Mathematical modelling and ‘ethnic conflict’ in Colombia: the Impact of the Unit and the Level of Analysis

Reflecting on methodology means thinking about decisions a researcher makes when approaching the subject of study, in terms of questions posed, concepts embraced and choice of methods. There are also epistemological and ontological assumptions underlying the selection of particular methodologies and concepts, though this is less often explicitly considered. Therefore within these considerations, the definition of the unit of analysis and the levels at which the study is conducted, need to be made explicit and to be justified, for quality in research. This chapter discusses these issues taking the example of Colombia, to show how some violent conflicts fall between the cracks of currently dominant methodologies of mathematical modeling and the use of existing econometric datasets, on the one hand, and the ways ethnicity is included in those models and datasets. Ethnic identities have mostly been the focus of studies by anthropologists, often focused on local level struggle sand claims, whereas a national level perspective is still most common in the econometric rational choice approach to conflict studies. The later however use ethnicity as one of its variables. My point is that research results also change when the research focus changes from national to local level.