Replication Data for: Are Things Really Getting Better?: How To Validate Latent Variable Models of Human Rights

Has respect for human rights improved? Answering this question requires valid comparisons of repression levels in many different states across time. Theoretically informed latent variable models of difficult-to-measure concepts such as human rights and treaty compliance tell a different empirical story of patterns of repression and reform than models that do not employ the same methodology. The use of this methodology is controversial for this reason. This paper addresses several critiques of the use of these models by addressing the usefulness of the latent variable model parameters in deviant case analysis, by showing empirical support for the argument that human rights are actually improving on average in both democratic and non-democratic states, and finally, by demonstrating that new data sources further support the findings in Fariss (2014, 2015).