In the afternoon, I go to listen to Marie-Helene Pare on ‘ Convergence Strategies in Triangulation Design: A Review of Current Practices’.
She talks about how triangulation is overrated and misused in MMR. People often cite and state that they are doing triangulation as a main rationale but actually they are doing a whole set of different things. My thoughts: No wonder triangulation is being one of the most widely known and cited terms in social science methodological debate!
She mentions a subsequent review of 600 MM studies found 2/3 of the studies mentioned using a concurrent triangulation design. With so many published studies reporting to be doing the triangulation, we really need a solid and sound sense of what we mean by that. Precisely for this reason, Marie-Helene is trying to present some clarity in how the triangulation works in a research process. She looks at how the data integration is achieved when different datasets/findings converge or diverge during the research process. The objective of her contribution is to identify when data integration occurs during the research (the exact timing), to describe how data integration is achieved and … oh, I forgot the last one:>. She then presents the results of analysis of couple of papers. The objectives for triangulation are in MM are usually corroboration, dissonance or fuller picture of the results. Uff, fuller picture as on objective of triangulation? Here I would say we have to do with complementarity not triangulation. She concludes with the observation that there is a pressing need for systematic reporting about data integration in MM studies calming to have used ‘triangulation’ as their primary purpose, or design. I strongly agree. I believe that true triangulation is today rather rare motivation for combining qualitative and quantitative methods. The reason David Morgan offers for this is that most applied research simply cannot put so much effort into finding out the same thing twice.