, for research and for disseminating and sharing evidence. But could not such local projects also be ideal ways to test out, empirically, and in an evidence-based way, how local conditions shape the meaning of RCT outcomes developed elsewhere? Might not some of these projects into which teachers pour their heart and soul, which Goldacre criticises for turning out to be too small, lacking robust design (p.17), in fact be avenues for translating distant evidence into locally relevant forms?

A related point concerns the critical research literacies mentioned by Goldacre. These are of course important, and if teaching is to benefit from any kind of research evidence, there must be critical appraisals of that evidence. But that critique cannot be limited just to understanding RCTs and ‘what works?’ kinds of research. Such critical skills should also involve understanding different approaches and their value. Goldacre doesn’t close off on the kind of critical understanding he’s advocating for, but I think it’s important to be really clear that a narrow RCT-focused literacy will not suffice.

On the perils of misinterpretation

In what was mentioned above we find an example of how some of the language used might have (perhaps unintentionally) provoked the strong reaction from some qualitative researchers. There is potential for readers to infer from Goldacre’s wording an equation of ‘small’ with lacking robustness (and such readings are readily apparent in the comments on the Guardian webpage). If big sample = better research, then we should look away from RCTs and more to statistical analyses of existing datasets, and the use of various regression models to figure out which schools are performing best. Sample size is a poor proxy for research quality. Goldacre knows this, but not all of his readers appear to notice this point.

There are other moments, too, for example when Goldacre mentions the risk of pilot studies ‘misleading’ on benefits and harms. I agree such risks are real and important. But it is not the pilot itself that poses the risk. It is the flawed interpretation or application of findings that poses the risks. Qualitative researchers might be forgiven for interpreting what was written as laying the problem at the door of qualitative research itself, rather than at the door of those who mis-use or abuse its outcomes. Yes RCTs are the only true ‘fair test’ but this doesn’t make other approaches ‘unfair’ provided they’re not doing a certain kind of test. Goldacre knows this. Many of his readers may miss this point.

Then there is the different language used in the 20 minute talk, which was less precise in its wording and thus more open to misinterpretation. Goldacre spoke of people being ‘horribly misled by weaker forms of evidence’. Any evidence, from an RCT or otherwise, has the potential to horribly mislead. Any evidence, from an RCT or otherwise, may be strong or weak depending on the question. The care Goldacre took in his written paper to manage these issues was less evident in the speech, in which listeners could easily be led into equating RCT with strong evidence, and other approaches as misleading and weak. This only applies to ‘what works’ questions. This is reinforced by phrasing that links good quality evidence with RCTs, again without explicitly placing a caveat of ‘only if we are asking is X better than Y’. And again in talk of swinging the balance towards more robust quantitative research. More robust than what? The potential for listeners to interpret this as a slight against qualitative research, or as a suggestion that qualitative evidence by definition lacks robustness, is clear.

As an aside, Goldacre also contrasts ‘nerdy academics’ with ‘teachers on the ground’ – setting up another potentially damaging binary. In particular this kind of talk fuels the Govian anti-academic rhetoric and misleads the public into outdated conceptions of ivory tower academics (see Pat Thomson’s blog on this). Many educational researchers are in schools week in, week out, working with new and experienced teachers. They are ‘on the ground’. They are also ‘on the ground’ because most academics are also teachers, themselves. This applies not only, but particularly, to educational researchers. If having a scholarly or theoretical interest in learning and pedagogy makes us nerds, then I’ll wear the nerd badge with pride. But I do take issue with characterisations that reinforce notions of aloof nerdiness against on the ground realism.

Another binary set up in Goldacre’s talk is between evidence-based practice on the one hand, and leaving everything to individual professional judgement. I’m convinced Goldacre has a more sophisticated view of practice than this – his writing about the need for critical appraisal of research suggests so – but again this kind of phrase can provoke defensive reactions, and risks being taken up in unhelpful ways if not set in a wider context.

On the risk of over-promising

Finally, there are some other very real risks that Goldacre himself acknowledges. He rightly says that evidence based practice isn’t about telling teachers what to do. As if evidence (from RCTs or otherwise) could ever be so prescriptive. Goldacre imagines a greater role of RCTs and networker participation of teachers in research, supported by experts, and feeding into two-way information architectures of setting the profession free from governments. Forgive me if I don’t hold my breath. For the simple reason that even if we were to achieve everything Goldacre sets out, it would offer few guarantees to children’s outcomes or teacher professional independence. Goldacre does not imply otherwise, but does not engage adequately with other features of the political-practice landscape.

Many teachers and educational researchers share a view that the education system in the US, which Goldacre notes funds way more RCTs in educational research than the UK, is straining – with many school buildings in urgent need of renewal, and high-stakes testing policies asserting a significant influence on practice. Outcomes from the What Works Clearinghouse are undoubtedly valuable, but do not land in a tabula rasa. And of course, not all RCTs change practice.

Even with more, better RCTs, and research cultures and information architectures of the sort Goldacre imagines, without stability in other aspects of the education system (for example in curriculum content, examinations, accountability structures, inspection regimes), any knowledge of ‘what works’ seems likely to be reduced in value either through short lifespan (it worked in the old system, but not the one now), or by simply failing to register on the radar in a profession that is straining from incessant change. I agree in theory, what Goldacre proposes might play an important role in emancipating the profession from ‘the odd spectacle of governments telling teachers how to teach’ (p. 19). I wonder how likely the promise of this is to hold true.

Some of the protectionism that Goldacre is surprised to see, and so strongly puts down as reflecting poorly on our profession, may in fact be understood as people with passionate commitments to precisely the same aims and improvements that Goldacre wishes to see, differing in their confidence in the whole of his vision becoming a reality, and clear in their understanding that even if it were to be realised, it would not be enough to secure the kind of conditions they feel best serve children and teachers.

I do educational research (and I’ll confess, it is of a qualitative kind most of the time) because I think research-based evidence has a lot to offer teaching and learning. Like Goldacre I don’t think we have exclusive rights to this kind of influence, nor do I think there is no space for political ideology either. I’m all for more evidence, better evidence, greater research literacies, more joined up research, and weaker divides between academe and schools. But please let us treat visions such as that set out by Goldacre with the careful and critical reading to which we should subject research.


Mortimore, P. 2000, ‘Does educational research matter?’, British Educational Research Journal, vol. 26, no. 1, pp. 5-24.

Martyn Hammersley’s framework for critical reading of (ethnographic) research: why I like it

This is just a short blog post to accompany a linked podcast, video, and prezi that go into these issues and the framework in more depth.

I’m often involved in teaching students about critical appraisal of educational / social science research.  I’m not convinced by arguments that we should judge research only by the criteria that apply within a particular perspective or paradigm. Notwithstanding my prior post, based on Schatzki’s arguments, about why ontology is important and how it changes the game in terms of judging research, I do believe that there are some dimensions of research that can be subject to a broader-based critique.

This refers to a framework presented by Martyn Hammersley in chapter 2 of: Reading ethnographic research: a critical guide, published by Longman (eg. 1998 2nd edition).

I think Hammersley’s framework (originally written with a focus on ethnographic research) provides a sound basis for precisely such an approach. The content does not overly prescribe what good research is, nor does it replace rules, conventions and quality criteria associated with particular perspectives or approaches.

But, as I say in the podcast, I’ve yet to come across a piece of social science research where asking probing questions about the focus, empirical context / case, methods, claims [and their links to the case] and conclusions [and their links to the focus] have not been useful as a means for assessing research quality.

Reader-listeners will detect my strong attachment to the idea of ‘evidence’ in educational / social  science research. I doubt everyone shares this, and I’d be surprised if everyone agrees with the views expressed in the podcast.

Part of my motivation for the podcast was a reaction to constructions of Hammersley (and others like him) as rather old-fashioned empiricists. I hope the podcast shows how a concern for evidence, quality of evidence, and relationships between claims and evidence does not automatically position one as a naive realist who’s never heard of the crisis of representation etc.

I conclude the podcast by arguing that the aesthetic dimension of research (something I’ve blogged about elsewhere, too), is something that is not excluded from Hammersley’s framework, but isn’t given the emphasis that it might deserve. I suggest that incorporating aesthetics into assessments of research quality (inspired by Silvia Gherardi, Antonio Strati and others), follows through on the original spirit of Hammersley’s framework. Hammersley is very careful in setting up a position that rejects a doctrine of immaculate perception, and has an explicit role for modes of writing, relationships between researchers and participants, and varying degrees of insight, inference and so on. I simply suggest that highlighting these complements and enriches a focus on claims and evidence.

In summary: a lot can be achieved in terms of critical appraisal of educational or social science research by thinking about:

1. The Focus (wider topic), its articulation (scope, boundaries), importance, relevance

2. The Case(s) studied [not that all research is a case study] – the spatially and temporally limited aspect of the wider focus that is the actual subject of empirical research

3. Methods – including processes through which data are generated [and not collected: see Pat Thomson’s blog for more on this], relationships between researchers and participants, analytic techniques etc.

4. Claims made about the case – different kinds of claim and the different kinds of evidence that would warrant them

5. Conclusions drawn – not letting go of evidence completely, but saying something about the wider focus, moving beyond the specific case (eg. via theoretical inference, empirical generalisation).

If we concern ourselves with these questions, and relationships between focus, case, methods, claims, and conclusions, while keeping a close eye on evidence (whatever that may look like), we can’t go far wrong. And if we are sensitive to aesthetic dimensions when we do this, too, so much the better!

Theodore Schatzki on why ontology matters in educational & social research

I was inspired to write this, and base much of the content on:

Schatzki T R (2003) A new societist social ontology. Philosophy of the Social Sciences 33(2), 174-202.

Schatzki considers whether social [and educational] researchers should simply to implement methodological strategies for investigating social affairs and to avoid ontology altogether—for ontologies are nothing but unnecessary and empirically unconfirmable presumptions.

In my words, what he is saying here is: it could be argued that ontology is an abstract philosophical concept best left to philosophers, while social scientists get on with rigorous empirical enquiry.

When Schatzki talks about the ‘social’ he does not mean social in terms of the opposite of anti-social (ie sociable). The social refers to things pertaining to human existence. All educational research is thus social research on these terms.

Schatzki: Types of ontology however, have implications for method.

Schatzki tells us that ontology affects:

  1. The choice and use of particular methods
  2. The inferences that are made from observations and measurements to statements about social matters [ie. how we interpret evidence]
  3. The formulation of these statements [ie. The kinds of knowledge claims we make, and the degree of certainty and universality that apply to them]
  4. What of the social is thought to be directly experiencable, observable, and measurable [ie. Which phenomena we assume we can directly see or measure. This gets messy quite quickly when, as in educational research, we are making claims about social phenomena such as learning, educational achievement, boredom, interest, motivation, emotions etc. Can we see these things? What does evidence of these look like?

Schatzki says: these matters will vary depending on whether a researcher is an individualist [someone who believes individual free will or agency control the world] or believes in social structures [things like class and race as primary influences on human life]… or social facts distinct from facts about individuals (and maybe their relations).

He then considers: An investigator might proceed oblivious to this dependency and simply carry on research. How he proceeds, however, will implicate stands on these issues that collectively affirm at least some general type of ontology (e.g., individualism).

What Schatzki is saying here is that like it or not, and whether we are aware of it or not, as soon as we embark on research, we are making ontological assumptions. They are not optional extras. They are inevitably, always already part of the process. You don’t choose to have an ontology or not. (You can choose which kind of ontology you work with).

Schatzki, as a philosopher, then tells us of the benefits to social research of being explicit about our ontologies, and exploring differences between them. He calls this ‘the advantage of ontological self-consciousness and choice when studying the social world’.

When I was a student I was often sceptical of the value of such self-consciousness, which seemed at times like naval-gazing, abstract introspection ladled with technical terms designed to confuse and frustrate (see another of my posts for more on my love/hate relationship with research perspectives as a set of concepts).

Schatzki then considers how a ‘methodologist objector’ (someone who is less convinced of the advantage Schatzki speaks of) might reply:

OK, a variety of ontologies might inform social research. There is still no reason to argue for and against particular ontologies. Justification in social science is empirical validity, and ontologies cannot be empirically tested.

What is being argued here is that the advantage of one ontology over the other is not something that can be put to empirical test. (NB. This is not the same as the advantage of ontological self-consciousness, which is being aware and explicit about ontology).

As social scientists we care about empirical evidence in leading us to make some claims about the world and to reject or question others. If we can’t test our ontological assumptions this way, then they are just arbitrary choices. Not very ‘scientific’ at all.

Schatzki considers one way of testing ontologies: to see which is the most successful research, and then go with whatever ontology that research is based on.

Great. Easy.

Too easy (as they say in Australia 😉 )

But. Big BUT…

Schatzki says: There are at least two problems with this response. First, unanimity does not exist in social science about what counts as a ‘successful’ research program.

In other words, not all social research projects have the same success criteria. They have many different aims. So establishing which is the most successful research isn’t going to be easy. Or fair.

Schatzki again: Second, and more specifically, what counts as success often reflects ontology.

What good research looks like (and I’ve written about this before) depends on the ontological assumptions upon which it is based, not just our aims (though these are linked).

The person who says ‘identify the best research and follow whatever ontology it uses’ is asking us to identify what ‘best’, but ‘best’ cannot be judged independent of ontology. It’s a circular argument that gets us nowhere.

Schatzki then offers us a genius phrase: So empirical validity is not ontologically innocent.

He adds: And because ontology is tied up with social research, there is room and need in the overall enterprise of social research for ontological disputation.

For me at least, he has won the argument. Even as a researcher who really cares and worries about data and evidence (in a way that some would regard as old fashioned), I simply can’t escape the need to be thoughtful and clear about questions of ontology.

Don’t like philosophy? Don’t like long words that end in –ology or –ism? Want to just get on and do some proper social research? Tough sh!t – ontology is coming your way whether you like it or not. In fact, it’s already there.

Why the idea of research perspectives is brilliant and annoying at the same time

What are research perspectives?

This is a term that is often used to describe different approaches to research. These differences are generally understood as being rooted not simply in the focus or methodology, but in deeper ontological and epistemological foundations of research. Ontology concerns assumptions we make about reality. Epistemology is our theory of knowledge, and how what we come to know relates to that reality (or those realities). Methodological approaches and study foci can in some ways be seen to flow from these deeper (philosophical) points of view.


Brilliance lies in the fact that acknowledging different research perspectives (others may use the term paradigm, though it’s not quite the same thing), we are forced into a number of important realisations:

  1. Not all researchers understand this thing called ‘knowledge’ in the same way
  2. So… the enterprise of doing research in order to advance knowledge is understood very differently. Some are looking to discover knowledge that gets close to a single truth. Others are looking to create knowledge that provides different possible answers to the same questions.
  3. By implication, what it means to do research well changes according to the kind of research we are doing. I like metaphors, so let’s think of this in terms of Olympic runners. We can compare the running style of a sprinter and that of a long distance runner. Who is the better runner? The sprinter moves quicker, and develops and uses her body to effectively cover short distances. The marathon runner develops and uses her body differently. It’s not fair to say one is better than the other because they are trying to do different things. What ‘good’ running is depends on which race you’re running in. What ‘good’ research looks like depends on the perspective taken. Bing!
  4. Finally brilliance in the idea of perspectives lies in the fact it forces us not to take knowledge, evidence, methods, data, and truth for granted. If we are in the business of producing new knowledge we need to take these things seriously, not brush them under the carpet.


But we have to be clever in the way we work with these concepts. Why?

  1. These are conceptual categories that have a mixed, sometimes quite problematic relationship with actual research practices. Many studies don’t fit neatly into one or other category.
  2. Categories tend to turn messy, blurred boundaries into neat, separate entities. Most people who write about research perspectives acknowledge this – we need the concepts as sign posts and to give us some clarity of thought; but at the same time we need to be flexible and hold them loosely. Aargh!
  3. People can develop a security in applying long words that end in …ism as a kind of badge or label that fits their research, or even themselves as a researcher. But like many things in the social world, research isn’t a stable activity, and projects may evolve, researchers may change their views, or hold contradictory views at the same time. Badges have their limitations.
  4. Badges or terms like ‘interpretivism’ also turn into chunks what might be better conceived as a continuum, or even a big set of splodges and squiggles (like a modern art painting maybe). Many books write of positivism, post-positivism, interpretivism, feminism, critical approaches, poststructuralism etc (and the terms used to describe the same, or near-same things vary; that’s another issue!). But however long the list, there’s always more. Practice theory has a ‘site’ ontology (see the annotated bibliography of Schatzki). Actor-Network Theory makes other ontological assumptions again. But they have some points of overlap.
  5. I’ve often been asked, what about a big quantitative study that studies gender inequality in schools? The quantitative stuff perhaps signals a post-positivist perspective. But the gender inequality might be redolent of critical or feminist approaches. Which is it? To return to our metaphor: it might not be clear which race is being run: Is there a sprinter warming up on the starting line for the marathon? Or is there some hybrid or complex combination going on? (Here’s where the metaphor runs dry, [excuse the pun]). That’s the difficulty when we set up categories like all the …isms. But at least the categories have been useful in getting us to think about the assumptions made by the researchers and their purposes.

So caveat emptor – buyer beware: use these concepts cleverly, and with caution.

PS. There is nothing new here. I’m by no means the first person to write about these issues.

Wise words on originality and knowledge

Just a short one today, posting a link to David Baume’s latest in his series of blogs about originality and knowledge.

I’m going to have to think about links between originality, aesthetics, creativity, and beauty in research, and in the knowledge that research generates.

Was prompted by discussions with Kathryn Sutherland yesterday (@SutherKa on twitter) to think of alternative to ‘impact’ as a measure of quality in research. She asks can we find the joy in our research? Good question!

Quality, parsimony and beauty in educational research

So… my first post aimed at supporting students studying research perspectives (UTS 013952) – which covers issues about quality in educational research, philosophy, what it means to produce new knowledge etc. At the heart of this is learning to be critical – not to take what you read for granted or at face value.

One thing I have often noticed is that people leap quickly onto the critical part of critique (ie. picking  holes, identifying limitations or shortcomings), and forget the equally important part: giving credit where it is due, identifying strengths. Think about it in terms of food shopping – we could go round the supermarket giving reasons why all the products are rubbish, but without a sense of what good food is and an ability to know it when we see it, our trolley would always be empty.

One really useful text for getting started on these issues is (chapter 1 in particular):

Yates, L. (2004). What does good education research look like? Situating a field and its practices. Maidenhead: Open University Press.

In the opening chapter, Yates notes that many people give one or more of three responses when asked what good research is:

1. It is technically good – systematic, tight, well designed etc.

2. Makes a contribution to knowledge – shows something we didn’t know before.

3. Achieves something that matters – which in education people often take to mean makes a difference to teaching and learning in classrooms.

Yates (quite rightly in my view) debunks each of these. When I’ve asked students in the past about good research, many have used words like ‘objective’, or ‘unbiased’. These point to the technical theme (1) from Yates. Research should be done well. We can’t cut corners, be sloppy with our concepts or methods. We have to think carefully about things like samples, the tools we use to generate data, and processes for analysis. But I would like to complicate ideas about what technical quality might look like. Can something be subjective and still technically good? In some circumstances, yes! It depends on your perspective – what ontology and epistemology you are working with.

Couldn’t a piece of research be technically good but still rubbish? Let’s think about designing something, maybe a car. It could have perfect components, a finely tuned engine, but be shockingly ugly, too wide for roads, too long for car parking spaces, too high to pass under bridges. No beauty. No utility.

Beauty? Utility? In research? Well the utility part links to the 3rd response above – making a difference to something that matters. I agree – there is an infinite number of questions we could ask about education, and I don’t think all of them are equally worthy of our attention as researchers (and the money of the people who fund research, who are often taxpayers!). But research can be useful in many ways, not just identifying ‘what works’ (here I am poking at a major preference in the USA for a particular kind of research that promises this kind of outcome; see here for more info). As we continue in class and in this blog, we’ll think more richly about what utility might mean.

What of beauty? I think good research does have an aesthetic quality that is often overlooked. A kind of elegance that comes from a great question that cuts through to the nub of an issue; snappy, tight concepts that give us something to work with without over-complicating (I’ve read hundreds of studies that tell me X or Y issue is more complex than we thought. Yawn); a neat design (size and scale aren’t everything); and a focused, insightful analysis. Statisticians don’t just apply rigid mathematical formulae; they make judgements when they build models of the world, and one of them can be framed in terms of parsimony – striking a balance between explanatory power and complexity. I think parsimony is a quality of all good research. But it’s not a 1 or a 0 kind of concept. More a question of grey areas than black and white: judgement; aesthetics; beauty.

The quick-witted amongst you have noticed I have ignored the 2nd response, that of making a contribution to knowledge. But have I? When I said not all of the infinite research questions are equally worthy of our attention, didn’t I imply that not all new knowledge is equally valuable? Research that doesn’t lead to new knowledge isn’t (by my definition) research, let alone good research. But that doesn’t mean all contributions to new knowledge are good research. What if I ‘found out’ that students do best if 100% of their classes are 1:1 with teachers with PhDs in their subject area? It might be ‘true’; based on a technically competent (even parsimonious) study; and as far as I know, no-one has shown this to be the case before. There’s my novelty. But so what? What’s the point at arriving at a conclusion that is so disengaged from realities of politices, budgets etc.?

Now I’m going to contradict myself, and leave you with the question that maybe good research questions and challenges the status quo, including dominant political ideas, assumptions about money and funding etc. Maybe something of the beauty in educational research is precisely the ability to take evidence and to use to imagine new possibilities, new ways of facilitating learning, to provoke new dreams of justice and equality? How else will we break persistent cycles of inequity if we don’t use research to do this? Adding a brick to the brick wall of existing ways of thinking is fair enough. Maybe good (beautiful) research lobs bricks through it, knocking a hole in the opaque edifice and giving us a glimpse of what might lie beyond?