A census is a count of a population in a specific region. In Canada, there are two types of censuses: the Census of Population and the Census of Agriculture. Both are conducted every five years by Statistics Canada, a department of the federal government. The larger of the two censuses, the Census of Population, gathers various demographic information, including where people live, as well as their age, sex, marital status and ethnic origin. This information is used by the government to establish electoral boundaries, to make federal transfer payments (money given to the provinces) and to monitor various social programs and policies (e.g. Canada Pension Plan, health care and education). In addition, the data is available to non-governmental organizations and to the general public; some older data is available to individuals interested in genealogical research.
The first colonial census in Canada was conducted between 1665 and 1666 by Jean Talon, intendant of New France, by order of Louis XIV. Each respondent’s name, age, sex, occupation, marital status and relationship to the head of the family were recorded. Talon did much of the data collection personally, visiting settlers throughout the colony.
DID YOU KNOW?
Talon’s first census revealed the population of New France to be 3,215, with nearly twice as many men as women (2,034 to 1,181). In order to grow the population, 900 unmarried women — known as Filles du Roi — arrived from France between 1665 and 1673. Following the 1671 census, Talon happily reported the results to the King. “The number of settlers has increased significantly as a result of the King’s decision to send young women from France,” he wrote. “In looking at the results of the 1671 Census, 700 children were born during the year.”
Including the 1665-66 enumeration, 98 censuses were conducted leading up to Canada’s first national census in 1871. Some of these documented particular regions, for example Newfoundland and Acadia, and later, Upper and Lower Canada, while others detailed the entire colony. Early censuses were primarily taken to assess the population’s capacity for taxes or military service. A count of the number of households indicated how many could be charged tax, while a count of the number of adult men indicated who could be drafted for military service. In addition to these standard population questions, some censuses asked different questions depending on the needs of the colony at the time. From 1667 on, for example, most censuses asked questions about agriculture, including number of cattle and lands cleared.
The last census under French rule was completed in 1739. When the British assumed control in 1763, both the quality and quantity of censuses declined. Only three (1765, 1784, 1790) were taken of the Province of Québec before the end of the century. Upper Canada had an annual census from 1824 to 1842; however, they were less frequent in Lower Canada, Nova Scotia and New Brunswick. In 1831, a census was taken in Assiniboia, marking the first time that an area that would become part of Western Canada was enumerated. Censuses in Assiniboia continued, about every three years between 1834 and 1856.
The Act of Union (1840) replaced the legislatures of Upper and Lower Canada and created a combined legislature for the new Province of Canada. The first census of the Province of Canada was conducted in 1841-42 for Canada East (currently Québec) and Canada West (currently Ontario). In 1847, the Act for taking the Census of this Province and Obtaining Statistical Information Therein, enacted by the Province of Canada legislature, provided for the censuses of 1851 and 1861 and legislated that a census occur every 10 years after that. The 1851 census included Canada East, Canada West, New Brunswick and Nova Scotia. Prince Edward Island was added to the list of enumerated regions in 1861.
Canada's First National Census
Canada’s first national census was completed in 1871, four years after Confederation. It compiled data for the country’s first four provinces: Nova Scotia, New Brunswick, Québec and Ontario. As with the colonial censuses, enumerators were sent out to ask households a series of questions, as opposed to individuals filling out the census forms themselves. A total of 211 questions were asked.
According to the British North America Act of 1867, representation in the federal House of Commons was to be based on population. Census data would indicate how many members of parliament should be elected from each region based on the number of people who lived there.
The first census of Canada as a country of 10 provinces and two territories (at the time, Nunavut was still a part of the Northwest Territories) occurred in 1951, with Newfoundland having joined Confederation in 1949.
Establishment of Statistics Canada
Continuing demand for population data led to the establishment of a permanent Census and Statistics Office, in 1905, under the Department of Agriculture (the Department was responsible for statistics, immigration and public health at various times during the decades after Confederation). Following recommendations of a review commission in 1912, the Office of Dominion Statistician was created in 1915. Afterwards, the government formed the Dominion Bureau of Statistics in 1918, and in 1971 changed its name to Statistics Canada.
Types of Censuses
Census of Population
Since 1971, the Census of Population has involved two components: a short-form census and a long-form census. Today, the short form collects data from about 75 per cent of the Canadian population, while the long form is distributed to approximately 25 per cent of households. The shorter questionnaire is limited to questions about age, sex, marital status and mother tongue. Significantly more detailed, the longer questionnaire includes all questions on the short form, as well as questions regarding ethnic origin, Indigenous ancestry, visible minorities, changes in residence, education, health, employment and the home the respondent lives in. Both questionnaires are mandatory and are collected every five years.
Census of Agriculture
In addition to the Census of Population, a Census of Agriculture is also taken every five years. It’s completed by farm operators, and includes questions about the farm’s location and size; types of crops grown; use of pesticides, irrigation and manure; livestock; and value of farming operations. As with the Census of Population, completing the Census of Agriculture is mandatory.
The Census of Agriculture began in 1871 along with Canada’s first national Census of Population. Like the population questionnaire, the Census of Agriculture was originally conducted every 10 years; however, rapid growth of the prairie provinces led to its completion every five years in Manitoba beginning in 1896, and in Alberta and Saskatchewan beginning in 1906. In 1956, the Census of Agriculture began to be conducted every five years in all Canadian provinces.
Census data are supplemented and updated through numerous ongoing surveys. Some are monthly (e.g. labour-force, manufacturing and natural gas distribution surveys), others are annual (e.g. oil pipeline, community health and maple products surveys) and some are less frequent (e.g. the Aboriginal Peoples Survey, conducted every five years).
In addition, Statistics Canada relies on the collection of many administrative records — such as vital statistics (births, deaths and marriages), individual and corporate tax statements, immigration, and health and crime reports — for defining survey samples, checking the accuracy of statistics, direct tabulation, and supplementing other data.
Cancellation and Reinstatement of the Long-Form Census
On 26 June 2010, the Conservative government, under the leadership of Stephen Harper, announced that the long-form census would not be used in 2011. It would be replaced by the National Household Survey (NHS), a questionnaire similar to the long form, but one that could be filled on a voluntary, rather than mandatory, basis. The NHS would be distributed to 30 per cent of households as opposed to 20 per cent (the long form sample size used during the 2006 census).
The government cited respecting the privacy wishes of Canadians as the reason for their decision, despite the established practice of removing all personal identifiers from census data. The decision resulted in public controversy, with widespread criticism from social scientists and others. In 2010, Munir Sheikh, head of Statistics Canada, resigned in protest. Critics noted that a voluntary survey would mean a decline in data quality, while an increased sample size would mean a rise in cost.
When the results of the 2011 NHS survey were tabulated, the response rate was 68.6 per cent, significantly less than the 93.5 per cent response rate garnered by the 2006 long-form questionnaire. In addition, the cost of administering the survey had increased by $22 million.
On 19 October 2015 a Liberal government led by Justin Trudeau replaced Harper’s government. Shortly thereafter, on 5 November 2015, the Liberals announced that the 2016 census would reinstate the long- form census. The 2016 census was conducted with procedures similar to the 2006 and early censuses.
How Often is the Census Collected?
Both the Census of Population and Census of Agriculture are collected every five years. This frequency began in 1951; prior to that both censuses were conducted every 10 years. Between 1951 and 1981, mid-decade censuses were shorter and less detailed. The 1986 census was the first mid-decade census to repeat most of the questions asked five years earlier.
Censuses in Canada have collected different data at different times. Some information has always been gathered (e.g. the number of people in a region and their sex). Other data categories have been added or removed, or in certain instances, have been broadened to include more options. Decisions as to which questions are included or excluded are based on feedback from data users and testing of new questions with the public. Some of the consequences of well-meaning attempts to improve the nation's data collection systems include breaks in a historical data series and lack of comparability of data due to changes in basic definitions and concepts. Below are examples of major components of the Canadian Census of Population and how they’ve shifted over time.
In 1665, Canada’s first colonial census collected information on marital status under the heading “conjugal condition,” a phrase not generally seen today. Respondents were asked if they were married, unmarried or widowed. These options remained the same until the 1901 census when “divorced” was added to the list of possible responses. In 1911, “legally separated” was added. By 1981, a long-term, heterosexual couple living in the same household could list their status as common-law; however, these were treated as “married” responses when statistics on marriage were generated. In 1991 this changed, and common-law relationships became a statistical category. It took another 10 years, on the 2001 census, for same-sex common-law partners to be able to note their relationships.
Enumerators asked respondents for their ancestral origin for the first time in 1871. Responses were grouped into 19 categories: African, Dutch, English, French, German, Greek, “Half-Breed,” Hindoo, Indian, Irish, Italian, Jewish, Russian/Polish, Scandinavian, Scotch, Spanish/Portuguese, Swiss, Welsh and “Various other origins.” With the exception of 1891, a question about ethnic origin has been included in each census, though its definition and wording has changed over time. Before 1981, for example, respondents were only asked for their paternal ancestry. In addition, even if a respondent gave multiple origins, only one of these was documented by Statistics Canada. From 1981 on, the number of ethnic origins that respondents were permitted to include gradually increased with each census, as did the number documented. On the 2016 long-form census, respondents were invited to list as many origins as they liked.
Data on ethnic origin faces two difficult challenges. One is the need to revise and expand the list of ethnic origins as the Canadian population becomes more diverse. The other is recognizing the fluidity and ambiguity of ethnic origin for many individuals, stemming from changing identities and increased ethnic intermarriage. The breakup of Yugoslavia in the 1990s, for example, affected the ethnic identities of immigrants from this area: some may have thought of themselves as Yugoslavian, while others may have identified as Serbian, Croatian, Bosnian or another identity. For children of a Yugoslavian intermarriage, the ambiguity of what to report as their ethnic origin may have been further complicated.
Although self-reporting of ethnic origin is a great improvement over early methods that assumed ancestry determined ethnic origin — or that an individual necessarily report the same ethnic origin over their lifetime — the ethnic-origin question on Canadian censuses is not a perfect tool to measure the country’s diversity. Respondents’ answers depend on a variety of factors, including knowledge of their ancestry and the personal ways in which they define their ethnicity. In addition, because of the changing way in which the question has been asked over various censuses, it’s difficult to compare data over time.
A question about visible minorities was first asked on the 1996 census. The data gathered from the question informs programs and policies that aim to boost employment levels among visible minorities. The Employment Equity Act defines visible minorities as "persons, other than Aboriginal peoples, who are non-Caucasian in race or non-white in colour." In 2016, respondents were invited to circle as many minorities as applicable from the following list: South Asian, Chinese, Black, Filipino, Latin American, Arab, Southeast Asian, West Asian, Korean, Japanese, or “other.”
Although Canada collects census data on religion, other countries have struggled with the inclusion of this question. The United States, for example, reported data on “religious bodies” every 10 years from 1906 to 1936, asking ministers, priests, rabbis and other persons of authority about the membership of their local religious group; however, the U.S. government has never collected census data on religion from individuals. The United States considered introducing a census question on religion in the 1950s and 1960s, but concerns about the separation of church and state have prevented asking Americans about their religious affiliation in federal censuses.
In Canada many colonial censuses collected data on religion, and after the establishment of a national census in 1871, the question became a permanent part of the questionnaire. In 1951, when the census began to be collected every five years (as opposed to ten), the religion question continued to be asked every decade. For example, Canadians were asked about their religious affiliation in the 2001 and 2011, but not in 2006 or 2016. The census question asks respondents to list one religion or denomination, with the option of checking “no religion.”
How is the Census Collected?
The census questionnaire may be filled out online or on paper. In seniors’ homes, hospitals, remote areas and on Indian reserves, data collection may be done via in-person interviews. One adult member of the household responds on behalf of all other members. In 2016, the census was available in 11 Indigenous languages and 13 additional languages, as well as in formats to assist the visually and hearing impaired.
Prior to 1971, all data collection was done by enumerators going door-to-door conducting in-person interviews. Self-enumeration was introduced in 1971, and the option to complete the census online started in 2006.
Barry Edmonston and Eric Wai-Ching Fong, eds., The Changing Canadian Population (2011).