Uncertainty Assessment in Multiple Systems Estimation

Multiple systems estimation (MSE) methods have been developed over the past century in fields as diverse as ecology, demography, epidemiology, and human rights. This has resulted in a broad class of estimation methods which vary in terms of assumptions, model selection criteria, and uncertainty assessment. This talk will focus on multiple potential sources of uncertainty, such as matching and model selection, and the varying methods developed to account for these multiple sources. Examples of estimation methods include bootstrapping results across multiple matched datasets and Bayesian model averaging.  Case studies from human rights research will be presented to ground the discussion in tangible examples.