Dear Group,

Long time appreciative follower - first time writer.

I am working on my PhD which is a realist evaluation of a community health committee intervention in Uganda. This was done through an NGO health programme, and as such some logistical constraints limited the iterative case study components. Instead, I conducted case studies consecutively. Thus, for my analysis each case study was analysed separately. I used NVivo and followed the lead of Dalkin and Forster's Blog (link at bottom), which was extremely helpful. This resulted in refined programme theories specific to each case (each their own results chapter).

I then aimed to synthesis findings across the case studies, which is what my question is around. I can't find much out there on this, and I was wondering if anyone has experience or can offer advice. I took the following steps to synthesize across my case studies, to look for theories that are of middle range:

• Step 1:  I looked for for commonalities or demi-regularities across my PTs from the case studies by comparing the resulting PTs and their associated CMOCs. I didn't use NVivo for this though, as when I first attempted it proved more difficult (less interactive and informative) than lots of colourful pens and paper. 
  
• Step 2: The theories from Step 1 were then re-examined with additional data. I had 5 additional Key Informant Interviews that were specific to the whole intervention, not just individual case studies. I used these to support/refine the theories at this stage. I didn't 'throw out' any theories based on KII info, they were used more to refine, and draw further associations between the resulting PTs as the KIIs had a strong 'systems thinking' view of the intervention.
  
• Step 3: Situated the theories within the literature. Again, this didn't work to exclude any theories (or even refine that much), but related them back to 'formal theories' for additional support. This section, in the 'synthesis chapter', would be similar to part of the discussion chapter of a PhD I believe. The end of this process produced the theories of middle range for the study.
  

I am hoping that the process I followed sounds appropriate to the group or if anyone has any suggestions for improvement?

Thank you all for the time, and in advance for your considerations!

Kind regards,

Brynne

http://www.qsrinternational.com/blog/how-researchers-use-nvivo-to-enhance-transparency

--
Brynne Gilmore, BSc MSc
Centre for Global Health
7-9 Leinster St. South
Dublin 2, Ireland
(o) +353 (0)18964373
(m) +353 868468435
Skype: brynne.gilmore