Skip main navigation

New offer! Get 30% off your first 2 months of Unlimited Monthly. Start your subscription for just £35.99 £24.99. New subscribers only T&Cs apply

Find out more

Further explaining of realist reviews

Dr Geoff Wong talks about generative causation, middle-range theories, program theories, and realist reviews in this video.
Welcome to the presentation. My name is Dr. Geoff Wong. I’m a clinical research fellow in the Nuffield Department of Primary Care Health Sciences at the University of Oxford, and also a practicing NHS GP. So generative causation– generative causation is pretty much the sort of– one of the most important concepts in the sort of realist philosophy of science. Ultimately, what it’s saying is that it’s not interventions. It’s not what we do that makes things happen. It’s these particular processes called mechanisms. And one way of thinking about mechanisms, or sort of subset of mechanisms, certainly within health service type programmes, one can see them as being responses that individuals might have to resources or opportunities provided to them.
Now, here’s a sort of study, weird, strange bit, perhaps, mechanisms on the whole within realist reviews are thought to be as– they are thought to be hidden. They’re sensitive to variations in context, and they generate outcomes. Those are probably the three most common properties of mechanisms that could be agreed upon by pretty much all realists. So it may seem rather strange that we’re starting to talk about hidden things, particularly, as you know, we’re all scientists of some description. But actually, this is extremely common. One of the most common types of mechanisms which fit the description, or the conceptual descriptions of being sort of hidden and sensitive to variations in context and generating an outcome, would be gravity.
No one has ever seen gravity. We can see its outcomes, what it generates, which is, for example, gravity makes an object fall to the ground, so the outcome of falling to the ground. The mechanism is gravity. And the context in this case, if it fell to the ground, would be, we are on earth. If you then change the context in this case, the thing that is functioning as context, and we’ve actually moved into the International Space Station, then the object would not fall to the ground. So in this case, you could say that the mechanism which is causing things and objects to fall, in other words, gravity, is sensitive to context.
Now, the important thing is within social sciences, which is what we’re kind of looking at in terms of these complex interventions that we are embedding and putting within the health systems we look at, the mechanisms are much, much more different. So they might be, for example, much more emotional in nature, sort of people being fearful about things, people being scared by having a particular condition, all sorts of other things where people might, for example, discount the future and say, well, you know, there’s no point in me needing to treat my blood pressure now because I’m only 40. So the mechanisms are all slightly different.
But the most important thing to bear in mind is that they still share the same conceptual properties, which is that they’re hidden, sensitive to variations in context, and that they generate outcomes. So we then move on to this idea, which I said earlier on, about what a– a middle range theory. So the important thing about theories, which Pawson and Tilley have really tried to make clear, is that theories are not very useful if you can’t test them, ultimately. And they’ve drawn on this concept of a middle-range theory by Merton. a particular aspect of it which they highlight is that middle-range theories are basically theories which are at the correct level of abstraction to be useful and testable.
Now, realist evaluation and realist review have a very specific way of expressing the middle-range theories. And very much the last concept to cover, are programme theories. Now, this isn’t a concept which is specific to realism or the type of realism Pawson and Tilley have put forward, but it’s a concept that’s come from evaluation literature. Effectively, it’s a sort of– an abstracted description in our diagram that lays out what a programme or family of programmes or intervention comprises, and how it is expected to work. So there’s kind of two parts to this. There is, this is what it consists of. And of the things of which it is consisting of, this is how the outcomes happen within them.
So those are the two separate thing. In that diagram I’ve provided there, you have these interventional strategies. So think of that yellow box in the top left-hand corner. It’s an interventional strategy. It could, for example, be free telephone advice for people who have questions about their medications. So you have to think, well, so that’s– sort of that’s where that lives. I wonder what the outcome is from that. So by providing the device, what is the expected anticipated outcome, which would be the Os. And then how in the world did that particular outcome happen? That would be what causes it, the M.
And then these are only triggered under the particular context, C1 or C2 in this case of the yellow box. And of course, these particular things have to be situated and sequenced in some way. All interventions provide sequencing. Being the classic one, I always tend to say to people is, look, if you’re not watching this particular video, then you’re not going to be learning very much from me. You might learn something but the sequencing here would be the person, individual, in order to learn something from me, would have to sign up and read this or listen to this particular presentation, and listen and watch it. Otherwise, they’re not going to learn very much. So that’s one sequencing.
Now, whether they learn subsequently anything more from me, that again, will perhaps need them to continue to watch and listen. And then ultimately, if they then are going to put things into practise, that might also require them to do, as well as practise. So that’s a sort of very simplistic sequencing, for example, of how you might get someone to become more skilled at something. So that’s a programme theory. So what actually happens in the complex intervention? And this is probably when you start seeing the sociological leanings of both Pawson and Tilley. They’re both sociologists.
So when participants take part in complex interventions, or not, as the case might be– and not is just as important as our unintended consequences– they make choices about what actions to undertake. And these choices about actions give us their outcomes. Actually, it’s patient choice. But then the important thing is that patients and people and the people we deal with, our clients, et cetera, they don’t have an infinite range of choice available to them. In fact, nobody does. For example, we’re all mortal. And this is, again, exemplified by the bottom right-hand corner with that man. He can’t really stand up because he’s stuck in a box. And the world constrains us.
Within, certainly, health service here, you can’t have any kind of medicine you want. You know you can’t be dispensed medication at a higher dosage than you were prescribed it. So we do not have an infinite range of choices. And it’s– the of range of choices is limited. And that basically forms the context. These are sort of the world around us, and there are multiple things in the world around us that may or may not function as context, or may or may not trigger in a way the mechanisms that we’re interested in to give us the outcomes that we want that constrain our choices.
And actually, the thing that should have been mirroring on and layering on, there’s such a– the realist middle-range theory of complex and mechanism that gives you an outcome, it’s the mechanisms that sort of drive these choices. And the way you can then change the options available to an individual, in effect, is you change the context. You change the things that may be functioning as context around an individual in order to enable them to have the outcome that we would want or they would want, or perhaps we would all want. Again, CMOs, the middle-range theory. So ultimately, this is, again, a diagrammatic representation of this. And starting at the left-hand side, where you see the word intervention.
There’s some degree of debate and discussion within realist research. But my position is very much that interventions manipulate context. It’s really about the only thing that we can change. Not to sound too Orwellian about it, but in effect, if you change what is around an individual, it well– it may well give them– it may well be more likely that the mechanism, the causal process for the outcome we want, gets triggered. So what you do is you manipulate the context, hoping that the mechanism that you want to be triggered is triggered, and that gives you the outcome that you want. So context influences which mechanisms will fire. And not everything is context.
So a very important thing to bear in mind is that in realism, context is something that we would assign to a particular thing. So it might be, for example, that gender is context. But it isn’t always context. So putting it all together– so one way of thinking about things is that our heads, they contain various different mechanisms. And these mechanisms are triggered by contexts. These are things in the world around individuals that trigger particular mechanisms within them. And the world is full of context. And our interventions try to change these, so that the right mechanisms are triggered. And our job, basically, as realist reviewers are to say, look, these are our outcomes. It’s often easier to work back from outcomes.
So these are the outcomes we want. What do we need to do in order to get these outcomes? And you can basically show the working with your CMO configuration. Well, we’d need to get this mechanism triggered. Wonder when those mechanisms trigger? Well, it’s under these particular contexts. Right. And then what does the literature tell us in a realist review of what kinds of interventions we can use to change these contexts? This fourth bullet point about irregularities and things that are sort of patterns in a way. So patterns of CMO configurations might happen. In other words, people tend to believe– behave in certain ways. And the most important thing is you’ve got to build your CMO configurations. So that’s your working.
That’s what you’re saying. That’s how you’re going about explaining to people, this is why I think this happens. This is the causal process under these contexts. And if you can build those middle-range theories, then you’re onto– you’ve pretty much made it. So what’s the realist review? Not going to spend too much– too much time on this. The most important is the red writing there, which is just the training materials, which you can find at the RAMESES project website. And the most important thing probably with this slide is to see that the goal of any realist review is basically to collect data, to enable you to refine a programme theory. This is a theory of implementation, and a theory of change.
And within that theory of change, that’s where your CMO configurations sit. So it’s a way of trying to– the programme theory is a way of organising and explaining phenomena. And if you look very carefully at this, it’s an iterative process. We’re kind of constantly, bit like sort of detectives, gathering more and more data to allow us to build this programme theory and to test it and to enrich it. So ultimately, here’s the closing summaries. Realist reviews, they enable reviewers to use the broad range of different data sources. And that’s the other thing, theory building isn’t just numbers. It can also be from, for example, qualitative work, even surveys, even policy documents. It can be processed evaluations of different sorts.
Just– but effectively, what we’re trying to do is gather data to build programme theories to explain context-sensitive, complex interventions. Realist reviews are based on a realist understanding of how the world works as a realist philosophy of science. And it’s just a view of how the world works that enable us to make sense of context-sensitive, complex interventions. In other words, because there’s a very explicit link between an outcome via the sort of causal process of a mechanism and triggered in particular contexts, it’s fairly clear that if you can show that working, that there’s a clear justification for the sort of explanations that you would give.
But the most important first step is just to understand that it’s a very different way of thinking about it. So if you can kind of have a grip or a sort of light grasp, even, of a realist philosophy of science, then that’s a sort of first big important step to be able to take. So the suggested readings, here are the two documents that you may wish to read. So thank you for listening. We have a JiscMail list. That’s a email list that you’re welcome to join. Basically, you ask to join on the list, or email me, and I’ll pretty much add you on. And there are some training materials, as I said, on the RAMESES project website.
And there’s a course that I also run at Oxford a couple of times a year. Thanks again for listening, and good luck with the rest of the course.

In this video, Dr Geoff Wong explains how generative causation and mechanisms contribute to middle-range and program theories.

Mechanisms are usually hidden, sensitive to variations in context, and generate outcomes; for example, gravity. In social sciences, mechanisms have the same three conceptual properties, however they are generally more emotional (such as a fear of something or discounting the future).

He explains that for social sciences, the specific way of expressing a middle-range theory is C + M = O (or, context + mechanism = outcome); a middle-range theory explains the limitations of context on mechanisms.

A program theory can be defined as an abstracted description and/or diagram that lays out what a program (or family of programmes or interventions) consists of and how it is expected to work. Overall, it organises and explains phenomena.

Complex interventions manipulate context, and this is what can change the outcome. Different mechanisms can also alter an outcome, and this can be clearly seen in a program theory. Patterns of C + M = O can occur because people tend to behave in certain ways in specific situations.

Dr Geoff Wong has suggested some further reading and a website to visit, and these can be found in the see also section below.

Please find a pdf of the PowerPoint slides in the downloads section below.

In the comments below please let us know:

  • Have you heard of realist reviews before?
This article is from the free online

Tackling Antimicrobial Resistance: A Social Science Approach

Created by
FutureLearn - Learning For Life

Reach your personal and professional goals

Unlock access to hundreds of expert online courses and degrees from top universities and educators to gain accredited qualifications and professional CV-building certificates.

Join over 18 million learners to launch, switch or build upon your career, all at your own pace, across a wide range of topic areas.

Start Learning now