Research Portfolio Post #6

Until this week during which our class’ focus shifted to large-n neo-positivist research, I had mostly considered my puzzle in terms of a handful of cases I was familiar with. I was aware, however, that torture has been used in prisons worldwide for centuries. Because I am interested in investigating the causes for the closure of each of these facilities, I knew I needed to address the duration of operation in each prison utilizing torture. Therefore, to compare a larger number of cases in a large-n research project, I would use years of operation as my dependent variable, allowing me to have an interval-ratio value to analyze statistically if needed. Ideally the nature of my independent variable will demonstrate some kind of correlation to the number of years each institution remained open.

As for finding the data for my dependent variable, I could not find any datasets including information on duration of prison operation internationally, especially not regarding prisons of torture. Given the specificity and scale of my subject matter I expected to need to search for my own data and found a few sources that I might be able to pull from. First, I found that the United Nations has a Committee Against Torture which publishes factsheets on systems of torture and the organization’s actions in response.¹ While not all of the information on torture is specifically related to prisons, I can sift through the data and pick the cases relevant to my research. The limitation of the UN site is that the information is mostly about ongoing crises and there is not necessarily a precise duration in years listed for each case. Regarding the scope, there is some measure of information on almost every country’s involvement with the committee and/or record with torture.

Another organization that has poured significant resources into investigating the persistence of detainee torture is Amnesty International. As an NGO dedicated to achieving human rights globally, Amnesty International has published numerous reports detailing the facts surrounding various torture practices globally.² This is another source I can sift through to find cases matching the parameters of my research to use for my dependent variable. The potential limitation on this data is that it is funded by a private organization, so it will be essential for me to consider the organization’s motivations and perspectives in finding their data. As for scope, Amnesty International provides a slightly broader offering of cases going further back in time and they also describe issues at the sub-state level more frequently than the United Nations.

¹ “OHCHR | Committee against Torture,” United Nations Human Rights Office of the High Commissioner, accessed October 11, 2018,
² “Torture and Prison Abuse,” accessed October 11, 2018,

5 thoughts on “Research Portfolio Post #6

  1. Tina says:

    Hi Hannah,
    We are facing the similar challenge of dredging up data that is normally classified/not readily publicized. I had a hard time operationalizing my variable and I think that it was a creative solution for you to operationalize your DV as the number of years, or duration, that a prison remained functioning (interval-ratio). I applaud your ambition to manually sift through the archives of Amnesty International and OHCHR to find relevant data for your research. If I am following you correctly, the cases in your research are the different prisons around the world. I am sure you know this already, but it is nonetheless important to note, that you will have to find a substantial (>15) prisons that have data available on them (whether it be through the UN or Amnesty). I actually did some light research and found this “World Prison Brief” dataset. I am not sure if it addresses torture but it a offers a substantial list of Prison systems around the world by region and gives some statistics.

    Maybe it will some help to you:

    Best of Luck!

  2. Hannah says:

    Hannah, given your initial research question, I think you have operationalized your dependent variable very cleverly. Regarding the data itself, how do you plan on sifting through the data? Creating a spreadsheet seems like the most practical way to gather your data, but I am sure there are other methods at your disposal. Also, how many cases do you think you can gather from the data sources you mentioned? Will this number be sufficient for large-N research? Are there too many data points to sift through on these websites? Just some questions to keep in mind as you start your large-N research design.

  3. loriyounissess says:


    Your post, as well as Tina’s comment, made me realize how lucky I was to find an unclassified database that detailed data from of over 18,000 terrorist attacks. I agree with Hannah’s comment that you might need to start a spreadsheet since you’re pulling data from various sources. I would suggest maybe talking to your mentor or Dr. Boesenecker about whether you want to gather data on 15-20 specific cases, or try to find a dataset that is more extensive. I think this will highly depend on how your research question shifts for your Large-n research sketch.

    Good Luck!

  4. Hannah — the data source you discuss here are good sources for your project (overall and in this specific methodology). As we’ve discussed in class, there is often not a single dataset with all of the information that one might need for a project, so compiling data from a variety of datasets and data sources is a fairly common step in research. Using years of operation as your DV would seem to be plausible (as long as you can establish that there is sufficient variation here so as to pose a puzzle/problem). This would mean that individual prisons are your cases, correct?

Leave a Reply

Your email address will not be published.