A guest blog by Dr. Joe O’Mahoney (Reader, Cardiff Business School)
Dr. O’Mahoney is particularly interested in how to put critical realism into practice and how critical realism impacts the way we do research. He is the co-editor of the volume Studying Organisations Using Critical Realism which explores how critical realism impacts research methods and techniques including ethnography, case studies, discourse analysis, grounded theory, mixed methods, and probability analysis. Alongside his interest in critical realism and methodology, Joe’s work is centered around questions of organization with a particular interest in management consulting, innovation, and knowledge. He is also a member of the Centre for Critical Realism and is involved in the organization of the upcoming International Association for Critical Realism conference in Cardiff, July 20 – 22, 2016.
PART ONE: Why is it so difficult to put CR into practice?
I had the pleasure of writing this piece from Phuket, Thailand, where I spent a month learning Muay Thai. As a relative novice to the sport, I decided to do a little reading before I arrived – not just the history of the sport in Thailand, but also how to do deliver a decent attack whilst protecting yourself from your opponent. Needless to say, all this reading seemed a little pointless after around thirty seconds in the ring, where my own domain of the empirical rapidly seemed over-determined.
Recovering in my room afterwards, I turned to my holiday reading: Porpora’s excellent Reconstructing Sociology, and Lopez and Potter’s After Postmodernism (I know how to have a good time!). I hope it is not too tenuous a link to say that I was struck by the way critical realism, like Muay Thai, sounds entirely logical, sensible and understandable, until one has to engage with the real world – whereupon it becomes a little trickier. Contrary to many other philosophically informed fields, it seems that in the field of critical realism, the balance between excellent theoretical tomes and excellent empirical applications is very heavily weighted towards the former. Of course, there are an increasing number of empirical CR studies in a variety of disciplines, but of those I have read it sometimes appears that after CR has been mentioned briefly in the methodology, it is quickly forgotten. It is rare to read an empirical CR piece that actually names its entities and mechanisms, let alone the associated essences, levels of emergence, and potential and actualised powers.
In my own field at least (organisation studies), many explicitly critical realist empirical pieces spend a significant portion of their paper critiquing competing perspectives, but identify social structures only in the most general terms, with a nod towards class, gender or history. It is perhaps telling that many CR review pieces in organisation studies tend to use 1960s and 70s Marxist workplace studies (e.g. by Gouldner, Beynon, Burawoy, Pollert and Edwards) as exemplars, rather than more modern works. This is not, of course, to say that there are not many excellent empirical CR pieces, nor that it is a requirement of that CR researchers to use all the theoretical tools and concepts that Bhaskar and others have placed at their disposal. However, it does raise the questions: why is it so difficult to put CR into practice? How can we better put CR into practice?
There are several answers to the former. The first is the ostensible complexity of the CR canon: although Bhaskar was one of the easiest people to understand face-to-face, any new researcher brave enough to tackle his original work is likely to be in for a challenge. I certainly was (and still am). Certainly compared to flat ontologies, which have only to concern themselves with one phenomenon (such as discourses, practices or events), CR’s strength is that it can conceptually underpin and integrate all of these (albeit with a little modification to what ‘these’ are). Moreover, even when understood, some of the core concepts of CR (for example, that we can know – give or take – about invisible things that cannot be measured) are so far from traditional sociology that one is perhaps tempted to downplay their distinctiveness in one’s writing.
The second reason, is (dare I say!) Margaret Archer. Archer’s epic, intelligent and insightful tomes have developed critical realist sociology beyond measure and provided researchers with an array of models and tools to apply CR to social structure, identity, culture, the person, reflexivity and agency. So far so good. However, for younger researchers, building upon such a well-developed system of thought can be a near-impossible task, especially if they come to CR through a non-philosophical or non-sociological academic route. (I am reminded of the outstanding Thai restaurant a few meters from the camp at which I trained, where I took breakfast, lunch and dinner – meaning my own kitchen cupboards were all but bare).
The final reason is methods. As CR research tends to use the same data collection methods as interpretivist, constructivist, actualist (and sometimes positivist) research (i.e. case-studies, interviews, ethnography, surveys, participant observation et cetera) it is an easy mistake, especially for the new researcher, to assume that the practice and analysis of these methods should be the same. This is a common error, especially in PhD theses, but one I have also seen made in research methods programmes, where critical realism is taught in a Week 1 ontology class, and then data analysis is taught a month later along traditional inductive / deductive lines. Relatedly, the critical realist emphasis on abduction (inference to the best explanation) and retroduction (identifying the causes and conditions of one’s findings) and requires a relatively sophisticated knowledge of the potential theories out there: the plethora of potential answers to the question ‘why did the chicken cross the road?’ gives an indication of the vast numbers of potential explanations even for this simple scene, and the difficulties of choosing between them, let alone demonstrating to a reviewer or examiner that this choice is valid.
Given the difficulties of putting critical realism into practice, what might we do to help ourselves? The next section reflects on some hints from my own, and others’, experiences.
PART TWO: How can we better put CR into practice?
To help answer this question, Paul Edwards, Steve Vincent and I applied for ESRC funding to bring together a diverse group of CR scholars, all using CR in different fields and using different methods. This allowed us to run a number of seminars around Britain exploring what worked and what didn’t with CR. The result was an edited collection of different examples of CR research methods being used in research, together with reflections on what did and didn’t work. Rather than go into detail about how CR uses all the different methods, I thought I’d summarise five of insights that I gained both from editing the book, and my own (very limited) attempts to use CR in my empirical research.
- Immerse yourself
The critical realist obsession with causal mechanisms is simply a desire to understand why things happen the way they do. Although your choice of methods will be dependent upon how well you know the mechanism(s) you wish to study, a prerequisite for a deep form of understanding is for the researcher to immerse themselves in what they are trying to understand – not just interviewing, but where possible, experiencing, following and living the complexities which they seek to understand.
- Notice the absences
When doing retroduction, don’t just ask what must the world be like for my findings to occur, but also ask, what must it be like for my non-findings not to have occurred. There is more not happening in the world than there is happening, and asking why certain things have not happened may give you insights into the mechanisms or contexts which allowed this state of affairs.
- Know your sacred cows
One of the risks of using retroduction is the unquestioning reliance on theories (or assumptions). I have noticed (and I may be wrong) that many critical realists seem to be thoroughly nice people and often, for example, tend towards leftist assumptions that capitalism is bad, that feminism is good, or that a fully developed, emancipated person is a good person (do I need to say I’m simplifying here?). These may well approximate truths, but a reflexive researcher will not only know their own assumptions, but always ask ‘what evidence would be sufficient for me to question these?’
- Engage with alternatives
Critical realists like a scrap – often dedicating an unusual amount of effort to getting the boot into socio-materialists, post-modernists, actor-network theorists etcetera. Yet it is important that we recognise that the tenets of critical realism are themselves only (at best) epistemological representations (not to mention approximations) of the truth. Whilst we may disagree with the ontological theories of competing positions, we can often utilise (or even improve!) their ‘domain-level’ theorising. In fact, in my experience, empirical studies from competing ontological positions, despite claims to the contrary, often make de facto claims which are often quite close to critical realism (see for example, here).
- Prepare for failure
Every discipline has a reason (or excuse) regarding the difficulty of getting published. However, critical realism is perhaps doubly unfortunate. First, CR research tends to critique other, more established, traditions, with the result that papers are often sent to reviewers from the position which one has attacked, with the obvious consequences. Secondly, the CR reliance upon retroduction and abduction is difficult for many to grasp. From a personal perspective, I am currently attempting to persuade a positivist reviewer that three case-studies supported by quantitative data provides greater insights that that generated by a mere regression analysis. I do not hold out much hope.
- Be brave
Critical realism is still a relatively young entrant into social science which means there are great opportunities not only to use it, but also to develop it. For the latter to happen, however, we perhaps need to be braver about this world-view: being more explicit about our mechanisms, entities, levels, and domains will not only help us establish CR as a distinctive research tradition, but also help us identify better where the difficulties lie and how this might be overcome.
If you’ve got this far, you might be interested in coming to the International Association of Critical Realism Conference in 2016, or joining the Critical Realism Network. Along with Margarita Mooney of Yale University, we are also interested in creating a network of people who are particularly interested in developing a CR approach to methods. Please let either of us know if you’d like to be part of our dialogue.
For more on the question of putting critical realism into practice and questions of methodology we would also recommend reading the edited volume Studying Organisations Using Critical Realism, looking at our previous post on Critical Realist Research Methods written by Margarita Mooney.