Abstract
Publishing quantitative research with undergraduate students presents unique challenges and opportunities. In this case study, I discuss how a team of students and I worked together to develop a research question on post-conflict justice mechanisms in the aftermath of civil war. Then, I highlight some of the data and methodological challenges we encountered along the way. In particular, the case study looks at the complexity of analyzing six types of post-conflict justice—truth commissions, trials, amnesty, reparations, purges, and exiles—in a single analysis. Given that 64 post-conflict justice combinations are possible, our biggest challenge was to collapse the data down to a statistically and substantively meaningful number of outcomes. We accomplished this by recoding our six types into two categories: restorative justice and retributive justice. Combined with the absence of a post-conflict justice, we analyzed our outcomes using a multinomial logistic regression model. We also explored a linear combination in an ordinal logistic regression model. Last, I discuss other data reduction techniques, including principal components analysis and factor analysis.
Logit and probit models