New Research Opportunities Using Historical Census Micro-Data in the RDCs with Danielle Gauvreau

Transcription

1 CRDCN webinar series New Research Opportunities Using Historical Census Micro-Data in the RDCs with Danielle Gauvreau January 21,

2 The Canadian Research Data Centre Network 1) Improve access to Statistics Canada detailed microdata, including an increasing range of surveys, census and administrative data. 2) Expand the pool of skilled quantitative researchers 3) Make research count Visit our website: Follow us on Twitter 2

3 Danielle Gauvreau is professor in the Department of Sociology and Anthropology at Concordia University and the Academic Director of the Quebec Interuniversity Centre for Social Statistics (QICSS). She is the author of several scientific publications on demographic changes in Quebec, on fertility and marriage patterns in particular. Her presentation today is based on one of her current research projects on the causes of the baby boom in Quebec. 3

4 New Research Opportunities Using Historical Census Micro-Data in the RDCs Dr. Danielle Gauvreau Concordia University, Quebec Interuniversity Center for Social Statistics, and Centre interuniversitaire d études québécoises CRDCN Webinar, January

5 Outline of the presentation 1. Census data and historical research in Canada 2. Recent work and ongoing research 3. The Baby Boom that was a marriage boom: the Quebec experience, (Example I) 4. Trends and determinants of mixed marriages in Quebec, (Example II) 5. Concluding thoughts 6. Acknowledgements & references

6 1. Census data and historical research in Canada Historical censuses are important for research because they allow us: To know ourselves better and understand where we come from To measures change and develop a better understanding of how it unfolded and why it happened To establish comparisons with recent or contemporary situations and draw parallels with what happened before Access to micro-level data from historical censuses today opens up new research possibilities that go beyond what has been done when the data was first published: New research questions, based on new paradigms Varying analytical scales, for example, from individuals to families and communities, or at different geographic scales New and more powerful statistical techniques such as logistic regression, multilevel models, life transitions models Construction of long-term series based on similar concepts Detailed comparisons with other countries where similar data is available

7 Census data and historical research in Canada What s in the census Religion (1852 ) Birth place (1852 ) Ethnic origin (1871 ) School attendance (1871 ) Occupation (1871 ) Relationship to head (1891 ) Wages (1901 ) Linguistic skills (1901 ) Year of immigration (1901 ) Number of children ever born ( ) Age at marriage (1941 ) and a lot more Examples of research based on census data Bettina Bradbury, Working families: Age, Gender and Daily Survival in Industrializing Montreal, 1993 Peter Baskerville and Eric Sager, eds., Household counts: Canadian households and families in 1901, 2007 Valérie Laflamme, Vivre en ville et prendre pension à Québec aux XIXème et XXème siècles, 2007 Gauvreau, Danielle, Diane Gervais and Peter Gossage, La fécondité des Québécoises : d'une exception à l'autre, 2007 Lisa Dillon, The Shady Side of Fifty: Age and Old Age in Late Victorian Canada and the United States, 2008 Sherry Olson and Patricia Thornton, Peopling the North American City, Montreal , 2011 Gordon Darroch ed., The Dawn of Canada's Century: Hidden Histories, 2014

8 Census data and historical research Census data are available in different forms. Beyond the actual images of the census, very useful to genealogists for example, there are different ways to get aggregate statistical information from the census: 1. Published material All censuses 2. Publicly available microdata files 3. Confidential micro-data files available in secure locations only (Research Data Centres or RDCs) Most censuses from 1852 to 1911 (1921 soon), which are in the public domain 1971 to 2011 censuses through Data Liberation Initiative (DLI) through 1951 (CCRI) through 2011 censuses (Statistics Canada)

9 Census data and historical research Where do these files come from? Individual scholars (Darroch & Ornstein, Dillon, Inwood, Sager & Baskerville, ) and teams of researchers (the CFP, the CCRI, the Mormons and NAPP) created these files, and the joint initiative of the Canadian Research Data Centre Network (CRDCN) and Statistics Canada made the more recent micro-data available in the Research Data Centres. Thanks to the research agencies who have been supporting these initiatives. A few words about the Canadian Century Research Initiative (CCRI) because many new research opportunities stem from the recent constitution of representative files for the 1911 to 1951 censuses. Three main elements: Representative samples (5%, 4%, 3%) of the 1911 through 1951 censuses GIS tools to study the spatial dimension of various issues (see Moldofsky) Contextual information pertaining to the context in which censuses were taken CRDCN ( and Statistics Canada

10 2. Recent work and ongoing research with historical micro-data files available in RDCs A few examples taken from the Statistics Canada website Study of the long run distribution of income in Canada, David Alan Green , 1921, 1931, 1941, 1951 Learning and earning: The characteristics of student part-time wage earners in the 1951 Census of Canada, Katharine Rollwagen , 1931, 1941, 1951 Projections du réseau familial élargi pour identifier les aidants potentiels des personnes âgées en perte d'autonomie en vue du maintien à domicile, Yann Décarie , 1951, 1981, 1991, GSS 1996, 2002, 2007, 2012 The residential crowding of immigrants in Canada, , Michael Haan And three more projects presented more extensively in the rest of this talk Lisa Dillon and Damaris Rose, Home beyond the nuclear family in Canadian Cities : gender & the changing living arrangements of the non-married. See next slides: thanks to Damaris! Danielle Gauvreau and Benoît Laplante, Le baby-boom québécois: étude de cas et perspective comparée Danielle Gauvreau, Patricia Thornton, and Hélène Vézina, Diversity, cross-cultural relations, integration: A socio-demographic and demo-genetic study of the Quebec population,

11 Home beyond the nuclear family in Canadian Cities : gender & the changing living arrangements of the non-married (SSHRC Insight grant, ) Damaris Rose, INRS; Lisa Dillon, U. Mtl With whom & in what housing arrangements did single city dwellers live in the little-known decades after the onset of individualization & mass urbanization (late C19) but before the onset of the massive rise in living alone (later C20)? Did economic, housing & cultural dimensions of early C20 metropolitan modernity contribute to rising residential independence among single (nevermarried) women? Urban geography Household demography Gender studies Conceptual framework: hierarchical model of residential autonomy in terms of BOTH relationship to household head AND housing/dwelling context (self-contained unit lodging house institutional)

12 Why we were drawn to the CCRI census samples Existing knowledge stems only from localized case studies thus non-comparable between cities; also, city directory & tax roll data usually only report household head CCRI can render visible small & unusual groups marginalized in published sources complete enumeration of members of private households & noninstitutional collective dwellings can document peer-group sharing (roommates), also boarders, roomers Over-sample (20%) of households in multiple-unit large dwellings can explore hypothesis re over-representation of single women office workers as heads/roommates in apartment buildings Over-sample (10%) of residents of single-unit large dwellings can document who lived in institutional settings/large group quarters

13 Methodological challenges & contributions 2 years (of 3-4 yr project) of data cleaning, creating new variables to operationalize our questions, e.g. hierarchy of singles residential autonomy comparable over time (changing census concepts & enumeration practices re household ) New geographical units of analysis: metropolitan areas; urban places by size class ( + CCRI s urban/rural definition needed correction) Many derived variables: demog., socioeconomic, ethnoreligious Using CCRI requires high level work on conceptual level (limited role for junior RAs) Trial & error, & ingenuity to get meaningful frequency tables respecting StatCan rules Some regressions require proportional weighting procedures Authors will place all their syntax (new variables, recodes) in the public domain upon completion of project

14 3. The Baby Boom that was a marriage boom: the Quebec experience, (Example I) The Baby Boom and Its Causes: What We Know and What We Need to Know (van Bavel and Reher, PDR 2013) We know little about the causes of the BB The BB is as much about marriage as about fertility (if not more) We need more studies using micro-level data to assess the contribution of various groups to the BB in various settings Studying the Baby Boom is likely to contribute to a better understanding of the demographic trends in Western countries during the mid-20 th Century. The Baby Boom was an unexpected phenomenon which can nonetheless be considered as a key moment between the 1 st and 2 nd demographic transitions. Our project makes use of the newly available historical micro-level data to address the concerns above, both from censuses (1941 and 1981 so far) and from retrospective surveys. So far, we worked on the marriage trends during the baby boom, and the next slides tell you why it is such an important question, particularly in Quebec. (Excerpts from a paper presented at the Social Science History Conference in Toronto, November 2014)

15 Number of births Birth rate in Number of births and birth rates, Quebec , , Birth rate Number of births 35,0 30,0 25,0 20,0 15,0 10, , Year 0,0

16 Quebec and Ontario fertility trends compared 5,0 Total fertility rate, Quebec and Ontario, ,5 4,0 3,5 3,0 2,5 2,0 1,5 Québec Ontario 1,0 0,5 0,

17

18 Marriage rate Number of marriages and marriage rates ( ), Quebec , Marriage rate 10,0 N of marriages ,0 6, ,0 N of marriages ,0 0 0, Year

19 Source and methods 1981 Census and the question on age at marriage Micro-level census data available in Research Data Centres in Canada Hazard models (Royston-Parmar) run separately for men and women to analyse the risk of marriage according to the characteristics below Independent variables Ethno-religious group Level of education Country of birth & age at immigration Place of residence

20 Excerpts - Survivor function of marriage for women by cohort for four ethno-religious groups, Quebec % single FC at BC at BP at J at FC at BC at BP at J at FC = French Catholic, BC = British Catholic, BP = British Protestant, J = Jewish

21 Hazard function of marriage for French Catholic women by cohort, Quebec Quebec French Catholic 0.20 Rate Age Hazard function from Royston-Parmar model

22 Excerpts - Survivor function of marriage for French-Catholic women by cohort for three levels of schooling, Quebec % single L at M at H at L at M at H at L = Low, M = Medium, H = High. The definition of schooling varies with the cohorts to account for the significant increase in schooling during the period.

23 Effect of immigration status and area of residence on the hazard of female marriage by education level net of cohort for some ethno-religious groups in Quebec. Canadian 1981 Census. Royston-Parmar flexible parametric hazard model. Relative hazards. All Low Medium High Quebec French Catholic Landed before age Landed at or after age * Rural *** *** *** *** Urban area less than 10, *** *** *** Urban area 10,000 to less than 100, *** *** *** *** Urban area 100,000 to less than 500, * * Quebec British Catholic Landed before age Landed at or after age Rural *** * *** *** Urban area less than 10, *** *** * *** Urban area 10,000 to less than 100, * ** Urban area 100,000 to less than 500, Quebec British Protestant Landed before age Landed at or after age *** *** ** Rural *** *** *** *** Urban area less than 10, ** * Urban area 10,000 to less than 100, *** * ** Urban area 100,000 to less than 500, * Immigration status. The reference category is Born in Canada. 2 Area of residence. The reference category is Urban area at least 500,000. * p<0.05; ** p<0.01; *** p<0.001

24 Summary of results All groups affected by the push towards more marriage and earlier marriage, although not to the same extent. Women experienced a much greater increase than men in their propensity to marry but both groups married at an earlier age. The most educated were most affected by the changes. But highly-educated women remained the least likely to marry and to marry early while their male counterparts became the most likely to marry and no longer suffered any delay in doing so. All ethno-religious communities displayed similar trends but British and French Catholic women, who at first were the least likely to marry were the most affected by the changes.

25 Discussion How to interpret these results? The gender difference in marriage patterns is intriguing Economic interpretation: post-depression, post-wwii Youth culture (The Dominion of Youth by Cynthia Comacchio) Link with Dillon and Rose research on residential patterns of young single women for cohorts born at the beginning of the 20 th Century The rise of a companionate model of marriage and the increasing importance of sexuality within marriage More work to come with the previous censuses, particularly the 1941 census which also includes a question on age at first marriage.

26 4. Trends and determinants of mixed marriages in Quebec, (Example II) Diversity, cross-cultural relations, integration: A socio-demographic and demogenetic study of the Quebec population, with Pat Thornton (Concordia) et Hélène Vézina (UQAC) Among the objectives: Study the opportunities and actual occurrences of contact between cultural communities, both from a geographical and a social perspective, and within households and families. The study of mixed marriages is one important dimension of this work. Outcomes Geography of Encounter: Immigration, ethnic diversity, and interethnic relations within Quebec, , in The Dawn of Canada s Century, edited by Gordon Darroch (2014). Marrying the other : Trends and determinants of culturally mixed marriages in Québec, forthcoming in Canadian Ethnic Studies

27 Mixed marriages Mixed marriages are viewed as a particularly intimate form of integration between two people of different cultural background The literature shows that endogamy is always the dominant mode of marriage patterns For a mixed marriage to take place, the following three conditions need to be met: There has to be opportunities for people to meet, both socially and spatially (structural factors) There must not be any formal barrier, either at the individual or the group level There must exist a certain individual willingness to mix with the other (cultural factors) In this study, mixed marriages are defined in reference to ethno-religious communities

28 Data & methods Data The entire 1881 census (Mormons and NAPP) The CCRI samples for the 1911 census, publicly available, and for the 1941 census, available in the RDCs Censuses contain rich ethno-cultural information Birthplace and religion since 1852 Ethnic origin since 1871 Linguistic characteristics since 1901 More importantly, the 1941 included new questions that will prove useful for this work: number of years of schooling, age at marriage and number of children ever born for women.

29 Benchmarks in history of immigration and diversity Waves of immigration and settlement French Regime until Aboriginal population and the French population. British colony after 1760: British and Loyalist immigration during the last decades of 18 th Century. Mostly English, Irish (Great Famine) and Scottish immigration during the 19th Century. Continued movement across the US border. Immigration becomes more diverse towards the end of the 19th Century. Jewish and other population from Eastern Europe, Italians, Chinese immigrants. Era of massive immigration during the first three decades of the 20 th Century. Main destination is Montreal. Internal movements contribute to mixing Periods of disruption ( and 1840s) saw massive movement (militia, migrant labourers, colonizers) Spilling out of French Canadians from the long settled seigneurial lands after 1840s Rural-urban migration (predominantly French Canadian) since the 1860s

30 Some aspects of cultural diversity (based on published material) Percentage All Quebec Montreal All Quebec Montreal All Quebec Montreal Immigrant 5,8% 15,9% 7,3% 21,2% 6,7% 14,9% Non French origins 21,1% 40,8% 19,3% 36,5% 19,1% 37,4% Other origins in % of non-french 8,9% 6,5% 22,6% 29,4% 28,9% 35,7% Total population 1,358, ,945 2,003, ,528 3,331,882 1,047,205

31 Distribution and mixing of the three major ethnic groups in Quebec by census subdivision in 1911 (CCRI spatial tools)

32 Ethno-religious group of spouses, 1911 sample (*the equivalent table for 1941 cannot be published) Husbands Wives FC BC OC BP OP J Total FC BC OC BP OP J Total

33 Intermarriage Exogamy rates by ethno-religious background Husbands Wives 60% 60% 50% 50% 40% 30% 20% 10% % 30% 20% 10% 0% FC BC OC BP OP J 0% FC BC OC BP OP J Composition of population Imbalanced sex ratios FC about 80% of the population BC overly female in 1881 BP about 10% OC overly male at all times BC declining from 6 to 4% Others increasing throughout the period

34 Multilevel logistic regressions of the probability to be part of a mixed marriage Ethno-religious group Ability to speak both French and English Age Birth place Years of schooling Habitat Diversity index Ethno-religious group and habitat Ethno-religious group and diversity Individual variables - FC + If can speak both F & E - If speak F only (women) + Among younger people - If born abroad + When more schooling Contextual variables + In urban and rural non farm + When more diversity Interactions Same + impact of habitat in urban & rural non farm and + impact of diversity for FC and BP only - Impact for the other groups The amount of variation explained at the contextual level varies from 14 to 21% and remains significant when individual variables are taken into account.

35 Complementary analyses French Catholics only Similar results (contextual variables, linguistic skills, habitat) but more direct assessment of structural factors with the log of the relative size of the FC group, which proves to be a very significant factor Number of years of schooling no longer significant and a few other minor differences Focusing on Montreal Similar results but contextual level not significant (ethno-religious group, linguistic skills, years of schooling) More intermarriage among FC when they live in more diverse milieux No impact of birth place

36 Discussion General stability of the results throughout the period (may be for different reasons though) Mixity compounds on itself Importance of structural variables: size and sex ratios + diversity level within contextual units. The context and specific history of each group matters Unique position of the British Catholics. The French Catholic and Jewish community contrasted The British Protestants: a Quebec minority but a majority within Canada Other groups behaviour affected by structural factors and their history of immigration

37 5. Concluding thoughts on the analytical potential of historical census data in the RDCs A range of research questions can be addressed in the long-term, over more than a century but be prepared for serious critique of the sources and a lot of work to make sure that you are dealing with comparable information across time! The newest developments have to do with our ability to study the mid-20 th Century and with new opportunities to connect with both the period before and after. Access to detailed micro-level data opens up new possibilities of research based on new paradigms and making use of the most powerful statistical tools that exist today A few final remarks How can we access the data in RDCs? QICSS 2015 one-week summer school on the use historical censuses (in French)

38 6. Acknowledgements & references Co-researchers and research assistants for the Baby Boom project : Benoît Laplante (INRS-UCS), Cindy Flick (INRS-UCS), and Meagan Wierda (Concordia University) Co-researchers and research assistants for the Cultural Diversity project: Pat Thornton (Concordia University), Hélène Vézina (UQAC), Sherry Olson (McGill University), Laurent Richard (Université Laval), Brian Armstrong (Concordia University), Anne Bourgeois (Université de Montréal), and et Marc-Antoine Côté-Marcil (Concordia University) Financial support

39 Acknowledgements The analyses presented as examples in this paper were conducted at the Quebec Interuniversity Centre for Social Statistics which is part of the Canadian Research Data Centre Network (CRDCN). The services and activities provided by the QICSS are made possible by the financial or in-kind support of the Social Sciences and Humanities Research Council (SSHRC), the Canadian Institutes of Health Research (CIHR), the Canada Foundation for Innovation (CFI), Statistics Canada, the Fonds de recherche du Québec - Société et culture (FRQSC) and the Quebec universities. The views expressed in this paper are those of the authors, and not necessarily those of the CRDCN or its partners.

40 Partial list of references Baskerville Peter and Eric Sager, eds., Household counts: Canadian households and families in 1901, Toronto, University of Toronto Press, 2007 Bradbury Bradbury, Working families: Age, Gender and Daily Survival in Industrializing Montreal, McLelland and Stewart,1993 Comacchio Cynthia, The Dominion of Youth: Adolescence and the Making of Modern Canada, , Wilfrid Laurier University Press, 2006 Curtis, Bruce, The Politics of Population, State Formation, Statistics and the Census of Canada, Toronto, University of Toronto Press, 2000 Gordon Darroch ed., The Dawn of Canada's Century: Hidden Histories, Montreal & Kingston, McGill-Queen s University Press, 2014 Lisa Dillon, The Shady Side of Fifty: Age and Old Age in Late Victorian Canada and the United States, Montreal & Kingston, McGill-Queen s University Press, 2008 Gaffield Chad, Conceptualizing and constructing the Canadian Century Research Infrastructure, Historical Methods: A Journal of Quantitative and Interdisciplinary History, 40, 2: Gauvreau, Danielle, Diane Gervais and Peter Gossage, La fécondité des Québécoises : d'une exception à l'autre, Montréal, Boréal, 2007

41 References (continued) Gauvreau Danielle et Patricia Thornton, Marrying the other : Trends and determinants of culturally mixed marriages in Quebec, , forthcoming in Canadian Ethnic Studies. Haan Michael, The residential crowding of immigrants in Canada, , Michael Haan, Journal of Ethnic and Migration Studies, 2011, 37: Henripin Jacques, Tendances et facteurs de la fécondité au Canada, Ottawa, Bureau fédéral de la statistique, Valérie Laflamme, Vivre en ville et prendre pension à Québec aux XIXème et XXème siècles, Paris, L Harmattan, 2007 Moldofsky Byron, Exploring Historical Geography Using Census Microdata: The Canadian Century Research Infrastructure (CCRI) Project, Chapter 13 in Historical GIS Research in Canada, Jennifer Bonnell and Marcel Fortin eds., Calgary, University of Calgary Press, 2014 Sherry Olson and Patricia Thornton, Peopling of the North American City, Montreal , Montreal & Kingston, McGill-Queen s University Press, 2011 Van Bavel, Jan and David S. Reher, The Baby Boom and Its Causes: What We Know and What We Need to Know, Population and Development Review, vol. 39, n o 2, 2013, p

42 To know more: All our webinars are available on You Tube: Visit our website to download our publications, subscribe to our newsletter, The Networker, and find more about social statistics in Canada: Acknowledgements: The services and activities provided by the CRDCN are made possible by the financial or in-kind support of the SSHRC, the CIHR, the CFI, Statistics Canada and participating universities which we gratefully acknowledge. Thank you for joining in!