Research Portfolio Post #6: Quantitative Data Sources

I am proposing to research intraethnic conflict because I want to find out what explains the different rates of violence in intraethnic conflict in order to help my reader better understand how intraethnic conflict might be mitigated or enhanced by different political, social, and economic factors.

To posit my puzzle as a large-n statistical question: What explains the variation of violence, or lack thereof, in intra-ethnic conflict?

While his article does not include a large-n data set, Brown creates a theoretical framework that can easilybe applied to a dataset. [1] Brown lists 12 possible factors that may create interethnic conflict: weak states, intra-state security concerns, ethnic geography, discriminatory political institutions, exclusionary national ideologies, inter-group politics, elite politics, economic problems, discriminatory economic systems, economic development and modernization, patterns of cultural discrimination, and problematic group histories.[2] In putting these factors within the context of a large-n statistical analysis, I would make these factors nominal independent variables. In creating my own data set I would get rid of “ethnic geography” since within intraethnic conflict there are no “ethnic” differences. I would also most likely make some variables more operational such as the degree to which there is economic development, not if it’s there or not. I would probably do the same with “weak states,” Another nominal variable I would add is external involvement in the area and break it down by economic, military, and political aid or intervention.

The dependent variable is intraethnic conflict occurring. To operationalize the variable, I would make it nominal by marking 1, 2, or 3 to denote non-violent conflict, semi-violent conflict, and violent conflict in each of my cases. I would then analyze patterns of the independent variables to see how the dependent variable is effected. Something I’d be particularly interested in analyzing is if there is an overbearing factor in every violent conflict, or if it’s rather that violence is a result only when there are x amount of factors involved.

The coverage of this dataset would include different intraethnic conflicts around the world, no matter the region in which they are taking place. The limitations of the data set is that with so many independent variables, there might not be data on every single variable for every case.

I’m excited to explore my puzzle from the neo-positivist lens for I have only been thinking about it from the interpretivist perspective.

 

Footnotes:

  1. Michael E. Brown. “The Causes of Internal Conflict” in Nationalism and Ethnic Conflict: Revised Edition, eds. Michael E. Brown, Owen R. Cote, Jr., Sean M. Lynn-Jones, & Steven E. Miller (Cambridge: The MIT Press, 2001), 5-13.
  2. Ibid, 5.

Bibliography

Brown, Michael E. “The Causes of Internal Conflict” in Nationalism and Ethnic Conflict: Revised Edition, eds. Michael E. Brown, Owen R. Cote, Jr., Sean M. Lynn-Jones, & Steven E. Miller, 3-25. Cambridge: The MIT Press, 2001.

2 thoughts to “Research Portfolio Post #6: Quantitative Data Sources”

  1. Hey Ben,
    I like how much depth you give in describing how you would assemble your dataset. Your examination of the variables, informed of course by scholarship is especially insightful. When you discuss that “Brown lists 12 possible factors that may create interethnic conflict…” (and then go on to explicitly name them), I’m wondering if you have sources of datasets in mind where you might find the information or “raw” data for the variables which you have already given much thought to. For instance, you make it a point to note that you seek to operationalize one of these variables as “the degree to which there is economic development, not if it’s there or not”. It seems to me that you can find a unit of measurement, say GDP per capita or really any other indicator that would make the most sense (a choice you would then justify). The information for this example seems like something you can gather from the World Bank’s datasets.

    In keeping with that example, this use of indicator comes with currency (say US dollars) which can serve as a useful unit of measurement/provide you with interval numbers. If you could do the same to the use of violence in intra-ethnic conflict somehow, that would allow you to use other statistical tools. Instead of the nominal “markings 1, 2, or 3 to denote non-violent conflict, semi-violent conflict, and violent conflict in each of [your] cases”, you could use fatalities or injuries or any other marker. If you manage your dependent variable into an ordinal or interval number, that might be useful!

    Finally, similar to Edelstein’s definition of occupation versus annexations or invasions, the way you decide to define what qualifies as intra-ethnic conflict (again, related to your choices and the literature) can inform your case selection. For my case, I’m thinking through what qualifies as a “humanitarian crisis” as it will define what cases I could study, and of course, if I have a sufficient number to conduct large-n study on the topic. I’m wondering how you would use the definition of intra-ethnic conflict to this end.

  2. Ben — even though creating your own dataset using the concepts from the Brown article could be an option (albeit a very time intensive option) for this methodology, Mohammad makes a very important point in his comments: you still need to identify the specific places where you would would find the actual data. That is one of the key first steps in operationalizing your DV (and this, in turn, is an essential first step for the large-n methodology). We have a whole range of datasets in our Bender Library listings that cover conflict, violence, and the like. Which of these would you use to collect and organize the raw data for your DV? As you think about this, remember that we really need an interval/ratio DV to maximize the power of this methodology. How could you operationalize the idea of “violence” in this way? Are you thinking of levels of violence? Number of incidents? Some other way to capture the concept.

Leave a Reply

Your email address will not be published. Required fields are marked *