All
Interesting that it always 'mechanisms' that create debate on RAMESES. I guess it is because the term evokes some fundamental atom of explanation. Here's a bit of good news - 'evidence-based mechanistic reasoning' has reached the clinical sciences. See a paper by Jeremy Howick et al under that title in J R Soc Med. There's also a book by the same fella and well as a growing number of other contributions. They have inspired me to have a closer look at the parallels between EBP and EBM.
May the debate continue. Hope to see many of you in Liverpool.
RAY
________________________________________
From: Realist and Meta-narrative Evidence Synthesis: Evolving Standards [[log in to unmask]] On Behalf Of Nick Emmel [[log in to unmask]]
Sent: Friday, July 11, 2014 10:48 AM
To: [log in to unmask]
Subject: Re: A new participant question regarding mechanisms
Dear Barbara (and all)
Below, an extract from my recent book, ‘Sampling and choosing cases in qualitative research: a realist approach, in which I discuss generative mechanisms in the context of sampling strategies in qualitative research, which I hope will add another dimension to this debate, which Geoff has so rightly characterised as a reoccurring challenge for us all.
Generative mechanisms in sampling
For realists, generative mechanisms govern, mediate, impede, and facilitate sampling choices. These are the powers that describe sampling. The ways in which these causal powers are implicated in explanations and the ways in which they differ from events and experiences are often explained by realists through considering the capacity of gunpowder to explode. In part, this is accounted for through its chemical composition of potassium nitrate, charcoal, and sulphur. But, as Sayer (1981) reminds us, it only has this capacity to produce an exothermic reaction and a loud bang if certain powers are applied to the gunpowder. To explode it needs the right kind of spark, under the right atmospheric conditions; it must be packed to the correct density, and so on. But it needs more than just the right chemical and physical conditions – social conditions are needed too. To use a parochial example, which will be more familiar to readers in England: if, as happened on the night of the 5th November 1605, Guy Fawkes with matches and touchwood in his pocket and gunpowder ready and primed, was arrested, then the gunpowder’s potential to explode and blow up the British House of Lords will not be realised. Gunpowder has the determined capacity to explode, but this is not pre-determined. To account for when and where it will explode we must explain the mechanisms, or causal relations and powers inherent in the system. The failure of the Gunpowder Plot can only be explained if we account for the substances in the system; the chemical (gunpowder), the agents (Fawkes and his associates), an anonymous tip-off to the 4th Baron Monteagle, and the structures (King James I and the guards he sent to search the cellars of Parliament, and who subsequently arrested Fawkes).
Every English child knows the story of Guy Fawkes and his failed attempt to blow up the English Parliament. We persist in burning Fawkes’ effigy on the 5th of November and sing the nursery rhyme ‘Remember, remember, the 5th of November, gunpowder, treason, and plot … ’; and for those who are unfamiliar with this story, substitute any recent act of terrorism, most likely planned with more sophisticated explosives. The underlying realist account remains the same. An explosive outcome is not pre-determined, but determined by powers and liabilities. These physical, social, and structural mechanisms all have to be disclosed for the explosion to happen. As with the disclosure of mechanisms to realise the explosive potential of gunpowder, sampling in qualitative research discloses a range of powers or liabilities as mechanisms for it to happen. Some of these mechanisms are external powers to the research and others internal. In identifying these we are able to elaborate on how these powers inform and control choices about who or what we want and are able to sample in the research.
It is tempting to draw on a metaphor of Russian dolls to explain this disclosure, in which the various physical, social, and structural mechanisms nest within the others to realise a particular sample in research. But the traditional matryoshka doll, where each is a scaled up version of its smaller sibling does not quite work as an adequate metaphor. More modern matryoshka dolls, such as those readily available in the markets behind Red Square, which nest a doll of Lenin, within a representation of Stalin, and so on all the way to Vladimir Putin, better explain the internal and external powers and liabilities that must be disclosed to interpret and explain social process and social objects in realist research. Vladimir Vladimirovich Putin is Vladimir Ilyich Lenin’s heir, even though many of the powers and dispositions that connect them are certainly not visible. Similarly, researchers are asked to explain the powers, dispositions, and liabilities that frame the choice of cases in research. Researchers can never be fully reflexive of course, an issue discussed in the conclusion to this chapter, but their challenge is to disclose the external and internal powers of sampling in the research as best they can.
External and internal powers in sampling
External powers are contingent to sampling. They are tangential, touching upon the sampling and the research, exerting their liabilities in particular ways, and shaping or defining the possibilities for a sample to be chosen or controlled by researchers. The institutions for whom we do research – universities, funding bodies, research governance structures, ethical review bodies, and so on – have powers to define the kinds of research that are possible, the research topics we may investigate, and the ways in which these are investigated. These mechanisms govern the kinds of research that can be done, and therefore the research questions we may ask. Subsequently, the samples we may choose to answer these research questions are described through powers and liabilities. Our sample is an outcome contingent upon external relations, which are embodied in the institutions for whom we do research.
In a similar way, a gatekeeper in the research has the determined potential power to shape who we may or may not talk with, thus shaping the possibilities for research. This determined power, is, like the explosive potential of gunpowder, not pre-determined. We may spend considerable resources developing a relationship of trust and negotiating a different outcome to the one the gatekeeper first intended. Nonetheless, gatekeepers have liabilities (and are often seen as a liability to the research), through refusing introductions or ensuring the researchers talk to particular kinds of people or access particular resources, which allow them to control the sample in particular ways.
External liabilities have a significant impact on the choices that can be made about who or what to sample in the research. These powers and liabilities can modify each other as well. That a researcher is seen as representing a particular institution could, potentially, fashion the ways in which gatekeepers respond to researchers seeking access to a sample through their mediation, for instance. That I work for a large university in a northern English city can be seen to confer credibility and legitimacy with some audiences. With others, the institution for which I work may well be perceived as a distant, irrelevant, ivory tower that does not address their needs.
Powers are relational, they are also wide ranging. Wendy Olsen (2004) describes how her association with low-caste women in a village in Andhra Pradesh, India, barred her from talking with higher caste women and their husbands, for instance. These external powers are mechanisms that allow for one sample in the research at the possible expense of a quite different one. These external powers motivate, constrain, discourage, and enable certain sorts of samples, they interpose in the choices to sample in particular ways.
The powers and liabilities of the caste system in rural India are built on ideas of touchability, untouchability, and the possibilities of transmission through a third person. In describing our sampling choices we will be interested to record these contingent experiences and events and how these frame in some way the possibilities to form a sample. There is invariably an element of constrained choice in the sampling choices we can make.
Unlike external powers, the generative mechanisms internal to the study are not contingent. These are the sets of ideas, or theories made by researchers that inform the choice of sample. Internal powers are explicitly theory laden. They have the potential to modify the sample chosen in quite significant and fundamental ways.
Dr Nick Emmel
Director of Research
School of Sociology and Social Policy
University of Leeds
Leeds
LS2 9JT
+44 (0) 113 343 6958
Twitter @NickEmmel
Blog http://realistmethods.wordpress.com/
[cid:image001.jpg@01CF9CF4.79890590]
Nick Emmel (2013) Sampling and choosing cases in qualitative research: a realist approach. London, SAGE.
Teaching and papers http://www.sociology.leeds.ac.uk/about/staff/emmel.php
Connected Lives http://www.reallifemethods.ac.uk/research/connected/
Timescapes http://www.timescapes.leeds.ac.uk<http://www.timescapes.leeds.ac.uk/>
From: Realist and Meta-narrative Evidence Synthesis: Evolving Standards [mailto:[log in to unmask]] On Behalf Of Simon Carroll
Sent: 11 July 2014 02:21
To: [log in to unmask]
Subject: Re: A new particiant question regarding mechanisms
Agree with this comment strongly Geoff...as we move towards the conference being organized in Liverpool, we need to be open and have a discussion about how we should not be too worried that a 'normal science' has not yet settled (if that type of conceptualization has validity for this approach) for RE and RS...a healthy debate and dialogue around the core concepts is a sign of openness and potential progress, rather than stagnation or dissolution....the attitude of irony and irreverence among the leaders here as opposed to high church proselytizing means that such openness can be expected to continue.
Cheers, Simon
On Thu, Jul 10, 2014 at 12:07 PM, Geoff Wong <[log in to unmask]<mailto:[log in to unmask]>> wrote:
A reoccurring challenge for all of us revolves around the issue of mechanisms- what they are, what are not and where they are located and so on.
The state of things as they are now in realist evaluation (RE) and realist review (RR) are there there is still a degree of disagreement and misunderstanding around the nature of mechanisms. As stated by Trish in another posting, we have not as yet reached what Kuhn would call 'normal science' - around this concept.
My suspicion is that this has led to a degree of variation in the application of RE and RR in published literature and from reading these, further confusion can result.
Some of you will have seen this paper already, but Marchal et al.'s paper on the promise of RE (attached) highlights this point about variations in application and understanding of the underpinning assumptions around RE.
Taking this one step further, Pawson and Manzano-Santaella's provide three examples of where improvements could have been made in published REs (also attached).
For those who want to read about mechanisms, Astbury and Leeuw's paper on mechanisms is attached (Unpacking Black Boxes in Evaluation).
Finally to pick up on Gill's helpful heuristic for analysis, there is a learning point from working backwards from the outcome. Context (C), Mechanisms(M) and Outcomes (O) are 'tied' to each other. They are not 'free-floating' C, M or O - a point raised in Pawson and Manzano-Santaella's paper as they write about "Configurations not catalogues".
Happy reading :-)
Geoff
On 10 July 2014 15:04, Trish Greenhalgh <[log in to unmask]<mailto:[log in to unmask]>> wrote:
Thanks Ceri - but didnąt we all argue about the difference between a
context and a mechanism at the time???
Trish Greenhalgh
Professor of Primary Health Care and Dean for Research Impact
Barts and the London School of Medicine and Dentistry
58 Turner St
London E1 2AB
UK
+44 20 7882 7325<tel:%2B44%2020%207882%207325>
[log in to unmask]<mailto:[log in to unmask]>
@trishgreenhalgh
On 10/07/2014 08:58, "Butler, Ceri" <[log in to unmask]<mailto:[log in to unmask]>> wrote:
>Dear all
>
>Apologies if Trish or others have already shared this article, but this
>might be of interest to some of you interested in undertaking a realist
>evaluation of a service transformation:
>
>Trisha Greenhalgh, Charlotte Humphrey, Jane Hughes, Fraser Macfarlane,
>Ceri Butler, and Ray Pawson. How Do You Modernize a Health Service? A
>Realist Evaluation of Whole-Scale Transformation in London. Milbank Q.
>Jun 2009; 87(2): 391-416.
>http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2881448/
>
>It is a few years old now but is still highly relevant, particularly
>looking at both the context and mechanisms of change in detail.
>
>Best wishes
>
>Ceri
>
>
>Ceri Butler
>
>Academic Lead, PG Programmes
>UCL Medical School
>University College London
>Room GF/664
>Royal Free Hospital
>London NW3 2PF
>
>Tel: 020 3108 9214<tel:020%203108%209214> (via Jeannine Attreed - Postgraduate Scholarship
>Administrator)
>
>UCL Medical School Postgraduate Activities:
>http://www.ucl.ac.uk/medicalschool/postgraduate/postgraduate-activities/
>
>UCL Medical School Education Consultancy:
>http://www.ucl.ac.uk/medicalschool/education-consultancy
>
>
>
>
>-----Original Message-----
>From: Realist and Meta-narrative Evidence Synthesis: Evolving Standards
>[mailto:[log in to unmask]<mailto:[log in to unmask]>] On Behalf Of Sandy Oliver
>Sent: 10 July 2014 07:10
>To: [log in to unmask]<mailto:[log in to unmask]>
>Subject: Re: A new particiant question regarding mechanisms
>
>Dear Gill
>
>Thank you very much for explaining how you identify mechanisms in a
>theory building realist review. That's helped me clarify my thinking for
>a couple of pieces of work: primary research about institutional
>mechanisms for producing policy-relevant reviews; and a review about
>effective and efficient committees.
>
>Best wishes, Sandy
>
>Sandy Oliver, PhD, Professor of Public Policy Social Science Research
>Unit and EPPI-Centre, Institute of Education, University of London.
>
>Public engagement with academic research: outsiders bring
>(a) independence for oversight
>(b) experiential knowledge for designing studies
>(c) practical and problem solving skills for data collection and
>analysis, and
>(d) an inquiring mind for research informed citizenship.
>http://bit.ly/YeT0w2 Twitter @profsandyoliver
>________________________________________
>From: Realist and Meta-narrative Evidence Synthesis: Evolving Standards
>[[log in to unmask]<mailto:[log in to unmask]>] On Behalf Of Gill Westhorp
>[[log in to unmask]<mailto:[log in to unmask]>]
>Sent: 10 July 2014 02:10
>To: [log in to unmask]<mailto:[log in to unmask]>
>Subject: Re: A new particiant question regarding mechanisms
>
>Hi all
>
>I think there are a couple of things to keep clear in this discussion.
>
>One is the distinction between realist synthesis and realist evaluation.
>It's certainly possible to do both in the same study, but they're quite
>separate things and they can require quite different processes to identify
>mechanisms. So - with thanks for the kind words Justin - the link we
>just
>posted is to a theory-building realist synthesis (rather than a realist
>evaluation). By that I mean - a realist synthesis where we could not,
>because of the terms of the contract and the stage of development of
>existing theory - posit the mechanisms in advance and investigate a
>little segment of program theory in depth. Rather, we had to begin the
>process of building realist theory from the existing literature and the
>focus was on breadth rather than depth.
>
>So my first question for you Barbara is - are you undertaking a realist
>evaluation or a realist synthesis, or both?
>
>The second thing I think we need to keep clear is the two uses of the
>term mechanism. It's unfortunate but the word was already used in much
>research and evaluation to mean 'program strategy' or 'feature of the
>intervention'.
>Realists of course mean 'underlying causal process'.
>
>That second one gets intricate for two reasons. Firstly, there are
>multiple different ways of thinking about and understanding underlying
>causal processes, of which Pawson and Tilley's 'reasoning and resources'
>is just one. Secondly, using P&T's construct, features of the
>intervention can in some cases be the resource, or provide the resource,
>in response to which the participant (or other decision-maker) reasons.
>
>Consequently, features of the intervention can concurrently be 'strategy'
>and 'context' and 'mechanism'. It all depends which particular little
>piece of the analysis you're doing at a particular moment in time. My
>mental trick for this is to start by identifying the current outcome of
>interest (which may of course be an immediate, short term, intermediate
>or long term outcome); work back from that to identify the mechanism(s)
>(it or they were underlying causes of the outcome), and then identify
>features of context that affect that mechanism.
>
>Context can of course influence outcomes in many ways. The 'purist'
>realist view is that it affects which mechanisms 'fire'. It also
>influences implementation, of course (which in turn can influence which
>mechanisms can
>fire) and it influences whether or not participants are able to put
>intended decisions into action or maintain them (implementing the
>decision may or may not be included in the description of the mechanism,
>depending on which construct of mechanism one is using).
>
>But in direct answer to your question Barbara - my perspective would be:
>yes, there are very many studies that are stronger on context than
>mechanism. However from a 'pure' realist perspective - it's a bit
>difficult to say that context has been dealt with 'well' if mechanism
>hasn't been, because the focus is supposed to be on how context affects
>mechanisms.
>
>I certainly wouldn't put it up as an exemplar of good realist practice
>but I attempted a realist evaluation of a pilot program for which the
>report is
>publically available: http://familybyfamily.org.au/ It does at least
>provide a CMOC table! Again, this was as much realist theory building as
>theory testing. We're hoping that future evaluations of this program may
>get a little further on theory testing.
>
>Cheers
>Gill
>
>-----Original Message-----
>From: Realist and Meta-narrative Evidence Synthesis: Evolving Standards
>[mailto:[log in to unmask]<mailto:[log in to unmask]>] On Behalf Of Jagosh, Justin
>Sent: Thursday, 10 July 2014 10:02 AM
>To: [log in to unmask]<mailto:[log in to unmask]>
>Subject: A new particiant question regarding mechanisms
>
>Dear Barbara,
>
>Your observation about realist evaluations emphasizing context over
>mechanism is something that I've noticed and thought about as well. There
>are indeed good examples out there. Purva's is one. Gill Westhorpe just
>posted a link to her recent realist evaluation in which the mechanisms
>are quite well defined.
>Yet there really is no set way of doing the analysis. It maybe the case
>that the research team decides that exploring contextual determinants is
>more useful than exploring mechanisms. Perhaps that is just fine -
>especially if we approach research with the realist principle that all
>research produces partial knowledge at best.
>
>Personally, my research orientation is to begin synthesis by searching
>for, and theorizing upon the mechanisms - that is, the range of intended
>and unintended resources created + stakeholder responses, for the program
>in question, or the cohort of programs. The idea here is that an
>understanding of contextual factors can then be built subsequently from
>an initial exploration into how the program works. In other words, first
>to 'break down' the program into its underlying mechanisms, and then ask
>the question:
>what contextual factors do we see impacting these mechanisms?
>
>Doing it the other way around (i.e., starting with studying context),
>might not yield the same analysis. An additional point is that
>contextual determinants can seem endless. Having a grasp of the candidate
>mechanisms can help focus and pull out the relevant aspects of context
>that would seem to make sense. The overall picture of the CMO
>configuration process can then inform the theoretical output of the
>research.
>There is no one way to tackle this, but it's what makes sense to me. I'd
>be interested in hearing other thoughts.
>
>sincerely,
>Justin
>
>
>Justin Jagosh, Ph.D
>Senior Research Fellow
>Centre for Advancement in Realist Evaluation and Synthesis (CARES)
>University of Liverpool, United Kingdom
>
>Phone:
>(in Canada)
>00-1-604-822-3814<tel:00-1-604-822-3814> (w)
>00-1-778-846-4589<tel:00-1-778-846-4589> (m)
>
>________________________________________
>From: Realist and Meta-narrative Evidence Synthesis: Evolving Standards
>[[log in to unmask]<mailto:[log in to unmask]>] on behalf of Barbara Maxwell [[log in to unmask]<mailto:[log in to unmask]>]
>Sent: July 8, 2014 23:59
>To: [log in to unmask]<mailto:[log in to unmask]>
>Subject: A new particiant question regarding mechanisms
>
>Dear fellow RAMESES participants,
>this is my first post on the list serve ( oh my, that does sound like an
>introduction at an AA meeting) and I was hoping for a little advice.
>I am engaged in a study of interprofessional clinical education for
>pre-qualification healthcare professional students and am attempting to
>conduct a realist evaluation. The design is structured around the
>application of the realist cycle. In reviewing the literature where a
>realist evaluation has been stated as being used in IPE evaluative
>studies, I have noticed that context is dealt with well in the majority
>of studies, but mechanisms are not. It appears that many studies elude to
>mechanisms but the items they identify are more characteristic of
>features of the program design, or could be argued to be context items,
>and do not describe the interaction of resources and reasoning. The
>studies also do not appear go to the stage of identifying CMOC theories.
>I wanted to ask if this a common issue in realist evaluation studies and
>ask for some recommendations for studies that the RAMESES community would
>recommend as exemplars of realist evaluation.
>
>
>Barbara Maxwell
>Professor & University Director of Interprofessional Education &
>Collaboration A. T. Still University
>5850 E Still Circle
>Mesa AZ 85142
>USA
>[log in to unmask]<mailto:[log in to unmask]>
>480 219 6109<tel:480%20219%206109>
>
>The Institute of Education: Number 1 worldwide for Education, 2014 QS
>World University Rankings
|
