Center for Human Rights Science Affiliates discuss their work at Joint Statistical Meeting in Boston, MA, August 4-7, 2014-Center for Human Rights Science - Carnegie Mellon University

Center for Human Rights Science Affiliates discuss their work at Joint Statistical Meeting in Boston, MA, August 4-7, 2014

Tuesday, August 5, 2014: Estimating Undocumented Deaths During the Syrian Conflict

Patrick Ball and Anita Gohdes and Megan Price

Abstract: The conflict in Syria has been tremendously well-documented, yet despite this, we still do not know how many people have been killed from conflict-related violence. This talk will present work from a collaboration with the United Nations Office of the High Commissioner for Human Rights evaluating multiple sources of records of documented deaths during the Syrian conflict. We will also present recent capture-recapture analyses aimed at estimating the total number of conflict-related deaths, including those that currently remain undocumented. Preliminary estimates indicate that including undocumented cases may as much as double the current death toll.

Thursday August 7, 2014: Detecting Duplicate Homicide Records Using a Bayesian Partitioning Model

Mauricio Sadinle

Abstract: Finding duplicates in homicide registries is an important step in keeping an accurate account of lethal violence. The task of finding duplicate records in a datafile can be postulated as partitioning the file into groups of coreferent records, where two records are called coreferent if both refer to the same entity. Traditional approaches to duplicate detection output independent decisions on the coreference status of each pair of records, which often leads to non-transitive decisions that have to be solved in some ad-hoc fashion. We present an approach that targets the partition of the file as the parameter of interest, thereby ensuring transitive decisions. Our Bayesian implementation allows us to incorporate prior information into the duplicate detection process, which is specially useful when no training data are available, and also provides a proper account of the uncertainty of the duplicate detection decisions. We present an application of this methodology to the detection of killings that were reported multiple times to the United Nations Truth Commission for El Salvador.

August 7, 2014: Special Session: Human Rights Violations: How Do We Begin Counting the Dead?


Abstract: Violent conflicts, such as in Kosovo, El Savador, and Syria, are not just controversial in the media; understanding the underlying causes and conflicts falls to statisticians, human rights researchers, and historians who are all asking the same question “how do we count the dead?”

As recently described by Maggie Koerth-Baker “How do you find 40,000 extra dead bodies? How do you even start to determine which groups killed which people at a time when everybody with a gun seemed to be shooting civilians?” Koerth-Baker describes a complicated civil war in Peru, from 1980 and 2000, where the aftermath of counting the dead was left to statisticians, among them Daniel Manrique-Vallier, who researched deaths for Peru’s Truth and Reconciliation Commission. Their results, very different from the initial report, showed evidence of nearly 70,000 deaths, 30 percent of which could be attributed to the Peruvian government.
There are new challenges regarding ongoing conflicts, such as in Syria, where death counts come in daily. Press and media releases come weekly, begging the question all around of how do we deal with such demands and challenges regarding record linkage in near real time. In this session, we address past and present challenges to this effect in estimating death counts and associated error rates. As a core to this work, CMU and HRDAG are working collaboratively to address these issues in Syria.  This session gave an overview of the collaboration and presents some of the findings coming out of it.

  • Large-Scale Clustering Approaches for Identifying Unique Human Rights Violations — Samuel Ventura, Carnegie Mellon
  • Bayesian Multiple-Recapture Estimation of Casualties in Armed Conflicts Using Nonparametric Mixtures — Daniel Manrique-Vallier, Indiana University (formerly of Carnegie Mellon)
  • Discussant: Jay D. Aronson, Carnegie Mellon
  • Discussant: Patrick Ball, Human Rights Data Analysis Group
  • Discussant: William Winkler, U.S. Census Bureau