Abstract
This article introduces the new dataset Cohort Component Population Estimates for Ireland, 1911–1920 (www.doi.org/10.5255/UKDA-SN-854673). The dataset provides vital statistics and cohort component population estimates at a spatially-disaggregated level for the island of Ireland for the period 1911–1920. The raw data were digitised by the authors using official UK government statistics. The population estimates were then derived by the cohort component method. These data provide novel intercensal population estimates at the county level that will be beneficial for researchers working in historical demography, as well as in economic and social history. The data provided can be readily reused and extended by other researchers to produce further series and indicators. An example application of the data in this manner is Colvin and McLaughlin (2021), who combine these population estimates with mortality statistics from the Spanish flu pandemic to demonstrate how demographic composition affects the interpretation of data on public health crises.
Online publication date: 11-05-2022
- –Related data set “Cohort Component Population Estimates for Ireland, 1911–1920” with doi www.doi.org/10.5255/UKDA-SN-854673 in repository “UK Data Service”
1. Introduction
- –Cohort Component Population Estimates for Ireland deposited at UK Data Service – doi:www.doi.org/10.5255/UKDA-SN-854673
- –Temporal coverage: 1911–1920
This article presents a novel historical demography dataset for Ireland in the period 1911–1920 inclusive, intended to serve as a comprehensive account of population change up to the point of Irish partition following the Government of Ireland Act 1920. The dataset contains new disaggregated estimates for the mid-year population and population composition for all 32 counties of Ireland, and with Belfast and Dublin reported separately, for each year between 1911 and 1920 (see the Appendix for the specifications table). Table 1 summarises this series, showing the average population estimates across all counties in each year by age group and sex.
The dataset also contains the various components of population change which are used to construct the main population estimates. These include: the number of births in each country per year (1911–1920); the number of deaths by age in each county per year (1911–1920), given in five- or ten-year age bins; the number of emigrants leaving from each county per year (1911–1919), from all countries; estimates of the number of international immigrants entering each county per year (1911–1919), to all countries; and estimates of troop movements into and out of each county per year (1914–1919), which are treated as migration movements, as per convention, in the population estimates. These series are formatted as individual tables within the dataset, with column headings displaying a shorthand county id, county name, county province, year, data value, and age range group (where appropriate). Figure 1 reports the trends for total births and total deaths in Ireland over the period. Total deaths fluctuate around an 1911–1917 average of 72,994 before peaking during the years of the Spanish flu pandemic in 1918–1919 and falling sharply thereafter. Total births steadily decline over the period and remain depressed before rebounding following the end of World War i and the Spanish flu pandemic.
Total births and deaths in Ireland, 1911–1920
Citation: Research Data Journal for the Humanities and Social Sciences 7, 1 (2022) ; 10.1163/24523666-bja10022
registrar general’s reports for ireland (british parliamentary papers, 1912b, 1913c, 1914b, 1915b, 1916c, 1917b, 1918b, 1919b, 1920b, 1921b).We believe our dataset will be beneficial to researchers working in Irish economic history, social history, and historical demography. Our population estimates can be readily used to complement and enrich new studies, as well as provide opportunities for researchers to combine them with other data, in order to produce useful quantitative indicators. The time period covered by the dataset overlaps with noteworthy historical events such as the Irish revolutionary period, World War i, and the Spanish flu pandemic. These are areas where existing studies have operationalised demographic data – for example, see Fitzpatrick (1978) and Hart (1997) on the geographic distribution of Irish nationalism and revolutionary behaviour – and, additionally, where researchers must consider the influence of demography in their analysis.
Existing research that seeks to exploit historical Irish demographic data has, for the most part, relied on official government statistics as provided in Ireland’s decennial census reports, beginning in 1821. Vaughan and Fitzpatrick’s Irish Historical Statistics (1978) – an ancillary publication to the multi-volume New History of Ireland project – has proved a venerable resource to this end, compiling the statistical tables on populations and vital statistics between the 1821 and the 1971 censuses. While such data are appropriate in many cases, and hitherto have been the only option for researchers, the decennial frequency of census data calls into question their reliability when analysing events occurring in intercensal periods.
Demographic data can be an invaluable tool for quantitatively driven analyses of historical events, however a failure to use the most accurate and relevant data can potentially mislead and give rise to biased results. Colvin and McLaughlin (2021) provide an example of this using the Irish experience of the Spanish flu pandemic of 1918–1919, showing that a reliance on the nearest census for population data can distort the denominator used in mortality statistics. Our dataset is designed to address these concerns and we recommend its usage in future studies into the aforementioned areas of Irish history.
There is also a great deal of scholarly interest in Irish historical demography and geography for their own sake. Tracing the development of, and exploring the causes behind, the substantial and sustained Irish depopulation following the Great Famine into the twentieth century has received interdisciplinary attention; such exercises also extend beyond the academe, with the government of the Irish Republic having established a Commission on Emigration and other Population Problems in 1948. Guinnane (1997), an exhaustive analysis of post-Famine population change focussed primarily on rural Ireland, is a more recent example in this area. Analysis of Irish population change and its relation to the economy has also extended to the pre-Famine period, such as in Graham and Proudfoot (1993), aided by various attempts at pre-censal population estimation. The construction of this dataset is also our effort in improving Irish population history, where we are seeking to address a lack of readily available demographic information for Ireland between the 1911 and 1926 censuses.
2. Experimental Design, Materials and Methods
Our population estimates are constructed using the cohort component method, described in Smith et al. (2002). This is a commonly used method in demography for future population projection, where the various components of population change – births, deaths, migration, and ageing – are aggregated to update both the total population amount as well as the population composition (presented by age-sex cohorts here) in each year. But rather than projecting future populations, we use this method to fill in gaps in the historical data and to address the lack of intercensal demographic information for Ireland between the 1911 and 1926 censuses. We stop our analysis in 1920 due to disruptions to statistical reporting during the transition to Irish partition.
The historical data that we have collected and digitised affords a relatively simple approach to the cohort component method that is largely atheoretical, requiring few assumptions or complex methods for projecting the components of population change. We start with the population provided in the 1911 Census of Ireland (British Parliamentary Papers, 1913a), the last all-Ireland census undertaken by the British administration, which was digitised by Clarkson et al. (1997a, 1997b). By this time the census reports had become highly accurate in their population counts. The main, relevant concern with the reliability of the 1911 census specifically is its biased age-distribution, resulting from the deliberately inaccurate self-reporting of age among the elderly (Budd and Guinnane, 1991); we discuss how we deal with this below.
where Psact is the population of sex s, at age-range a, in county c, and at time t; and B, D, I and E are the various components of population change, namely within-year births, deaths, immigration, and emigration, respectively. Where possible, we have gathered official, contemporary statistics relating to each of these components; all data sources and estimation methods are described below.
2.1. Births and Deaths
We gathered data on the numbers of births and deaths in each county in each year from the Registrar General’s Reports for Ireland. These report the number of births and deaths within each of the Poor Law Unions of Ireland, which were a second-level administrative unit used for the collection of vital statistics. These annual reports are compatible with the decennial census reports and can therefore be used in conjunction with one another. Walsh (1970) shows that, by 1911, disparities between census data and data reported in the Registrar General’s reports had been near eliminated.
The births and deaths within each Poor Law Union were then aggregated to the 32 counties that constitute the country’s first-level administrative units. These were then further aggregated to Ireland’s four historic provinces (Connacht, Leinster, Munster, and Ulster); while not an official administrative unit, these are considered socially and culturally important. Finally, these were aggregated to the national total.
Both the birth and death statistics are broken down by sex, and death statistics are additionally reported by age in five-year bins. Together, these allow us to also estimate the change in the population by distinct age-sex cohorts.
We manually digitised the birth and death statistics used in our cohort component estimates from each of the Registrar General’s reports between 1911 and 1920 (British Parliamentary Papers, 1912b, 1913c, 1914b, 1915b, 1916c, 1917b, 1918b, 1919b, 1920b, 1921b).
2.2. Migration
Similarly, emigration statistics were also gathered from contemporary British parliamentary records, namely the Emigration Statistics of Ireland reports. These contain the number of emigrants from each of the 32 counties of Ireland, which again are broken down by sex and age groups. We digitised the reports corresponding to each year between 1911 and 1919 (British Parliamentary Papers, 1912a, 1913b, 1914a, 1915a, 1916a, 1917a, 1918a, 1919a, 1920a).
Contemporary statistics for immigration into Ireland do not exist. Instead, we base our estimates on information on annual immigration figures reported by Department of Social Welfare (1955, p. 326). We assume that immigration follows a similar spatial pattern to emigration, and allocate immigrates to counties based on county emigration patterns. This is not a heroic assumption given that immigration at the time mostly constituted returning emigrants, who would return to their families in their home counties (Fernihough and Ó Gráda, 2019). Indeed, from over 14,000 Irish returnees that appear in US passport document data for the years 1890–1920, almost three quarters claim family-related reasons as their return motivation (Dunnigan, 2012).
Table 2 reports the numbers of emigrants, in addition to our estimates on the number of immigrants, for each year between 1911 and 1920.
A limitation of our population estimates is the lack of accounting for internal migration within Ireland. We believe, however, that internal migration is of relatively lower concern for Ireland in this period. Historically, Ireland has had much lower levels of internal migration and urbanisation relative to other countries, owing largely to its slower pace of industrialisation (Guinnane, 1997, p. 122). A comparison of the 1901 and 1911 censuses gives a sense of the limited extent of Irish internal migration; only 13.2 per cent of people were living in a county outside that in which they were born in 1911, compared to 11.4 per cent in 1901 (British Parliamentary Papers, 1901, 1913a). Additionally, World War i further limited the movement of people for most of the intercensal period we deal with.
Internal migration becomes more pronounced following partition of the island, particularly the exodus of minority Protestants and Catholics from the South and North of Ireland respectively (Lynch, 2019, ch. 7). This is another reason – alongside disruption to statistical reporting during partition – why we end the dataset in 1920, as the accuracy of our population estimates would likely be impaired without accounting for this increased internal migration.
2.3. Military Personnel
We account for military enlistment during World War i and treat troop movements as migratory behaviour; Irish soldiers leaving for war is treated as emigration, and subsequent demobilisation is treated as (return) immigration. Military enlistment is treated in the same way in the population estimates provided by the Registrar General’s Reports; for example, the male population is estimated to have decreased by 49,881 between 1914 and 1915, and then increased by 63,000 between 1918 and 1919 (British Parliamentary Papers, 1921b, p. 40).
We estimate total Irish enlistment using contemporary military sources (British Parliamentary Papers, 1921a, p. 9; War Office, 1922, p. 363) and we disaggregate our total enlistment estimate to the county-level using an additional parliamentary source (British Parliamentary Papers, 1916b) that contains data on military age men in Ireland and which covers 97 per cent of the total Irish recruits. War Office statistics are used to estimate military mortality and these imply a mortality rate of 14 per cent for all military personnel.
Figure 2 is a county-level map of military recruitment intensity for the period 1914–1916. This is calculated by taking the estimated total number of recruits in each county as a share of the at-risk county population, defined as the male population aged 20–44 in 1911. Relatively higher levels of military recruitment is observed in urban areas and the province of Ulster.
Estimated relative military recruitment, 1914–1916
Citation: Research Data Journal for the Humanities and Social Sciences 7, 1 (2022) ; 10.1163/24523666-bja10022
Note: See text for calculation of values. Dublin City is included in Dublin County; Belfast in Antrim.colvin et al. (2021); 1911 census of ireland (british parliamentary papers, 1913a).Finally, Table 3 reports the weights used to distribute migrants and military enlistment by age cohort, with migrants weights sourced from Akenson (1993) and military enlistment weights sourced from a contemporary government report (British Parliamentary Papers, 1921a). Contemporary estimates of recruitment discuss both the fitness for general service and the age of the population. It was noted at the time that Ireland’s lower share of the male population enlisting in the military was because “Ireland is chiefly composed of young men up to 18 years of age and of men over 50, as a large proportion of the remainder emigrate to the United States and Colonies” (British Parliamentary Papers, 1921a, p. 9).
The age of military service was between 19 and 41. We have decided to exclude 19-year-olds to avoid introducing complicated assumptions about the proportion of 15–19 year olds eligible for recruitment. For the same reasons, we have included 42–44-year-olds from the 35–44 age group. The exclusion of 19-year-olds and the inclusion of 42–44-year-olds should largely balance one another out in the total recruit estimates. We allocated military recruitment between ages 20 and 44 assuming the age of recruits is proportional to the share of the 20–44 county population.
2.4. Adjusting for Ageing
Once we have accounted for all deaths and migration flows within an age-sex cohort, we then move a share of each age group forward by one year, assuming a survival rate for the last year in the age group. This effectively ages our population and updates the age structure in each year. For example, in the 25–29 age group the ‘forwarded’ share represents the 29-year-olds that have aged and moved to the 30–34 group; these are then replaced in the 25–29 group by a share of the 20–24 group, representing the 24-year-olds who have aged. We derive assumed survival rates of the last year in any age group using weights calculated from the same age group in the 1926 census, which was then conducted and published separately for the UK-polity of Northern Ireland (Government of Northern Ireland, 1929) and the Irish Free State (Roinn Tionscail agus Tráchtála, 1928, 1929). Table 4 reports the weights that are applied to the upper boundaries of the age groups, these being the relevant ages for the ageing adjustment.
2.5. Age-Sex Cohort Categorisation
We mostly follow the age groups that are reported in the Registrar General’s Reports and in the Irish censuses, but we instead choose to group all the under-5-years ages together into one single age band of 0–4 years and to collapse all ages over 65 into one single age band of 65+ years. The latter is less impactful in terms of our data’s granularity, but the choice was made due to perceived discrepancies with age statements in the 1911 census, our base year for the population estimates.
We elaborate on our choice to collapse the over-65s into a single age band: While Ireland’s census commissioners believed the self-reporting of age was more accurate in the 1911 census than in previous censuses (British Parliamentary Papers 1913a, p. 25), Budd and Guinnane (1991) raise concerns about the deliberate overstatement of one’s age to avail of the 1908 Old-Age Pensions Act, resulting in a biased age-distribution. We choose to collapse all age groups over 65 years-of-age as a way to avoid potential distortions in the 1911 census, that may arise from the overstatement of ages, while not trying to manipulate the underlying census data or introducing additional assumptions.
3. Conclusion
We believe this dataset has significant reuse potential for research into modern Irish history, while also constituting a novel contribution to Irish demographic history in its own right. The importance of using accurate and relevant demographic data in quantitative historical research is shown in Colvin and McLaughlin (2021) – part of an ongoing research project on the Spanish flu in Ireland – and we recommend the usage of this dataset for historical analyses of Ireland in this time period.
Notably, the data allow for a re-examination of the Irish revolutionary period with new, intercensal demographic information. This can greatly enhance our understanding of the causes and geography of the Irish War of Independence, complementing existing quantitative analyses by Fitzpatrick (1978) and Hart (1997). For example, the authors – alongside Robin Adams – are using the data to investigate the ‘youth bulge’ hypothesis (Weber, 2019), which links the population age structure to episodes of political violence, in the Irish case in the 1910s and 1920s.
A key theme in the historiography of this period is Irish military recruitment and the Irish home front in World War i, where research has sought to explain military volunteering patterns and disparities across Ireland, evaluate Ireland’s contribution to the war effort, and investigate events such as the 1918 conscription crisis (Bowman et al., 2020; Fitzpatrick, 1995; Gallagher, 2019; Pennell, 2012; Ward, 1974). This is another example of a research area where the dataset will be beneficial, providing spatially disaggregated estimates of population and military recruitment that can be easily integrated and visualised (see Figure 2).
Acknowledgements
We thank Thomas Boys for excellent research assistance. We thank Christiaan van Bochove and two referees for helpful comments on our original submission.
Conflict of Interest
The authors declare that they have no known competing financial interests or personal relationships which have or could be perceived to have influenced the work reported in this article.
References
Datasets
Clarkson, L. A., Kennedy, L., Crawford, E. M., & Dowling, M.W. (1997a). Database of Irish historical statistics: Age, 1821–1911. [Data Collection]. Colchester, University of Essex, UK Data Service, SN: 3574. www.doi.org/10.5255/UKDA-SN-3574-1.
Clarkson, L. A., Kennedy, L., Crawford, E. M., & Dowling, M. W. (1997b). Database of Irish historical statistics: Population, 1821–1911. [Data Collection]. Colchester, University of Essex, UK Data Service, SN: 3578. www.doi.org/10.5255/UKDA-SN-3578-1.
Colvin, C. L., McLaughlin, E., & Richmond, K. J. J. (2021). Cohort component population estimates for Ireland, 1911–1920. [Data Collection]. Colchester, University of Essex, UK Data Service, ReShare Repository. www.doi.org/10.5255/UKDA-SN-854673[Revised February 2022].
Official Publications
British Parliamentary Papers. (1901). Census of Ireland Part i: Area, Houses, and Population. HM Stationery Office [Cd. 847].
British Parliamentary Papers. (1912a). Emigration statistics of Ireland for the year 1911. HM Stationery Office. [Cd. 6131].
British Parliamentary Papers. (1912b). Forty-eighth detailed annual report of the Registrar General for Ireland containing a general abstract of the numbers of marriages, births and deaths registered in Ireland during the year 1911. HM Stationery Office. [Cd. 6313].
British Parliamentary Papers. (1913a). Census of Ireland, 1911. General report with tables and appendix. HM Stationery Office, Dublin [Cd. 6663].
British Parliamentary Papers. (1913b). Emigration statistics of Ireland for the year 1912. HM Stationery Office. [Cd. 6727].
British Parliamentary Papers. (1913c). Forty-ninth detailed annual report of the Registrar General for Ireland containing a general abstract of the numbers of marriages, births and deaths registered in Ireland during the year 1912. HM Stationery Office. [Cd. 6917].
British Parliamentary Papers. (1914a). Emigration statistics of Ireland for the year 1913. HM Stationery Office. [Cd. 7313].
British Parliamentary Papers. (1914b). Fiftieth detailed annual report of the Registrar General for Ireland containing a general abstract of the numbers of marriages, births and deaths registered in Ireland during the year 1913. HM Stationery Office. [Cd. 7528].
British Parliamentary Papers. (1915a). Emigration statistics of Ireland for the year 1914. HM Stationery Office. [Cd. 7883].
British Parliamentary Papers. (1915b). Fifty-first detailed annual report of the Registrar General for Ireland containing a general abstract of the numbers of marriages, births and deaths registered in Ireland during the year 1914. HM Stationery Office. [Cd. 7991].
British Parliamentary Papers. (1916a). Emigration statistics of Ireland for the year 1915. HM Stationery Office. [Cd. 8230].
British Parliamentary Papers. (1916b). Statement giving the particulars regarding men of military age in Ireland. HM Stationery Office. [Cd. 8390].
British Parliamentary Papers. (1916c). Fifty-second detailed annual report of the Registrar General for Ireland containing a general abstract of the numbers of marriages, births and deaths registered in Ireland during the year 1915. HM Stationery Office. [Cd. 8416].
British Parliamentary Papers. (1917a). Emigration statistics of Ireland for the year 1916. HM Stationery Office. [Cd. 8520].
British Parliamentary Papers. (1917b). Fifty-third detailed annual report of the Registrar General for Ireland containing a general abstract of the numbers of marriages, births and deaths registered in Ireland during the year 1916. HM Stationery Office. [Cd. 8647].
British Parliamentary Papers. (1918a). Emigration statistics of Ireland for the year 1917. HM Stationery Office. [Cd. 9013].
British Parliamentary Papers. (1918b). Fifty-fourth detailed annual report of the Registrar General for Ireland containing a general abstract of the numbers of marriages, births and deaths registered in Ireland during the year 1917. HM Stationery Office. [Cd. 9123].
British Parliamentary Papers. (1919a). Emigration statistics of Ireland for the year 1918. HM Stationery Office. [Cmd. 77].
British Parliamentary Papers. (1919b). Fifty-fifth detailed annual report of the Registrar General for Ireland containing a general abstract of the numbers of marriages, births and deaths registered in Ireland during the year 1918. HM Stationery Office. [Cmd. 450].
British Parliamentary Papers. (1920a). Emigration statistics of Ireland for the year 1919. HM Stationery Office. [Cmd. 721].
British Parliamentary Papers. (1920b). Fifty-sixth detailed annual report of the Registrar General for Ireland containing a general abstract of the numbers of marriages, births and deaths registered in Ireland during the year 1919. HM Stationery Office. [Cmd. 997].
British Parliamentary Papers. (1921a). General annual reports on the British army for the period from 1 October 1913, to 30 September 1919. HM Stationery Office. [Cmd. 1193].
British Parliamentary Papers, 1921b. Fifty-seventh detailed annual report of the Registrar General for Ireland containing a general abstract of the numbers of marriages, births and deaths registered in Ireland during the year 1920. HM Stationery Office. [Cmd. 1532].
Department of Social Welfare. (1955). Report of Commission on emigration and other population problems (pp. 1948–1954 R. 84). Stationery Office. (Pr. 2541).
Government of Northern Ireland. (1929). Census of population of Northern Ireland 1926: General report. HM Stationery Office.
Roinn Tionscail agus Tráchtála. (1928). Census of population 1926 Vol. 1: Population, area and valuation of each district electoral division and of each large unit of area. The Stationery Office.
Roinn Tionscail agus Tráchtála. (1929). Census of population 1926 Vol. v, Part i: Ages, orphanhood and conjugal conditions. The Stationery Office.
War Office. (1922). Statistics of the military effort of the British Empire during the Great War, 1914–1920. HM Stationery Office.
Secondary Literature
Akenson, D. H. (1993). The Irish diaspora: A primer. P. D. Meany.
Bowman, T., Butler, W., & Wheatley, M. (2020). The disparity of sacrifice: Irish recruitment to the British armed forces, 1914–1918. Liverpool University Press.
Budd, J., & Guinnane, T. W. (1991). Intentional age-misreporting, age-heaping, and the 1908 old age pensions act in Ireland. Population Studies (Camb), 45(3), 497–518.
Colvin, C. L., & McLaughlin, E. (2021). Death, demography and the denominator: Age-adjusted Influenza-18 mortality in Ireland. Economics and Human Biology, 41, Article 100984. www.doi.org/10.1016/j.ehb.2021.100984.
Dunnigan, D. R. (2012). Irish return migration from America at the turn of the nineteenth century, 1890–1920. Unpublished PhD dissertation, National University of Ireland Maynooth. https://mural.maynoothuniversity.ie/4735/.
Fernihough, A., & Ó Gráda, C. (2019). Across the sea to Ireland: Return Atlantic migration before the First World War. ucd School of Economics Working Paper Series, University College Dublin. http://dx.doi.org/10.2139/ssrn.3491676.
Fitzpatrick, D. (1978). The geography of Irish nationalism 1910–1921. Past & Present, 78, 113–144.
Fitzpatrick, D. (1995). The logic of collective sacrifice: Ireland and the British army, 1914–1918. The Historical Journal, 38(4), 1017–1030.
Gallagher, N. (2019). Ireland and the Great War: A social and political history. Bloomsbury Publishing.
Graham, B. J., & Proudfoot, L. J. (Eds.). (1993). An historical geography of Ireland. Academic Press.
Guinnane, T. W. (1997). The vanishing Irish: Households, migrants, and the rural economy in Ireland, 1850–1914. Princeton University Press.
Hart, P. (1997). The geography of revolution in Ireland 1917–1923. Past & Present, 155, 142–176.
Lynch, R. J. (2019). The partition of Ireland: 1918–1925. Cambridge University Press.
Pennell, C. (2012). A kingdom united: Popular responses to the outbreak of the First World War in Britain and Ireland. Oxford University Press.
Smith, S. K., Tayman, J., & Swanson, D. A. (2002). Overview of the cohort-component method. State and Local Population Projections: Methodology and Analysis (pp. 43–48).
Vaughan, W. E., & Fitzpatrick, A. J. (Eds.). (1978). Irish historical statistics: Population, 1821–1971. Royal Irish Academy.
Walsh, B. M. (1970). Marriage rates and population pressure: Ireland, 1871 and 1911. The Economic History Review, 23(1), 148–162.
Ward, A. J. (1974). Lloyd George and the 1918 Irish conscription crisis. The Historical Journal, 17(1), 107–129.
Weber, H. (2019). Age structure and political violence: a re-assessment of the ‘youth bulge’ hypothesis. International Interactions, 45(1), 80–112.