Testing Methods of Record Linkage on Swedish Censuses

Abstract.

Research benefits a great deal when different kinds of registers can be combined. Record linkage is an important tool for connecting sources to create longitudinal databases of individual data. In this article, researchers test and evaluate different methods of record linkage used when linking two censuses. By comparing the results of the census linkage with other continuous Swedish parish registers, they find that applying constructed name variables and household links considerably increases the success rate without the risk of introducing bias. Missing links are mainly related to name problems, and appear most frequently among children and when the family structure has changed between the censuses. Faulty links are very few and must be regarded as only a marginal problem for analysis. The study underlines the importance of adapting the linkage process to the special characteristics of the sources.     

Linking the 1940 U.S. Census with modern data

Abstract

The U.S. Census Bureau has created a set of linkable census, survey, and administrative records that provides longitudinal data on the American population across the past eight decades. While these files include modern decennial censuses, Census Bureau surveys, and administrative records files from other federal agencies, the long time span is only possible with the addition of the complete count 1940 Census microdata. In this paper, we discuss the development of this linked data infrastructure and provide an overview of the record linkage techniques used. We primarily focus on the techniques used to produce a beta version of a linkable 1940 Census microdata file and discuss the potential to further document and extend the infrastructure.     

The Mormon Historical Demography Project

Abstract

This article describes the background and reconstruction of the Belgian 1961 census sample file. Recently, a complete codebook for the sample file has been drawn up, and the representativeness of the sample has been evaluated as accurate, except for the loss of cases that occurred for three Belgian subregions. With the addition of the 1961 census sample, cross-sectional microdata from the five successive population censuses since that year, covering a broad range of topics, are now available to the research community. The reconstruction of the 1961 census sample was part of the activities of the HISSTAT project, which is currently being run by a consortium of research institutions in Belgium to bring together cross-sectional aggregated data since 1800, aggregated longitudinal data since 1880, and cross-sectional microdata since 1961, all covering the entire country.     

ASPECTS OF DOMESTIC SERVICE IN GREAT BRITAIN AND IRELAND, 1841–1911

Abstract

Domestic service has been neglected by historians until comparatively recently. Partly the lack of attention has been due to the paucity of sources – although these tend to emerge when interest grows – and a lack of interest in women’s history. Some research has, however, been done and here a number of recent studies are discussed. These relate to the geographical origins of the servants employed in the six houses occupied by the 7^th Duke of Devonshire at the censuses of 1861-91; the changes in servant-keeping experienced by two upper-middle-class owners of a house in Hertfordshire, in the late 19^th century; aspects of ‘kin’ and ‘true’ servants in the Scottish textile town of Kilmarnock as revealed by the 1871 census; and some features of domestic servants working at the Roman Catholic seminary of St Patrick’s College, Maynooth, Ireland, based upon the household schedules of the 1911 census and a wages book

These and other studies reveal that it is from local, micro-level research that the most fruitful insights into the experience of domestic service in the past are likely to emerge. This is, in part, because of the variety of that experience; in part, because that is where the sources are to be found; and, in part, because the sources often require a local interpretation. For example, work on the census enumerators’ books is revealing that each 19^th-century census is best seen as a series of local censuses, often reflecting local interpretations of standard questions.     

Delineating the nation: South African censuses 1865–2007

The census plays a significant role in delineating the nation in statistical terms. The decisions as to whom to enumerate, what questions are to be asked and how the results are presented all modify the view of the population offered to contemporary observers and to posterity. Although census officials tend to be conservative in retaining a large body of questions in similar form from one enumeration to the next in order to promote inter-census comparisons, those concerned with identity have tended to shift with the political evolution of the state and nation. Nowhere has this been more in evidence than in South Africa where the state and nation have been redefined several times since the commencement of modern scientific censuses in 1865. Administrations run by the British Empire, Boer republics, Union of South Africa, apartheid republic, African ‘bantustans’ and now democratic republic have each brought their own concepts to national identification and the framing of the questions of national identity in the census. As a result the set of nearly forty censuses present an often contradictory and complex image of the South African population, ranging from comprehensive inclusive censuses to narrowly restrictive enumerations of a single ethnic group. There was thus little of the continuity in census taking between the colonial and post-colonial states noted elsewhere. South African censuses therefore offer an insight into how the nation was viewed at the time the census was undertaken.     

Record linkage in the Cape of Good Hope Panel*

Abstract

In this article, we describe the record linkage procedure to create a panel from Cape Colony census returns, or opgaafrolle, for 1787–1828, a dataset of 42,354 household-level observations. Based on a subset of manually linked records, we first evaluate statistical models and deterministic algorithms to best identify and match households over time. By using household-level characteristics in the linking process and near-annual data, we are able to create high-quality links for 84% of the dataset. We compare basic analyses on the linked panel dataset to the original cross-sectional data, evaluate the feasibility of the strategy when linking to supplementary sources, and discuss the scalability of our approach to the full Cape panel.     

Gender trouble: Women and the making of Italy's ‘active population’, 1861‐1936

Abstract

This article seeks to contribute to a better understanding of the complex sociocultural and ideological reasons that lay behind the well‐known fall in the number of women workers in the Italian population censuses from the post‐Unification period to Fascism. This was a time when the statisticians of the whole ‘civilized’ world were engaged in the definition of the modern statistical notion of ‘active population’. Through an examination of material published by Italian as well as non‐Italian statistical institutions and of the debates on women's work at the turn of the century, the article shows how the statisticians sought to make their data comparable to those of other, more modem countries and argues that an important turning point in the changing statistical representations of women's work in Italy occurred with the census of 1901. This coincided with growing interest in the question of women's work and the campaign for the introduction of legislation ‘protecting’ women workers. As state constructions, the censuses clearly contributed to the production of a more masculinized image of the labor force, an image which, however, eventually became too distant from reality even for the statisticians themselves.     

‘Ariane's Thread’: Sri Lankan Census Superintendents’ Reports as Guides to Identity Data

National censuses collect, classify and tabulate vast quantities of personal data on a regular basis. The resultant numerical tables require analysis and interpretation. It has been the census superintendents, charged with the collection and processing of the data, who made the first comments on their work. Their reports provide guides for users seeking to make sense of the complex numerical tabulations. An examination of the reports on the censuses of Sri Lanka demonstrates the evolution of the census superintendents’ approaches to the collection and presentation of data on ethnicity, religion and language. The classification schemes developed in the early colonial censuses have been retained, with modifications, until the present. Continuity and change are documented in the official commentaries, which have acted as guides to make the results, and the reasoning behind them, more accessible to those who not only filled in the questionnaires, but also used the statistics.     

EPICENTRES OF MISSING DAUGHTERS

The spatial spread of the female child deficit in India has been explored in terms of hot spots (with high deficits) and cold spots (with more female-favourable child sex ratios). It has been argued, using the Census of India data from individual censuses, that there is a contagion effect for both hot spots and cold spots. This paper takes this discussion forward by asking whether such an effect can be seen across censuses. To do so, it develops the concept of an epicentre to see whether a hot or cold spot in one census spreads shock waves across a wider region in later censuses. The longitudinal analysis of child sex ratios over three censuses—1991, 2001 and 2011—shows that the hot spots are epicentres for the spread of female child deficits, while the cold spots display a reverse effect.     

Trends in the residential mobility of seniors in Canada, 1961–2006

This article examines residential mobility for seniors 65 years of age and older in Canada using census data from 1961–2006. We addressed three questions. First, have seniors been increasingly likely to change their residential location within Canada or alternatively become increasingly likely to age‐in‐place? Second, has the in‐migration of seniors to Canada from other countries become more pronounced over the years? Third, does the residential mobility of seniors vary by age and sex? We used census data to calculate the percentages of seniors who changed their residence in the five‐year periods prior to each of the 1961–2006 censuses and the percentages of seniors who moved in the previous year for the 1991–2006 censuses. We calculated the percentages of seniors making local moves, longer distance moves within the same province, moves from one province to another, and moves to Canada from another country. We found that rates of residential mobility for seniors tended to increase in the 1961–1981 period but have been lower and relatively consistent from 1986–2006. We found no evidence to suggest a pattern of sustained increase in residential mobility of seniors. We conclude that Canadian seniors tend to age‐in‐place and that when seniors do change residence, the likelihood of residential mobility decreases with the distance of the move and decreases with age. Nevertheless, the likelihood of changing residence may increase for seniors 75+ years of age who need assistance and are at risk of institutionalization. We found that senior women were more likely to change residence locally than senior men. Finally, we found that from 1961 to 2006 between 0.8 percent and 1.4 percent of seniors had migrated to Canada in the five years prior to each census from other countries and that this pattern has fluctuated over the past half century with no clear trend.     

The quest for a census of the British Empire c.1840–1940

Regular population censuses are an integral part of the oversight functions of the modern state. Whereas the United Kingdom instituted a decennial enumeration in 1801, it was not extended to the overseas Empire. The administration of the British Empire was decentralised and early nineteenth-century census taking was subject to local initiatives. However, from the 1840s successive Colonial Secretaries considered a unified imperial census necessary to fulfil their oversight functions and demanded the taking of censuses by colonial governments to coincide with that in the United Kingdom. Initial responses were mixed, but with each decade the coverage improved, although no agreement was reached on the questions posed and the classification systems employed. These remained severe drawbacks to the creation of a comprehensive imperial view. Only in 1906 was an official volume entitled Report on the Census of the British Empire published. The experience gained in its preparation, collating the diverse colonial reports prompted the undertaking of a fully coordinated enumeration in 1911. The First World War intervened and prevented publication. Although subsequent attempts were made to revive the concept, little was achieved. The combination of the pursuit of local interests, accentuated by greater political independence, and the lack of adequate resources at the General Register Office in London to coordinate and analyse the results ensured that the quest for a unified census was finally abandoned with the onset of the Second World War.     

The ethnic classification in the late 19th‐century censuses

The topic of this paper is the ethnic classification used in the last three censuses of the 19th century, i.e. the censuses taken in 1875, 1891 and 1900. The aim is to study the content of this classification diachronically, based on new and supplementary evidence from the parish of Astafjord in the southern part of Troms county, northern Norway. This evidence primarily comes from the parish registers from one decade which, unusually, contain rather elaborate characterizations of ethnic affiliation. This feature of the parish registers gives a unique opportunity to check the corresponding characterizations in the censuses.     

Simple strategies for improving inference with linked data: a case study of the 1850–1930 IPUMS linked representative historical samples

Abstract

New large-scale linked data are revolutionizing quantitative history and demography. This paper proposes two complementary strategies for improving inference with linked historical data: the use of validation variables to identify higher quality links and a simple, regression-based weighting procedure to increase the representativeness of custom research samples. We demonstrate the potential value of these strategies using the 1850–1930 Integrated Public Use Microdata Series Linked Representative Samples (IPUMS-LRS)—a high quality, publicly available linked historical dataset. We show that, while incorrect linking rates appear low in the IPUMS-LRS, researchers can reduce error rates further using validation variables. We also show how researchers can reweight linked samples to balance observed characteristics in the linked sample with those in a reference population using a simple regression-based procedure.     

The use of Forestry Commission censuses for the study of woodland change

The value of the first two completed Forestry Commission censuses, of 1924 and 1947/9, for the study of woodland change is assessed, and the contribution they might make to the investigation of past planting trends, the history of individual woods and post-war changes in woodland type and species is discussed. The original returns for the 1947/9 census have recently been released under the “thirty years rule” and problems involved in using this source together with possible avenues for further research are outlined.