My thanks to those who contributed to this thread and also to those who gave us the example.

A number of interesting methodological points came to inlight in this thread. Feel free to comment on any of these :-)

MECHANISMS
This one never goes away. "What is it and where does it reside?" are probably the two most common questions asked about it!
A suggestion made included:

- people's reactions to programmes
- "...the word mechanism is that it’s shorthand for “that which causes the outcome”"
- "...internal to people that is caused to 'fire' (in certain contexts..) by the programme."
- "...“the inherent powers and liabilities of things” (as per Sayer)..."
- "...“reasoning and resources” (as per Pawson and Tilley)..."
- "...‘Mechanisms existing whether they’re fired or not’ ..."

Perhaps mechanisms are all of these and more and that the word is nothing more than a collective noun. Just as 'car' (a self-propelled road vehicle designed to carry passengers) is a collective noun with lots of variant and types within it, a mechanism is a concept designed to generate outcomes (sorry for adding another  definition)??
If in a realist world mechanisms are everywhere, then there are bound to be different types and which ones we look at will depend on which outcomes we are interested in?
Maybe what we should be expecting is not necessarily an upfront definition of a mechanism, but one that comes later - for example, outcome X was generated by mechanism Y and this is one of those internal in people mechanisms?

INTERPRETATIONS
Whilst in the review the team were cautious in the extent of interpretations they made, there was much agreement that this is the value that realist synthesis (RS) adds - "...for complex (social) inquiry where the goal is building explanations (rather than 'presenting facts') then interpretation is vital."
An additional point raise was that it was then vital to check any interpretations made not only against the included studies but also the "...broader / less direct literature." especially if the includes studies did not contain enough of the relevant detail.

MIDDLE-RANGE THEORY
A comment was made that policy and decision makers may be less interested in middle-range theory (MRT), but there was agreement that its use in RS was vital. RS seeks to explain and one device it uses is the MRT. If you are interested, within this thread were examples of how different 'candidate' MRTs might have been used to make sense of the findings in the example review.

INITIAL STAGES OF A RS
At the start of a RS, the decisions made could have important consequences. Some of these decisions may be beyond the control of the review team, but the ideal scenario would be to be able to influence the following:
- frame the question so it can be answered using RS
- set up conditions so that a question can be answered (e.g. sufficient resources and time available, appropriate focus). An example drawn from this thread was that if the focus of a RS was set to include only randomised controlled trials, then this might have limited the review process.

Once again, I hope this summary helps rather than hinders and do feel free to comment.


Geoff