Application and Interpretation of Multiple Systems Estimation Methods in Human Rights Research

• Activity Number: 499
• Type: Topic Contributed
• Date/Time: Wednesday, August 3, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Social Statistics Section
• Author(s): Megan Price*+ and Daniel Manrique and Anita Gohdes
• Companies: Benetech and Duke University and Humboldt University
• Keywords: multiple systems estimation ; capture-recapture ; human rights

Abstract

Multiple systems estimation (MSE) methods are often introduced with the classic two-system estimator, using a motivating example from ecology. This estimator relies on four strong assumptions, which are typically then explored and adjusted for with the generalization to three or more systems and a variety of MSE methods and estimators. We find that the implications of violating these four classic assumptions, in particular the homogeneity and independence assumptions, do not translate clearly to the more-than-three-systems case. We use case studies from human rights research to present the subtle ways that misunderstanding these assumptions can lead to misapplication and misinterpretation of MSE methods. Additionally, we examine how MSE methods rely on patterns of inclusion or capture to represent the underlying population of interest even when the individual samples themselves are not representative.