Eric W. Sager, Peter Baskerville, eds. Household Counts: Canadian Households and Families in 1901. Toronto: University of Toronto Press, 2007. 486 pp. $45.00 (paper), ISBN 978-0-8020-3802-9; $85.00 (cloth), ISBN 978-0-8020-3860-9.
Reviewed by Francoise Noel (Department of History, Nipissing University)
Published on H-Canada (June, 2008)
Counting More Than Canadian Families
Household Counts, edited by Eric W. Sager and Peter Baskerville, brings together a number of studies by members of the Canadian Families Project (CFP), active between 1996 and 2001. The project created a 5 percent random sample of dwellings of the 1901 census in Canada using both Schedule 1 and Schedule 2. It, therefore, combines information on living persons and property enumeration. The database was created to be compatible with the American Integrated Public Use Microdata Series (IPUMS), allowing for comparisons between the two countries. While numerous publications by the individual members of the project based on this data have appeared in the past, including an important special issue of Historical Methods: A Journal of Quantitative and Interdisciplinary History (fall 2000), this is the first collection of essays based on this project to appear. Most of the essays are based primarily on the CFP 1901 database, but some extend their study across time, one makes a comparison with American data, and others use it only as context for other research. While the essays share a focus on the study of family history, the wide range of topics covered demonstrates that census data as "a social survey" can be used to illuminate questions relating to religion, gender, race, ethnicity, class, and region, as well as family and household.
The book is divided into five sections. The first provides a national demographic context. The second through fourth sections explore family geography and social history, and the fifth part examines the cultural context of the 1901 census. The essays share a perspective that places the emphasis on the individual rather than the household, and that sees "families and households as fluid and transitional microsocial communities" (p. 9). Although the nuclear family household was the most common form overall in 1901, individuals lived in several different types of family or household structures in the course of their lives; and historians, sociologists, and demographers are able to extract much information from the census that speaks to the diversity of society rather than the norm.
The three essays in part 1 provide a national perspective on household and family structure, fertility, and family geography. Stacie Burke provides context for the demographic structure of household and family in 1901 by comparing it with 1991. The typology of household structures and the definition of census family used by the 1991 census, although more restrictive, is used to compare the two sets of data. As a preliminary basis for comparison, Burke provides an overview of the structure of the population at these two points, showing the effects of the demographic transition quite clearly. In 1901, there were many children and few people over the age of fifty-five in the population; in 1991, there were fewer children and a much larger proportion of adults over sixty-five years of age. The comparison of household types in 1901 and 1991 shows that, overall, the family was more central to households in 1901 than in 1991 but much more likely to include additional persons. There were considerably more husband/wife households in 1991 and many more people who lived completely alone. The essay also examines who the additional persons were in 1901; the geographical distribution of household types by province; households through the life course; and the likelihood of alternative types of households based on geography, rural/urban, occupation, age, sex, and crowding. In 1901, 63 percent of households included a nuclear family, whereas in 1991 only 38 percent of households did. In 1901, families were more open to having others in the household, whereas the focus on privacy in 1991 has led to fewer of these types of households. Less time was spent as a member of a nuclear family in 1991 than in 1901. While the nuclear family remained important in 1991, other household forms were clearly much more so than in 1901. For this study, data on households of more than thirty persons were eliminated because these would be institutions rather than households. Thus, while this provides an overview of the changing structure of households, it is not possible to answer whether or not the older population was more likely to be living in institutions in 1991 than in 1901.
Peter Gossage and Danielle Gauvreau examine the diversity of the Canadian population in terms of their reproductive behavior, first geographically, at both the provincial and the census district level; and secondly based on certain key socioeconomic factors, such as religion, mother tongue, place of birth, and rural/urban distinctions. The child-women ratio was calculated for all married women between the ages of fifteen and forty-nine for the entire sample. This was then mapped according to census district. While this confirms the high fertility rates in Quebec, especially rural Quebec, the authors argue that to focus exclusively on the "two solitudes" would mask other significant diversities, which are also evident from this exercise. The very low fertility rates in southern Ontario are also evident. The remainder of the country, however, shows many variations. Fertility is relatively high in much of the Prairie Provinces and parts of New Brunswick and Nova Scotia. British Columbia, in contrast, is a low fertility area. The latter difference, they argue, is as interesting as that between French and English Canada. Exploring the impact of several different factors on fertility ratios more thoroughly, the authors conclude that we are unlikely to discover a model that would explain the fertility transition across societies, as regional variations cannot be explained using the same model, even within Canada. The CFP 5 percent sample, however, is an effective tool for exploring Canadian diversity at the turn of the twentieth century.
In their study, Larry McCann, Ian Biuck, and Ole Heggen examine the demographic characteristics of Canadian families from a geographic perspective. This is a cartographic essay in which basic information about families is mapped by census district. The authors include average family size overall. For rural and urban families, they examine the percentage of nuclear, extended, and solitary families, as well as fertility ratios, percentage widowed, and dependency ratios (dependency on labor of family members over the age of fourteen). These maps establish visually many of the patterns discussed in the previous two essays.
Part 2 examines urban families. Two essays, by McCann, Biuck, and Heggen, continue the cartographic presentation begun in chapter 3. The third contribution, by Kenneth Sylvester, focuses on rural to urban migration. In chapter 4, McCann, Biuck, and Heggen examine characteristics of families that lived in urban places of at least twenty-five hundred people. These were much more numerous in eastern Canada than on the prairies and in British Columbia. With few exceptions, in western Canada, however, the patterns of family size, type, fertility, and work strategies was remarkably similar and did not exhibit variations based on a core-periphery differentiation. In chapter 6, the authors continue their cartographic display of family characteristics, moving their discussion to Montreal and its suburban municipalities, Canada's metropolis in 1901. Using Amalgamated Census Areas to have at least thirty families in each area (and over sampling where necessary), the authors display information with considerable spatial precision. While there appears to be a sharp demarcation between the central core and the suburbs on some variables, such as cohabiting with a lodger or a boarder, other variables, such as average family size, vary more with ethnicity. The methods used here bring greater precision to models of urban social zones than previously possible, and they could be applied to other urban places as well.
In chapter 5, Sylvester examines family change in the context of rural to urban migration. He begins with an investigation, on a national scale, of the census question relating to whether or not people were rural or urban born. By far the largest proportion of native-born Canadians was persisters--in both the urban and rural setting. Regionally, there were far more urban to rural movers in the west than elsewhere and far fewer in Quebec. There were slightly fewer rural to urban movers in the Maritimes than in other regions. Looking at family living arrangements in rural Canada, Sylvester finds that migrants were more likely to live in complex families than persisters, and that complex households were almost twice as likely in the west and half as likely in Quebec than in Ontario. For urban dwellers, city size did not have an impact on family types. Individuals with professional and clerical occupations were more likely to live in complex families than general laborers. The data show that while rural migrants might have moved to the city seeking employment and household independence, the latter was not easily achieved. Overall, rural to urban migration remained modest in 1901. In this chapter, the CFP data is used to test various aspects of modernization theory. The results are not all as expected, suggesting the utility of using this data to test models in the Canadian context.
Focusing on young and old populations, part 3 includes essays by Gordon Darroch, Bettina Bradbury, and Lisa Dillon. Darroch is interested in the transition to adulthood and the extent to which children continued, in the late nineteenth and early twentieth centuries, to live in households other than their parents'. This phenomenon among youth over the age of fifteen, seen as boarders and lodgers, has been studied to a greater extent than the fostering, apprenticeship, and service of younger children. According to the author, the number of youth and children involved, however, challenges conventional histories of family and childhood. The nature of growing up is placed in the context of the rise of the liberal order in Canada and the idealized nuclear family. The study is based on the Darroch and Michael Ornstein 1871 census sample and the 1901 CFP sample. Between 5 and 7 percent of children ten to fourteen years of age lived in a household other than their parents in both 1871 and 1901. Not surprisingly, the number increased considerably for older age groups. Of those single individuals living away from home, about one-quarter were related to the head of the household in which they lived. Another 11 percent in 1871 and 13 percent in 1901 were heads of household. The number of domestic servants increased from 13 percent in 1871 to 22 percent in 1901. The remainder was unrelated others: 51 percent in 1871 and 40 percent in 1901. Darroch notes in particular two groups that have been understudied: single women over thirty years old who headed their own households and women who boarded and lodged. Fostering children in homes of relatives had declined but certainly not disappeared by 1901. Domestic service was age related, peaking in the late teens and early twenties. More females (up to 50 percent) than males (up to 36 percent) were in service. The large numbers in service "flew in the face of the celebration of individual autonomy" (p. 214). The practice of boarding and lodging children to send them to school also continued. Darroch concludes that the "implantation of a liberal cultural formation" was hampered by this continuation of informal fostering, boarding, and lodging and the increased need for domestic service (p. 236).
Bradbury examines children who lived with only one parent in 1901. Based on the CFP sample figures, she estimates that there were about one hundred thousand in the fifteen years or younger age group across the country. These children have not received much attention. She uses the CFP database to explore how their lives were different from children in two parent families. They were a diverse group and existed in every part of the country and among every racial and ethnic group. Almost one-quarter lived with one parent who was listed as married. They were more likely to be living with their mother than their father, and those living with their mother were more likely to be raised in poverty. All children in single parent families were also more likely to leave school early and to live in a home that included relatives, other families, or boarders. Ethnicity and race, religion, and class affected their lives as well. Their history demonstrates that the prevalence of single parent families has a long history in Canada.
Dillon is interested in the stages of old age and compares these boundaries between men in the United States and Canada using 1870/1871 and 1900/1901 census data. Young-old age is characterized by a continuation of mid-life patterns, such as living with dependent children and economic independence, usually between the ages of fifty-five and sixty-five or seventy. Most men then experienced a transitional period in which they began to lose their independence. She looks first at questions relating to household head status, owning or renting, listing an occupation, living with a spouse, and living with children. Dillon examines the likelihood of holding household head status based on independent socioeconomic variables, including region, age, occupation, and presence of a spouse. Those with economic advantages, farmers, and those with a spouse were more likely to retain household head status longer. Men in their seventies experienced a decade of transition, which suggests that old age was more diverse in the late nineteenth and early twentieth centuries than today.
Part 4, aptly referred to as "New Interpretations: Family and Social History," includes essays by Sager, Lynn Marks, and Baskerville, which deal with the working class, religion or the lack thereof, and midwives. Sager examines the Canadian working class in 1901, particularly in terms of earnings. The 1901 census question on wages provides a unique opportunity to study this issue for 1901. Sager begins with two hypotheses: that inequality of earnings was greater in urban places and that earnings varied with stages of the family cycle, even when other factors were taken into account. He finds that wages were unequal nationally and provincially, with no significant differences between urban and rural areas. The greatest inequality was for clerical work where women workers were numerous. Based on a regression analysis of many different factors contributing to per-person family earnings, he finds that family type mattered a great deal. As he points out, however, the fragility of earnings based on stages of the life cycle does not negate the importance of occupation and class. "The point is that among working-class families large proportions were vulnerable to poverty, and vulnerability increased as one moved through the family cycle" (p. 359). While sex, occupation classification, and ethnicity all contributed to the fracturing of the working class in 1901, Sager concludes that inequalities in earnings "were far more extreme than in our time, and so extreme as to be fundamental to all divisions within and between social classes" (p. 360). One of the interesting details in this essay is table 10.7, which shows the occupation and earnings hierarchy in Ontario manufacturing and transportation sectors. It uses the skill designations given in the census, such as manager, master, superintendent, foreman, hand, and assistant, to divide the workforce into five categories. Mean age and mean annual earnings were calculated for each group. In the top category, earnings averaged $1,168 compared to only $184 in the lowest. The mean age was 41.3 compared to 20.2. The way in which CFP data were coded makes this possible and would certainly have application in other contexts as well.
Marks looks at "unbelief" in British Columbia. In 1901, 1.5 percent of the population of British Columbia stated that they were atheists, agnostics, or had no religion. This was about ten times more than in Canada overall where the number was only 0.16 percent (p. 373). Using the CFP database, Marks provides an overview of Canadian church participation by mapping the percentage difference between the number of communicants and total adherents relative to the Canadian average. Most census districts in the north and west were below average, and British Columbia had the lowest level of participation for Protestants in the country. This lower participation held true for Catholics as well. Religious periodicals tended to see the overrepresentation of young men in the population as the major reason for this discrepancy. By using a sample of atheists in Ontario for comparison, Marks shows that while men were more likely than women to be atheists, they were ten times more likely to be atheists in British Columbia than in Ontario. Women in British Columbia were six times more likely to be atheists than women in Ontario; this applied to married as well as single persons. The working class was more atheist than the middle class. Miners were particularly overrepresented, especially in the Kootenays. This, however, cannot be accounted for just on the basis of the overrepresentation of young single men in that group as married men and women in these areas were also atheists. Marks suggests that the culture of labor radicalism helped support the choice to publicly identify as atheist. Ethnicity was also likely a factor. By looking at the ethnicity of atheists as compared to the total population, Marks finds that several groups were overrepresented: Finns/Swedes, Germans, and Americans. Of these, Americans were the largest group in the population and therefore contributed the greatest number of atheists. Among the British, there was no over- or under-representation. Marks concludes: "I hope that this chapter has demonstrated that in trying to understand the nature of radicalism among BC workers it would be useful for labour historians to explore questions of religion and of religious belief" (p. 397). British Columbia also stands out as the first and the most secular province in the country. "Canadian historians of religion and secularization cannot just dismiss the B.C. experience as a strange exception to Canadian patterns," Marks claims (p. 397). By studying how and why this pattern first emerged here, she argues, one might shed light on the process elsewhere in Canada.
Baskerville examines the place of midwives in the medical marketplace of Victoria from 1880 to 1901 using data from birth registrations. This microhistory approach allows him to test the general statements often made about the decline of midwifery at the turn of the century. In the period, births attended by midwives declined from 30 percent of births to 14 percent. While the literature on the decline of midwifery suggests that the Canadianization of the population accounts for the decline, Baskerville uses data from birth registrations and the census to show that birthplace alone does not account for the decline. Occupational status is more closely linked to the decline than place of birth. In looking at religion, he finds that the use of midwives was almost entirely limited to Protestants. This finding cannot be explained from the sources. Other findings show that midwives sometimes attended births with doctors, and that clients often changed midwives or moved between midwives and doctors from one birth to another. These findings suggest that the decline in the use of midwives was more complex than the general literature suggests. While Baskerville bases his study primarily on birth registration information, he shows the utility of having the census database available for context.
Part 5, "The Importance of Cultural Context," includes a contribution by Chad Gaffield on constructing identity and one by Annalee Lepp on constructing normality. Questions on language were added to the Canadian decennial census surveys in 1901. The question was asked using three columns: "mother tongue," "English spoken," and "French spoken." This structure did not allow for more than one mother tongue. Gaffield has previously shown, based on the CFP data, that the language of children was usually assumed to be that of the father. In this study, he examines the discussion in political and journalistic sources around the introduction of this question in the census. The desire to categorize the population on the basis of "origin" went back to the 1871 census, but how to obtain the desired results was less clear. The 1891 census, in particular, which had used birthplace of parents and a special column for French Canadians, had resulted in many persons of "French" origin being enumerated as "English" in origin. This explains, in part, the addition of the question on language in 1901. The question on nationality was to counter the argument that persons should be allowed to call themselves "Canadian" if they so chose. Despite the intentions, however, the language question did not simplify the question of personal identity. Gaffield points out that instead of creating the "officially anticipated clarity," it "added another layer of complexity" (p. 437). Officials hoped that the language question would allow a count of French Canadians outside of Quebec, but instead it "revealed the complex, competing, and contested concepts of individual and ancestral identities across Canada in the decades following Confederation" (p. 437). While this clarity was not forthcoming, Gaffield concludes, language was increasingly used in public debate at the turn of the century. When the results of the 1901 census showed the spread of French in New Brunswick and up the Ottawa Valley, one journalist worried about this trend and wondered what could be done about it. As Gaffield remarks, Ontario responded by the introduction of Regulation 17 a few years later. This is a fascinating essay that brings the study of the census into a different realm, that of identity. It sheds useful light on the language question, which was changed again slightly in 1911. It should be noted that while the book is remarkably free of typographical errors, the reference to the 1901 census on page 426 should probably be to the 1891 census as it relates to the under-enumeration of French Canadians.
Lepp explores the ideological climate of the turn of the century and asks whether or not the 1901 census "promoted a particular version or official representation of social reality up to and including marital, familial, and household arrangements" (p. 445). She does so by looking at the "implicit assumptions and exclusions" found in the prescriptive categories of the census (p. 445). Looking more particularly at the question on conjugal condition, Lepp compares the number of divorced persons listed in the published census reports with those listed in the manuscript census. Since the CFP sample was inadequate for this purpose, information from the manuscript census on all divorced persons was used instead. This showed a total of 1,215 divorcees compared to only 661 in the published report. The lack of correspondence between the two is particularly notable in Quebec, where the official report indicated only 58 divorces but the manuscript census showed 368. Officials had made a conscious effort to minimize the numbers of those categorized as divorced by asking enumerators not to include those 'separate as to bed and board' in that category. Although there is no evidence of how the re-tabulation was done, it appears that cases of separation indicated as divorced in the manuscript census were changed to married for the published report, particularly for Catholics living in Quebec. The categories provided by the census caused difficulties for some women, as a small number indicated their relationship to the head of household as "wife" yet indicated they were divorced. Official data were, therefore, shaped by dominant ideological notions of marriage and family, but "individuals may have resisted classifying their marital identities within the confines of prevailing state and religious ideologies and the organizational categories of the census schedules" (p. 465).
Household Counts adds immeasurably to our knowledge of the family and understanding of the diversity of Canadian society at the turn of the twentieth century. Findings based on the national sample confirm the different stages of population and development in the various regions of Canada in 1901. Many of the differences in household structure, fertility, and other variables relate to this diversity. While regional differences in the many essays that use regression analysis tend to be at the level of the province, those studies that map national results by census districts allow us to visualize not only differences between the longer settled and more industrialized east and the more recently settled west, but also many similarities that might appear between regions smaller than the provincial boundaries. Sections of New Brunswick, Manitoba, and Quebec, where the concentrations of French Catholics was high, for example, display similarities for some variables. The northern sections of Ontario and Quebec often resemble the prairies and British Columbia more than they do the southern sections of the same provinces. Such results challenge our traditional ways of seeing national differences, which are so much more complex than the often cited dichotomy between French and English Canada suggests. The particularity of British Columbia on several fronts emerges most clearly.
Many of these studies identify issues that deserve further study. Darroch, for example, found an unexpected number of women lodging and boarding and single women over thirty who headed their own households. Although this was not the direct focus of his research, this finding might stimulate more research in this area. The methodological approaches the authors use here have done much to minimize the limitations of the census as a cross-sectional snapshot of society. That family and household types were not static and that individuals lived in a wide variety of family forms is evident. The range of questions that can be addressed using census data also emerges clearly from this collection.
Overall, Household Counts is a successful volume that will be of interest to researchers in many areas of social and cultural history, including those interested in family, childhood, youth, gender, the working class, religion, and identity. It includes valuable information for both upper level undergraduates and graduate students. I hope that this volume will stimulate more use of the CFP database, not just by CFP researchers but also by others. On that note, it would have been useful if information on where to access the database had been included, perhaps in the introduction. When I tried to download it several years ago, there were errors in the variable list which had to be corrected before I could download properly. I hope that this has been corrected and that there is now a site making it readily available. Since H-Net encourages responses to reviews, it would also be interesting to hear from list members if and how they are using Household Counts and other essays based on the CFP database in their teaching, and whether or not the availability of the CFP database for public use is sufficiently well known that researchers not originally associated with the project are using it. It will only serve the original purpose of the project if it becomes widely used as a tool for researchers beyond the original project participants. Household Counts should help to make its availability and its potential more widely known.
If there is additional discussion of this review, you may access it through the list discussion logs at: http://h-net.msu.edu/cgi-bin/logbrowse.pl.
Francoise Noel. Review of Sager, Eric W.; Baskerville, Peter, eds., Household Counts: Canadian Households and Families in 1901.
H-Canada, H-Net Reviews.
Copyright © 2008 by H-Net, all rights reserved. H-Net permits the redistribution and reprinting of this work for nonprofit, educational purposes, with full and accurate attribution to the author, web location, date of publication, originating list, and H-Net: Humanities & Social Sciences Online. For any other proposed use, contact the Reviews editorial staff at firstname.lastname@example.org.