Online Program Home
My Program

Abstract Details

Activity Number: 538 - SPEED: Predictive Analytics with Social/Behavioral Science Applications: Spatial Modeling, Education Assessment, Population Behavior, and the Use of Multiple Data Sources
Type: Contributed
Date/Time: Wednesday, August 1, 2018 : 10:30 AM to 11:15 AM
Sponsor: Social Statistics Section
Abstract #332920
Title: Estimating the Size of a Hidden Finite Set: Large-Sample Behavior of Estimators
Author(s): Si Cheng* and Daniel J. Eck and Forrest W Crawford
Companies: Yale School of Public Health and and Yale School of Public Health
Keywords: large-sample; asymptotics; population size; capture-recapture; multiplier method; network scale-up
Abstract:

Estimating the size of a hidden finite set is an important task in a wide variety of scientific fields. Often it is infeasible to directly enumerate the elements, and statisticians have developed various indirect approaches using structural assumptions about ordering or network relationships between units. However, evaluating the large-sample and large-population properties of these estimators can be challenging. Most popular estimators are consistent under an "infill" asymptotic regime in which the population size is fixed, while the number of samples grows. But this scenario may be starkly different from the constraints in practice. We formalize two asymptotic regimes, "infill" and "outfill", to study the large-sample behaviors of those methods. The outfill regime allows the population size to grow together with the sample size, with the latter approaching a fixed proportion of the former. This framework provides a unifying perspective for understanding asymptotic claims on hidden set size estimators. We derive the asymptotic properties of popular estimators and show that, under the outfill setting, inconsistency is a common problem from which almost all estimators suffer.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program