Since my project will investigate the effect of nationalist protest on foreign policy decision making, my dependent variable will be the amount of nationalist protest activity (which, depending on the data available, I will hopefully end up operationalizing as the number of people who took place in protests). My independent variable will be the outcome of the international crisis that sparked the protests. It seems that the easiest part of operationalizing the variables involved in my research proposal is the dependent variable. I have chosen to operationalize this as the method of crisis resolution, which the International Crisis Behavior Project (ICBP) provides as nominal-level data (resolution via negotiation, resolution via nonmilitary pressure, resolution via violence, etc.) for 475 cases of international crises spanning from 1918-2015.(1) This data set is potentially very helpful because it codifies the involvement of great powers in each international crisis as an ordinal-level variable.(2) Thus, we can control for this variable that Trachtenberg painstakingly avoids in his case study analysis because it can hugely affect crisis outcome in terms of escalation or de-escalation.(3) Unfortunately, the amount and type of confrontational events that the ICBP has collected is so broad that often there is not a clear and neat confrontation with two or more sides having clearly defined and conflicting interests, suggesting that some events will need to be filtered out.(4)
Essentially, the main problem that I am encountering at this point is operationalizing my main dependent variable in a valid and meaningful way. It has been extremely difficult to gather comprehensive data on worldwide protests events. The World Values survey has a dataset that includes the willingness of a person in a given country to participate in a protest (through petition, boycott, or demonstration — each is measured) in a given year and the number of people who have recently participated in various forms of protest.(5) Unfortunately, this survey data does not say anything about the actual number of protests that year because it was performed once every several years — and even if it did, it does not distinguish between protest on foreign or domestic issues.(6) An ideal dataset would include the number of anti-foreign protests in a given country-year; or would have data for each of many major protest events worldwide (in such a way that I could parse out the ones motivated by anti-foreign sentiment in the runup to each of the international crises), but finding such a dataset is proving to be a challenge.
- Michael Brecher, Jonathan Wilkenfeld, Kyle Beardsley, Patrick James, and David Quinn. International Crisis Behavior Project data on conditions and outcomes of international crises 1918-2015. August 23, 2017. Raw data.
- Ibid.
- Trachtenberg, Marc. “Audience Costs: An Historical Analysis.” Security Studies 21, no. 1 (January 1, 2012): 3–42. https://doi.org/10.1080/09636412.2012.650590.
- Brecher et al., International Crisis Behavior Project data on conditions and outcomes of international crises
- Chai, Jingjing, and Jie Yan. World Values Survey data on the percentage of people in China who have participated or are willing to participate in various forms of protest. April 18, 2015. Raw data. China, Beijing.
- Ibid.
Jack — the datasets that you identify here (ICB and World Values) are clearly relevant to your research. As we’ve seen with class examples and in our readings so far, it is rare that one finds a complete dataset that has all of the information/coverage needed for a particular variable (and you of course note the limitations on the data sources that you discuss here). Some original data collection may be needed as well as, perhaps, some shifting of the RQ to account for data availability/limitations while still maximizing the leverage offered by large-n analysis.
Note that there does seem to be some confusion between IVs and the DV in your post that you’ll want to think about. If you are proposing to analyze the “the effect of nationalist protest on foreign policy decision making” then, by definition, “foreign policy making decision making” is the DV (the outcome that is shaped or impacted by other factors, such as “nationalist protest.” Make sure to work on sorting out the DV from the potential IVs as you work towards the first research design sketch and keep thinking about the additional data sources that you might examine to build your own dataset for your DV.