A not-uncommon research strategy in health and social care research is to generate different types of data and, through some process of transformation, bring these together into a coherent whole. The idea here is that combining data produces a more complete, detailed, analysis than can be created using one type of data alone. For example, in my doctorate, which focused on the system of mental health care and the division of labour, I conducted lots of qualitative interviews but also used written records as a source of data and observed people going about their day-to-day work. What people say, what people do, and what people write about they’ve done are not the same thing: knitting together a rich, or ‘thick’, description of a social setting is helped when different classes of data are available to be drawn upon. In more recent studies of care planning and coordination (see here and here) the research teams I’ve been a part of have variously combined interviews, documentary review, questionnaires and observations.
In a slow-burn kind of way, over a period of many months I’ve been working with members of the 3MDR project team to bring together data of very different types. The 3MDR study, led by Jon Bisson, is something I’ve written about before and involved examining the efficacy of a novel intervention for people with post-traumatic stress disorder. Across the project overall three, distinct, classes of data were generated: outcomes, derived from clinician-assessed and self-reported standardised measures; psychophysiological, including breathing and heart rate, walking pace, words and phrases used by participants during therapy, plus subjective unit of distress scores; and qualitative, namely post-therapy interviews where people talked about their views and experiences. Working particularly closely, in the first instance, with Robert van Deursen and Kali Barawi our task has been a mixed-methods data synthesis to explore the interrelationships between people, interventions and context and to investigate how factors within these three domains interact in specific outcome typologies.
This has been an interesting and challenging project, and we’re not yet done. Whilst many of the ideas underpinning this analysis are familiar ones (complexity, interconnections, the search for patterns) the combined dataset we’re mixing together is an unusual one. This work is also proving to be a reminder of how much can be found out through the detailed study of relatively small numbers of participants. Our data relate to ten people only, but our total dataset is both comprehensive and varied. At some point (but not quite yet) we’ll have a paper ready for journal submission, and I’ll be able to share more on this site.Follow @benhannigan
2 thoughts on “Synthesising data”
Very interesting, Ben. What a complex mix of data.
It really is, Alan: beyond anything I’ve had experience of working with before.