A new strategy for linking U.S. historical censuses : A case study for the IPUMS multigenerational longitudinal panel
(2022) In Historical Methods 55(1). p.12-29- Abstract
This paper presents a probabilistic method of record linkage, developed using the U.S. full count censuses of 1900 and 1910 but applicable to many sources of digitized historical records. The method links records using a two-step approach, first establishing high confidence matches among men by exploiting a comprehensive set of individual and contextual characteristics. The method then proceeds to link both men and women by leveraging links between households established in the first step. While only the first stage links can be directly comparable to other popular methods in research on the U.S., our method yields both considerably higher linkage rates and greater accuracy while only performing negligibly worse than other algorithms in... (More)
This paper presents a probabilistic method of record linkage, developed using the U.S. full count censuses of 1900 and 1910 but applicable to many sources of digitized historical records. The method links records using a two-step approach, first establishing high confidence matches among men by exploiting a comprehensive set of individual and contextual characteristics. The method then proceeds to link both men and women by leveraging links between households established in the first step. While only the first stage links can be directly comparable to other popular methods in research on the U.S., our method yields both considerably higher linkage rates and greater accuracy while only performing negligibly worse than other algorithms in resembling the target population.
(Less)
- author
- Helgertz, Jonas LU ; Price, Joseph ; Wellington, Jacob ; Thompson, Kelly J. ; Ruggles, Steven and Fitch, Catherine A.
- organization
- publishing date
- 2022
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- census data, machine learning, Record linkage, United States of America
- in
- Historical Methods
- volume
- 55
- issue
- 1
- pages
- 12 - 29
- publisher
- Heldref Publications
- external identifiers
-
- scopus:85119255501
- pmid:35846520
- ISSN
- 0161-5440
- DOI
- 10.1080/01615440.2021.1985027
- language
- English
- LU publication?
- yes
- id
- 2d78057c-9133-4b4f-ab2c-2f1306f7f322
- date added to LUP
- 2021-12-13 15:08:09
- date last changed
- 2024-06-15 22:36:50
@article{2d78057c-9133-4b4f-ab2c-2f1306f7f322,
abstract = {{<p>This paper presents a probabilistic method of record linkage, developed using the U.S. full count censuses of 1900 and 1910 but applicable to many sources of digitized historical records. The method links records using a two-step approach, first establishing high confidence matches among men by exploiting a comprehensive set of individual and contextual characteristics. The method then proceeds to link both men and women by leveraging links between households established in the first step. While only the first stage links can be directly comparable to other popular methods in research on the U.S., our method yields both considerably higher linkage rates and greater accuracy while only performing negligibly worse than other algorithms in resembling the target population.</p>}},
author = {{Helgertz, Jonas and Price, Joseph and Wellington, Jacob and Thompson, Kelly J. and Ruggles, Steven and Fitch, Catherine A.}},
issn = {{0161-5440}},
keywords = {{census data; machine learning; Record linkage; United States of America}},
language = {{eng}},
number = {{1}},
pages = {{12--29}},
publisher = {{Heldref Publications}},
series = {{Historical Methods}},
title = {{A new strategy for linking U.S. historical censuses : A case study for the IPUMS multigenerational longitudinal panel}},
url = {{http://dx.doi.org/10.1080/01615440.2021.1985027}},
doi = {{10.1080/01615440.2021.1985027}},
volume = {{55}},
year = {{2022}},
}