Abstract
A man, though wise, should never be ashamed of learning more, and must unbend his mind.
To start, it is worth appreciating that there are many “schools” of realist research with a bewildering array of names. Each has interpreted and applied the works of the “father” of realism, Roy Bhaskar (Bhaskar, 1978, 1979), slightly differently. The result is they each do things differently for differing reasons. The implication is that a researcher wanting to use a realist approach must first decide on which “school” to follow. Once selected, while there may be some leeway in interpretation and/or application, the followers of the chosen realist school will have certain ontological, epistemological, conceptual, and methodological expectations. Make sure you understand what these expectations are at the start of your journey. Following a certain school will take you down a different path, though these paths may overlap in parts.
The path I want to now travel along is that of the realist approaches, realist evaluation (Pawson & Tilley, 1997), and realist review (or synthesis—the terms are synonymous; Pawson, 2006). Some of what follows applies only to realist evaluation and/or realist reviews, but where there are wider implications, I will point these out.
Common to all realist approaches is a need to unbend the mind in terms of conceptualizing how the world is constituted or “works”—that is, our ontological assumptions. Most realists share the view that hidden, context-sensitive mechanisms cause outcomes, that the world is stratified, and that our knowledge can only ever be partial (Astbury & Leeuw, 2010). Everything else are embellishments of these assumptions. These assumptions drive much of the thinking among realists about knowledge claims and methodology. Failure to understand the realist assumptions within each school is something to be avoided as in my experience has often led to misapplication and/or unwarranted knowledge claims. The tip here is to prepare for the journey. Make sure you understand realism before setting out. In particular, make sure you know what realist mechanisms are and are not!
Specifically, for realist evaluation and realist reviews, I want to raise your awareness of five potential conceptual obstacles that you may face on your journey, namely—program theory, context (C), mechanism (M), outcome (O), and context–mechanism–outcome–configurations (CMOCs). Providing detailed methodological guidance is beyond the scope of this editorial, my intention is more to arm you with overarching strategies to ease your journey.
Program theory is very helpful—it can both guide what you need to do before and during your project and provide explanations of how and why intervention components work and how and why they may be optimally implemented. Take time to develop one (or more as needed). The key here is to have an educated guess early on, for example, in planning the project or once you have started. It does not matter if it is incomplete or even if it does not seem that realist in nature. The point of your realist evaluation or review project is to gradually develop and refine the program theory so that it is more detailed, realist in nature and the inferences within it are supported by data.
The dastardly gang of C, M, O, and CMOCs will at every opportunity harry you on your journey. First, don’t be intimidated by them! They are but tools to help you think about how to analyze your data so as to provide an explanation that is realist in nature. Sense making is not an easy process and takes both time and, in realist evaluation and reviews, is an iterative process. When you have data, rather than just jumping in straight away to conceptually categorize “things” as a C, M, or O, gather all seemingly relevant “thing” into one place. For example, it might be that you gather all data that seem to be related to an intervention strategy into one place prior to analysis. Or if in your program theory there is a concept that seems to important, gather all data that seems to be related to that concept together. This initial “sorting” process puts all seemingly relevant “things” in one place and can make it easier to move on to the next stage—iteratively developing CMOCs.
When developing CMOCs, remember that what you are doing is providing a realist explanation for an outcome. This outcome happens because of these mechanism(s) which are triggered under these context(s). Cs, Ms, and Os don’t free float—“something” is not worth calling a C unless you can infer that it is a “trigger” for M that causes an O you are interested in. In other words, Cs, Ms, and Os are configured—organized in such a way that they make sense. One potentially helpful trick is to think “backward” from the outcome. Ask yourself, for this outcome what caused it (the mechanism) and under what contexts was the mechanism triggered?
And don’t take you eye off seemingly innocent O. Within your program theory, there are likely to be multiple outcomes that may need to happen before your final desired outcome(s). This means that you are likely to need to construct more than one CMOC—one for each outcome “proximal” to the final desired one.
But what do you do if you become unstuck with the issues I have mentioned earlier or others? Well, there is no substitute for reading—though it is important to read the “right” stuff. For realist evaluation, there is Realistic Evaluation (Pawson & Tilley, 1997) and for realist reviews, Evidence-Based Policy: A Realist Perspective (Pawson, 2006). The Science of Evaluation: A Realist Manifesto delves deeper and provides the background into the origins of both approaches (Pawson, 2013). For more hands-on guidance, the RAMESES Project has a range of resources of realist reviews (and soon) realist evaluations (www.ramesesproject.org). Some find reading published examples of realist reviews or evaluations a helpful source to learn from, although my caveat would be that the methodological quality of some published examples may mislead more than inform.
Finally, there is often no substitute for asking others who have been on the journey. You do not necessarily have to be alone on your journey. Do seek out fellow travelers, band together, share the load, and enjoy the companionship. While we may all be separated geographically, technology does (on this occasion) come to the rescue. The RAMESES Project maintains an e-mail listserv that is open to all (www.jiscmail.ac.uk/RAMESES), and there is always the telephone, e-mail, and Internet-mediated communication. There are no silly questions, just those silly enough not to ask for help when they need it!
In summary, while realist research approaches may be the path less well trodden, you are not alone in your journey. Prepare well by first making sure you understand what realism is (and is not) and what implications there are on concepts, knowledge claims, and methodology. This may require you to unbend your mind, but you will be amply rewarded by the insights you gain. When you meet an obstacle and get stuck, seek help—there are other travelers on this journey and many will lend a helping hand. Most importantly, enjoy the journey of discovery and bask in the satisfaction of reaching your journey’s end.
Footnotes
Acknowledgments
Geoff Wong was the main keynote speaker at the International Institute of Qualitative Methodology’s (IIQM) 21st Qualitative Health Research Conference, Toronto, Canada October 2015.