Nominal record linkage of historical data : procedure and applications in a North Wales parish
Nominal record linkage of historical data : procedure and applications in a North Wales parish
During the late nineteenth century, Britain experienced two demographic phenomena of great importance. First, there was a substantial and irreversible fall in fertility rates; a phenomenon referred to as the fertility transition. Second, there was large scale population movements from rural to urban areas; this phenomenon was also largely irreversible.
Aggregative analyses of census data and to a lesser extent family reconstitution methods have been used to investigate these events. However, both are deficient in the questions that can be answered about these occurrences, because of the techniques and the data employed. To adequately study both phenomena, individual-level longitudinal data of sufficient quality is required. Little has been done in this field due to fears about the quality of ecclesiastical registration data during the late nineteenth century. This is overcome in this study by using civil registration data. Also the technique employed to obtain individual-level longitudinal data known as record linkage is extremely laborious when done manually. This can be overcome by automating the linkage process.
This thesis considers the development of an automated record linkage system using a computer database. The issues involved in automated record linkage are discussed at length; especially that of the problem of name variations on original manuscript sources. In this study the problem is overcome through the development of a program which combines the use of well established name standardisation procedures to identify name similarities.
The record linkage process itself mimics the actions of a manual record linker by employing probabilistic making rules. This is achieved through the assignment of weights to specific agreements and disagreements of identifying items on compared records. The calculation and assignment of these weights are discussed in detail. The problem of determining the accuracy of the linkage process is discussed and the results of an investigation into any inherent bias in the process are presented.
University of Southampton
1993
Davies, Hywel Rhodri
(1993)
Nominal record linkage of historical data : procedure and applications in a North Wales parish.
University of Southampton, Doctoral Thesis.
Record type:
Thesis
(Doctoral)
Abstract
During the late nineteenth century, Britain experienced two demographic phenomena of great importance. First, there was a substantial and irreversible fall in fertility rates; a phenomenon referred to as the fertility transition. Second, there was large scale population movements from rural to urban areas; this phenomenon was also largely irreversible.
Aggregative analyses of census data and to a lesser extent family reconstitution methods have been used to investigate these events. However, both are deficient in the questions that can be answered about these occurrences, because of the techniques and the data employed. To adequately study both phenomena, individual-level longitudinal data of sufficient quality is required. Little has been done in this field due to fears about the quality of ecclesiastical registration data during the late nineteenth century. This is overcome in this study by using civil registration data. Also the technique employed to obtain individual-level longitudinal data known as record linkage is extremely laborious when done manually. This can be overcome by automating the linkage process.
This thesis considers the development of an automated record linkage system using a computer database. The issues involved in automated record linkage are discussed at length; especially that of the problem of name variations on original manuscript sources. In this study the problem is overcome through the development of a program which combines the use of well established name standardisation procedures to identify name similarities.
The record linkage process itself mimics the actions of a manual record linker by employing probabilistic making rules. This is achieved through the assignment of weights to specific agreements and disagreements of identifying items on compared records. The calculation and assignment of these weights are discussed in detail. The problem of determining the accuracy of the linkage process is discussed and the results of an investigation into any inherent bias in the process are presented.
This record has no associated files available for download.
More information
Published date: 1993
Identifiers
Local EPrints ID: 462779
URI: http://eprints.soton.ac.uk/id/eprint/462779
PURE UUID: 12ac4e8b-e279-4f0f-bbeb-c95ee0008f85
Catalogue record
Date deposited: 04 Jul 2022 20:02
Last modified: 04 Jul 2022 20:02
Export record
Contributors
Author:
Hywel Rhodri Davies
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics