Just to continue this summary.
Majolein responded to the comments raised by Mark and Gill in the thread 'Reactions fo messages about realist reviews retention studies'. (8th July)
Discussion about reaction and response being part of what a mechanism was ensued and it was suggested that REASONING might be a better word as "... reasoning' implies more about the process of decision-making than 'response' or 'reaction'..."
Caution was raised that it was easy to confuse the bondary between mechanism and context. This may be because something we conceptualise as a mechanism may become a context. An example I came across whilst working with another review team was that 'trust' could be a mechanism (a person may trust another and so listens to them) but the next time they met, the previous encounter they had would have build up a degree of trust (now a context).
A question that came up was how does one best explain what a mechanism is.
The proposition put forward was that mechanisms operated at different levels of abstraction and at each different level of abstraction manisfested itself in different ways. So if a mechanism was 'that which causes outcomes' (see previous message) then what the nature of 'that' was in any one instance depends on the level of abstraction the outcome of interest resided in.
A more detailed (and better :-) explanation may be found in Gill's message of 11th July.
Unanswered questions remain about the usefulness in realist research of different ways of explaning what a mechanism is and also different ways of conceptualising and describing it.
My suspicion is that an answer may come with the explanatory usefulness of conceptualising a mechanism in a particular way. So when undertaking a review, it may be that we need to be prepared to change how we conceptualise what a mechanism is in our review and pick but define the conceptualisation that best explains the patterns we can see and are interested in explaining in the data?
Geoff
|