Historical demography abstracts

Strand organizer: Dr. Romola Davenport, University of Cambridge 

Mortality & morbidity in historical populations:  Monday 7 September 4.45pm 

Mortality differentials during the demographic transition: A linked-records study from Tartu (Estonia), 1897–1899
Hannaliis Jaadla, Tallinn University 

This study analyses mortality differentials among the Lutheran population in the urban Estonian environment of Tartu at the end of the 19th century. Death records linked to individual-level census data from 1897 and birth records from two subsequent years allow for an examination of mortality variation among sub-groups of the population, thereby providing insight into factors shaping survival over the life course at the height of mortality transition in Estonia. Poisson regression models were estimated separately for children, adults and elderly and differentiated between exogenous (infectious diseases and accidents) and endogenous (chronic diseases of later life) mortality. The findings indicate that socio-economic status was more important for men than for women in determining mortality risk. Socio-economic differences were not a significant source of variation in child mortality. Instead, sanitary conditions (water supply, location of privy) exerted a stronger influence on mortality during the first 15 years of life. Households that used the river as their main source of drinking water showed substantially lower survival rates for children. For adults, mortality levels among the foreign-born population were lower, a finding also demonstrated in other settings. In the late 1890s, 65 per cent of causes of death in Tartu could be included under exogenous mortality. The analysis found strong influence of socio-economic characteristics and sanitary conditions of the households on exogenous mortality. 

E-mail: hjaadla@tlu.ee

The Widowhood Effect in Polygamous Marriages: Evidence from Historical Utah 
Kieron Barclay1, Robyn Donrovich2, Heidi Hanson3, Ken R. Smith3; 1London School of Economics , 2KU Leuven, 3University of Utah

In this study we examine the impact of being in a polygamous marriage on mortality using data from frontier societies in what is now Utah in the United States. Furthermore, we examine whether the widowhood effect on mortality differed in polygamous marriages, and whether it varied, for the husband, according to the loss of multiple wives, and for the wife, whether it varied according to the loss of the husband or a sister-wife. Polygamy is still practiced in many countries today, and we are not aware of any research examining how mortality in polygamous marriage differs from mortality in non-polygamous marriages, nor of how the widowhood effect differs in polygamous marriages. We use data from the Utah Population Database on 44,868 individuals born between 1792 and 1900. We find that individuals in plural marriages had lower mortality, even after adjusting for membership of the Church of Latter-day Saints (Mormons). Consistent with previous work, we find that mortality was elevated after the death of a spouse in non-plural marriages. However, in plural marriages mortality was not elevated after the death of a spouse. For women in plural marriages, mortality was only significantly elevated after the death of the husband as well as the death of two sister wives. We believe that our results may be able to shed light on the explanatory mechanisms for the widowhood effect, as an increase in the risk of mortality following the death of a husband versus a sister-wife are likely to be explained differently. 

E-mail: k.j.barclay@lse.ac.uk 

A frame of persons affected by pellagra in an Italian area: strategies and practices of care and social security (late XIX-early XX century) 
Giovanna Da Molin, Maria Federighi, Department of Education, Psychology and Communication, University of Bari 

The paper deals with the issue of assistance practices for the benefit of persons affected by pellagra within an Italian area at the end of the 19th century, focussing on prevention, care and reclusion strategies provided by the authorities. The research seeks to understand and quantify the impact of this disease on population, the reasons, the social effects
and the action plan of the authorities. Since the early 19th century pellagra was endemic in most rural areas of northern and central Italy, with an agrarian economy marked by maize and a lower classes’ diet almost entirely based on this cereal. The paper aims to highlight the spread of pellagra and the measures adopted in terms of prevention, assistance and care, through the setting up of specific structures in order to guarantee social security.
The methodology, quantitative and qualitative, uses previously unknown archival sources between 1896 and 1903 stored in two Archives of Bologna. The cross analysis of sources allow us to follow life-stories of some persons suffering from pellagra. Such sources allow us to draw out information on the spreading of pellagra according to gender, age, employment, geography, medical certifications, distance from the nearest structures, stadium of the disease, status after the intensive care, as well as various types of assistance: from the subsidy up to admission to and permanent residence within lunatic asylums or specific structures. Following this completed study, we are currently investigating new archival sources on pellagra in different Italian areas. 

E-mail: giovanna.damolin@uniba.it


Fertility & kinship in historical populations: Tuesday 8 September 1.45pm 

Diversity in Demographic Outcomes: Scottish Chain Migration to Ontario, 1871-1901 
Sara Rose, Wilfrid Laurier University 

The Scottish migration experience is marked by its significant use of chain migration. This paper focuses on how migrant chains affected Scottish settlement patterns in Ontario in the late 19th century, highlighting their reliance on friends and family to successfully relocate on the opposite side of the Atlantic Ocean. Remittances, letters, and other forms of support point to the continued importance of kin and clam across borders. What differentiates Scottish migrants from chain migrants of other nationalities is the durability of their settlements. Chains from other origins typically produce durable settlements that persist over time. Scots, on the other hand, show significantly less settlement durability, with variation across the settlement styles of Highland and Lowland Scots. Census data is used to describe the concentration of Scottish immigrants over time within districts in Ontario, Canada, and how the degree of concentration changed over time. This data and the results described are illustrated with a series of maps and confirmed by regression analysis. 

E-mail: srose@balsillieschool.ca 

The low levels of fertility during the interbellum in France 
Sandra Brée, Centre for Demographic Research, Université catholique de Louvain 

Between 1920 and 1940, fertility in many European countries had reached very low levels, almost always below the replacement level. France was no exception to this rule and women born between 1890 and 1910 were characterized by very low completed childbearing, relatives to two children per woman, or even less. This papers aims to understand why and how such low levels of fertility were reached during the interbellum? Thanks to the retrospective approach of 20th century fertility from French censuses and Family surveys, it is possible to reconstruct women lifetime fertility. Indeed, in the 1946 census and in 1954 and 1975 “Family surveys”, each woman was asked about the number of live births she had and the birth date of each child. The size of siblings, the age at first marriage and at first birth (according to the rank of children) will be analysed to understand the weight of childlessness and low fertilities (one-child women) and the role of postponement of first marriage and first birth in the low levels of fertility observed. Another focus of this research is to get to the infranational levels (namely the French départements) to determine if childlessness and low fertilities are generalized or if some areas or environment are more concerned than others. 

E-mail: sandra.bree@uclouvain.be

Genealogical analysis of participants in the CARTaGENE project (Quebec, Canada) 
Marc Tremblay, Université du Québec à Chicoutimi 

The main objective of the CARTaGENE project (www.cartagene.qc.ca) is the creation of a large database on the Quebec (Canada) population, covering various aspects of public health. CARTaGENE is part of the P3G consortium (Public Population Project in Genomics and Society), an international organization that aims to provide the scientific community with access to the most current information in all areas of population genomics (http://P3G.org). The first phase of the CARTaGENE project involved an extensive survey carried out with 20,000 people aged between 40 and 69 years. Participants were recruited in four regions of the province: Montreal, Quebec City, Sherbrooke and Saguenay. In addition to biological samples stored at the Genome Quebec Biobank (www.genomequebec.com), data on more than 650 demographic and health variables were collected. Participants in the survey were also invited to transmit information to the BALSAC Project (http://balsac.uqac.ca) for genealogical purposes. The ascending genealogies of 5,095 participants were reconstructed in order to establish a demogenetic portrait of the four regional populations. Most genealogies extend back to the 17th century, with a mean genealogical depth of nearly 10 generations. Such a genealogical completeness allows for a very rich and precise measurement of inbreeding and kinship ties in the population, which may be crucial for studies aiming to identify genetic variations associated with diseases. Inbreeding and kinship coefficients were calculated at each generation level, up to the 13th generation. Results showed significant differences between the four regional populations. 

E-mail: marc.tremblay@ugac.ca   

Measuring Education from the present in the past - The construction of a global dataset on the educational attainment in the 20th Century 
Markus Speringer, Anne Goujon,  Samir K.C., Jakob Eder, Ramon Bauer, Michaela Potancokova, Wittgenstein Center for Demography and Global Human Capital (IIASA, VID/ÖAW, WU) 

This paper describes the first results and the methodology to develop a Historical Human Capital Database based on two major efforts: first, the collection, harmonization and validation of datasets on educational attainment and literacy by age and sex in the 20th century (Speringer et al. 2015) and second, the reconstruction of the data gaps in the time series based on the multi-dimensional cohort-component projection method, that can be used for back-projecting populations. (K.C. et al. 2015) The first base for the backprojection model is the data on educational attainment gathered for the Wittgenstein Centre new projection round which allows us to reconstruct the past until 1970 (Bauer et al. 2012; Lutz, Butz, and KC 2014). However more data points are needed. So far most endeavors aiming at collecting data plus systematically reconstructing and modelling detailed information on past levels of education and literacy (by age and sex) of the population have stopped at 1950 or after (Lutz et al. 2007; Barro and Lee 2013; de la Fuente and Doménech 2012). A longer perspective can advance our knowledge of human capital formation and its consequences on demographic, socio-economic, political and technological changes as human capital advancement is highly path dependent (Lutz et al. 2007; Goldin and Katz 2008). The resulting Historical Human Capital Database will enable to study these interdependencies more precisely and in longer time-span to possibly better understand and shape our future. 

E-mail: markus.speringer@oeaw.ac.at 


Marriage & death in Britain 18001911: Wednesday 9 September 9.00am 

Urban mortality and standards of living in Britain 1800-1860 
Romola Davenport, Cambridge Group for the History of Population and Social Structure, University of Cambridge 

The malign contribution of northern industrial cities to the stagnation of national life expectancy over the period 1820-1870 forms part of one of the most long-running debates in English economic history, regarding the impact of early industrialisation on living standards. The pessimistic view argues that industrial cities experienced a worsening of mortality especially in the second quarter of the nineteenth century, and that this was due to the peculiar conditions of industrialisation, including administrative breakdown and rising social inequalities. An alternative view is that any mortality rise that did occur was largely a function of population density and exogenous epidemiological change. This paper uses cause of death data from Glasgow and Manchester, and from the US cities of Boston, Baltimore and Philadelphia, to explore this debate. Our results indicate that, as expected, measles epidemics increased in frequency in rapidly growing towns and must have contributed to increases in early childhood mortality. It is widely accepted that scarlet fever increased in virulence in the third decade of the 19th century and this phenomenon is clear in all towns examined. Smallpox mortality, having fallen dramatically in the first decade of the 19th century with the advent of vaccination, appears to have risen again by the end of the second decade. This rise was relatively slight and also appears to have been widespread in both urban and rural areas suggesting that it was more a function of rising uncertainties about the efficacy and safety of vaccination than a specific failure of administrative will. 

E-mail: rjd23@cam.ac.uk 

Spatial modelling of rural infant mortality in nineteenth-century Britain 
Paul Atkinson, Ian Gregory, Catherine Porter, Brian Francis, Lancaster University 

Rural infant mortality in later nineteenth century Britain is an under-researched issue. We ask which places had similar experiences and what factors most strongly determined these. Latent class trajectory modelling of changes in infant mortality rate (IMR) over time is used to identify clusters of Registration Districts (RDs) with similar trajectories: we display these spatially using the Great Britain Historical Geographical Information System (GBHGIS). The paper then presents the largest longitudinal study so far conducted of British infant mortality, using data from 363 rural RDs of England and Wales spanning 1851-1911. Data are mainly drawn from the Decennial Supplements on Mortality of the Annual Report of the Registrar General, digitised in the GBHGIS. Preliminary results show striking contrasts between mainly northern and western RDs which began the period with the lowest IMRs of all and experienced an absolute increase in IMR, and clusters of RDs with a variety of falling IMR trajectories, generally – though by no means all – in the south and east of England. Longitudinal regression by cluster of IMR against a set of independent variables suggested by the literature suggest that the explanatory power of the usual candidates (population density, fertility, female literacy and living standards) varies for rural RDs, correlating nationally with 50% of the change in IMR. These findings have the potential to deepen our understanding of rural demographic change, pointing to a role for changing spatial patterns of female human and social capital and maternal health. 

E-mail: p.atkinson3@lancaster.ac.uk 

A detailed geography of early-age mortality in England and Wales, 1851-1911 
Alice Reid, University of Cambridge 

It is well known that there is considerable spatial variation in infant and early child mortality, and that the trajectories of decline in early age mortality were also differentiated by place. The most detailed spatial analyses of nineteenth century mortality on a large scale use decennial data for the 614 registration districts in England and Wales (Woods and Shelton 1997). However studies on selected smaller areas have shown that registration districts can contain considerable variation in the risk of death due to variations in population density, occupational and industrial make-up, environmental hazards and local-disease environments. This paper will use new data for infant mortality for the over 2,000 registration sub-districts of England and Wales to present detailed maps of mortality over the late nineteenth century. The paper will examine spatial and temporal variation in the seasonality of infant mortality to identify changes in the geographic pattern of seasonally specific causes of death over time. It will also explore differences in the relationship between infant and early child mortality over space and time and consider different methods of estimating early childhood mortality (age 1-4) for sub-registration districts. 

E-mail: amr1001@cam.ac.uk 

Marriage patterns in Victorian England and Wales: an analysis based on registration subdistrict data, 1851-1911 
Eilidh Garrett1,  Alice Reid2, Eddy Higgs1; 1University of Essex, 2University of Cambridge, 

This paper builds on work by Anderson (1976), Hinde and Woods (Woods and Hinde, 1985; Woods, 2000) and Day (unpublished PhD thesis) who have previously considered marriage patterns in Victorian Britain. In their work Anderson, Hinde and Woods lamented that the scale and detail of the statistics published on nuptiality were inadequate to the task of building up a comprehensive picture of marriage behaviour in Britain. Woods and Hinde were able to demonstrate how useful individual level data were when interpreting the broader statistics, but their supply of such data was very limited. Woods (2000, Figure 3.9) had to use Knodel’s data for Germany in 1880 to illustrate the relationship between three main measures of nuptiality: the percentage of women never married, female SMAM and Im, lamenting that such information was not available for English and Welsh registration districts. With the release of the Integrated Census Microdata (I-CeM) dataset, which includes all individuals enumerated in each census from 1851-1911 (bar 1871) it is now possible to calculate marriage profiles over space, through time and across occupational groups for England and Wales. This paper will report on an exercise using the ICeM data to assemble the measurements needed to reproduce Wood’s figure using sub-registration data for England and Wales at various time points across the 1851-1911 period. These will then be used to identify, and map, the presence and extent of different nuptiality regimes within the two countries. Finally, the conclusions of the previous authors will be reviewed and re-evaluated. 

E-mail: eilidh.garrett@btinternet.com