Thanks Eleanor. Realism is a broad methodological church. For me it's core is an acceptance of the reality of causal mechanisms, independent of the events that they generate. So if you conceive your natural experiment as involving an attempt to understand the generative mechanisms entailed in the intervention, it's natural and social context, as well as those possessed by the actors involved, in addition to measuring its impact, then a realist approach could well be fruitful.

Best wishes, Sam

> On 22 May 2017, at 18:17, Eleanor Grieve <[log in to unmask]> wrote:
>
> Thanks for raising this as I've always wondered how to reconcile establishing 'impact' or 'attribution' with a realist approach ie the outcome less an estimate of what would have happened anyway.
>
> I am not trying to set up a 'realist RCT' as randomisation is unfeasible by the nature of our intervention - but I am proposing drawing on methods used in natural experiments to 'model a counterfactual' and then using a realist approach to provide explanatory power for the outcomes.
>
> Very happy to be put right if this is not congruent to realist thought!
>
> Eleanor
>
> Eleanor Grieve
> Research Associate
>
> Tel:  +44(0)141 330 4510
> Fax: +44(0)141 330 5018
> [log in to unmask]
>
> Institute of Health and Wellbeing
> College of Medical, Veterinary and Life Sciences
>
> Health Economics and Health Technology Assessment
> University of Glasgow
> 1 Lilybank Gardens
> Glasgow
> G12 8RZ
>
> http://www.gla.ac.uk/hehta
>
> -----Original Message-----
> From: Realist and Meta-narrative Evidence Synthesis: Evolving Standards [mailto:[log in to unmask]] On Behalf Of Sam Porter
> Sent: 22 May 2017 12:17
> To: [log in to unmask]
> Subject: Re: Establishing the counter factual situation in realistic evaluation?
>
> Thanks Rasmus, good question about a contentious subject. I think it is possible but only if you can satisfy yourself that you have sufficiently controlled the context, so if all goes well, the control group is subject to the same sorts of contextual mechanisms as the experimental group. This means that you can be reasonably confident that any differences in outcome are the result of mechanisms within the intervention (though that begs the question of which specific mechanisms are making the difference, which is where RE strategies come in). I think we would need to have more details about precisely what it is you are planning to do before making a call on its feasibility.
>
> This opinion is of course predicated on my acceptance of the possibility of realist RCTs. If you reject that possibility, then I think it would be hard to construct an argument for using similar strategies.
>
> Hope this is of some help, Sam
>
> On 22 May 2017, at 09:59, Rasmus Ravn <[log in to unmask]<mailto:[log in to unmask]>> wrote:
>
> Dear all.
>
> For some time I have been reflecting on a rather simple question: Is it feasible to establish the counterfactual situation as part of a realistic evaluation?
>
> To elaborate, my question concerns whether it is in accordance with the principles of realist evaluation (primarily generative causality) to use control groups (established either by randomization or through statistical matching).
>
> Reading through the realist literature, my own impression is that factuals are compared in realistic evaluations (through inter- and intra-program variation)  and that the counterfactual situation is not established.
>
> I am aware of the discussion that followed the paper by Jamel et al. (2015) "The three stages of building and testing mid-level theories in a realist RCT: a theoretical and methodological case-example".
>
> There are of course differences of opinions but I cannot help by wonder if the critique put forward of the "realist RCT" would also extend to any type of evaluation that tries to establish the counterfactual situation?
>
> My initial thoughts on the subject would be that the critique would apply to every type of evaluation that establishes the counterfactual situation, because these evaluation approaches tries to "imitate" the RCT.
>
> One of the arguments against using the counterfactual situation in realistic evaluation could be that you cannot randomly or statistically be assigned to receive a mechanism.
>
> I am hoping some you might enlighten me.
>
> Kind regards,
>
> Rasmus Ravn, PhD Student, Aalborg University, Denmark BU is a Disability Confident Employer and has signed up to the Mindful Employer charter. Information about the accessibility of University buildings can be found on the BU DisabledGo webpages. This email is intended only for the person to whom it is addressed and may contain confidential information. If you have received this email in error, please notify the sender and delete this email, which must not be copied, distributed or disclosed to any other person. Any views or opinions presented are solely those of the author and do not necessarily represent those of Bournemouth University or its subsidiary companies. Nor can any contract be formed on behalf of the University or its subsidiary companies via email.
BU is a Disability Confident Employer and has signed up to the Mindful Employer charter. Information about the accessibility of University buildings can be found on the BU DisabledGo webpages. This email is intended only for the person to whom it is addressed and may contain confidential information. If you have received this email in error, please notify the sender and delete this email, which must not be copied, distributed or disclosed to any other person. Any views or opinions presented are solely those of the author and do not necessarily represent those of Bournemouth University or its subsidiary companies. Nor can any contract be formed on behalf of the University or its subsidiary companies via email.