<abstract xmlns="http://www.w3.org/1999/xhtml"><p>The Capture-recapture method is a well-known solution for evaluating the unknown size of a population. Administrative data represent sources of independent counts of a population and can be jointly exploited for applying the capture-recapture method. Of course, administrative sources are affected by over- or undercoverage when considered separately. The standard Petersen approach is based on strong assumptions, including perfect record linkage between lists. In reality, record linkage results can be affected by errors. A simple method for achieving linkage error-unbiased population total estimates is proposed in Ding and Fienberg (1994). In this article, an extension of the Ding and Fienberg model by relaxing their conditions is proposed. The procedures are illustrated for estimating the total number of road casualties, on the basis of a probabilistic record linkage between two administrative data sources. Moreover, a simulation study is developed, providing evidence that the adjusted estimator always performs better than the Petersen estimator.</p></abstract>

The Capture-recapture method is a well-known solution for evaluating the unknown size of a population. Administrative data represent sources of independent counts of a population and can be jointly exploited for applying the capture-recapture method. Of course, administrative sources are affected by over- or undercoverage when considered separately. The standard Petersen approach is based on strong assumptions, including perfect record linkage between lists. In reality, record linkage results can be affected by errors. A simple method for achieving linkage error-unbiased population total estimates is proposed in Ding and Fienberg (1994). In this article, an extension of the Ding and Fienberg model by relaxing their conditions is proposed. The procedures are illustrated for estimating the total number of road casualties, on the basis of a probabilistic record linkage between two administrative data sources. Moreover, a simulation study is developed, providing evidence that the adjusted estimator always performs better than the Petersen estimator.

Coverage Evaluation on Probabilistically Linked Data

Italian National Statistical Institute - Istat, Via Cesare Balbo, 16 00184 Rome, Italy

Journal of Official Statistics

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

{"article-title":"Coverage Evaluation on Probabilistically Linked Data"}

The Capture-recapture method is a well-known solution for evaluating the unknown size of a population. Administrative data represent sources of independent...

Coverage Evaluation on Probabilistically Linked Data

Publicado en línea: 01 sept 2015

Páginas: 415 - 429

Recibido: 01 feb 2014

Aceptado: 01 ene 2015

DOI: https://doi.org/10.1515/jos-2015-0025

Palabras clave
Linkage errors, capture-recapture method, Petersen estimator, administrative data

© by Loredana Di Consiglio

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

Coverage Evaluation on Probabilistically Linked Data

Publicado en línea: 01 sept 2015

Páginas: 415 - 429

Recibido: 01 feb 2014

Aceptado: 01 ene 2015

DOI: https://doi.org/10.1515/jos-2015-0025

Palabras claveLinkage errors, capture-recapture method, Petersen estimator, administrative data

© by Loredana Di Consiglio

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

Palabras clave
Linkage errors, capture-recapture method, Petersen estimator, administrative data