Triangulation and integration : Processes, claims and implications

Researchers who advocate the use of multiple methods often write interchangeably about 'integrating', 'combining' and 'mixing' methods, sometimes eliding these descriptors with 'triangulation', which itself encompasses several meanings. In this article we argue that such an elision is problematic since it obscures the difference between (a) the processes by which methods (or data) are brought into relationship with each other (combined, integrated, mixed) and (b) the claims made for the epistemological status of the resulting knowledge. Drawing on the literature for examples, we set out different rationales for using more than one method, then we develop a definition of integration of methods as a specific kind of relationship among methods. We also discuss different places in the research process where integration can occur: for instance, data from different sources can be integrated in the analysis stage, or findings from different sources at the point of theorizing.