Chapter 1: Setting the stage and introducing the early 1970s

Table of Contents

The evolution of Canada's national statistical agency

Historical works are excellent opportunities to peer into the past, not only to satisfy our curiosity about "the way things were," but also to see how far we have come and to learn from the past. For Statistics Canada, such works are also opportunities to commemorate the agency's contributions to Canada and its people. They also serve as a reminder that an institution such as Statistics Canada did not instantly materialize—it was shaped over many years by many influences, and it continues to evolve each and every day.

Two relatively recent significant works have been authored on the history of Statistics Canada, the first of which was published in 1993: 75 Years and counting: A history of Statistics Canada. This publication covered the early beginnings and the first 75 years of the agency. Six years later, a more academically-oriented work was published: The Dominion Bureau of Statistics: A History of Canada's Central Statistical Office and Its Antecedents 1841-1972. The latter, written by retired Assistant Chief Statistician David Worton, intertwined the agency's story within the larger economic, political and social context of the times. True to common practice among historians, the story ends in the far enough past to allow for sufficient objectivity. Two decades later, in 2018, Statistics Canada celebrates its 100th birthday and, in honour of the occasion, has published this new work. Although not as academically oriented, it takes flight from where Mr. Worton left off and, after a brief synopsis of the early years, tells the story of the next 36 years, until 2008.

Employees processing vital statistics in the coding room, circa 1930
Employees working in the mailroom, 1952

The advent of statistics in Canada

Canada celebrated the 150th anniversary of its confederation in 2017. Prior to Confederation, statistical activity hinged around measuring the population to gauge the progress of European colonization. The collection of statistics on what is now known as Canada goes back to the early 1600s, when Roman Catholic missionaries collected data on births, deaths, and marriages of early European settlers. In fact, one can go back as far as the year 1605 by delving into records of European settlement at Port Royal, or by pouring through the writings of French explorer Samuel de Champlain and missionary Gabriel Sagard. In the data gleaned from these historical records, deaths exceeded births until about 1638. Thereafter, the population grew—in fact, a history of Statistics Canada written in 1952 quite matter-of-factly pronounces that "indeed, the French Canadian population has always been remarkable for its high rate of natural increase."

Of course, the original occupants of the land we now call Canada were the Indigenous peoples. A book published in 1876 on the Censuses of Canada from 1665 to 1871 contains statistical tables on social and economic conditions in Canada over those 206 years, including population estimates of Indigenous people. The earliest counts are sourced from letters from Jesuit missionaries written in 1611, and include estimates of population and land by band. The first systematic enumeration of Indigenous people took place in the Census of 1871, which counted 102,358 Indigenous persons.

The first known systematic enumeration of the colony population in Canada was first conducted in the winter of 1666–1667 by the first Intendant of New France, Jean Talon, who ironically, was later to be granted the title of count—Count d'Orsainville. He counted 3,215 European settlers in the three settled districts of Quebec, Trois-Rivières, and Ville-Marie (or Montréal). Jean Talon took this count on the request of the Sun King, Louis XIV, who desired to know how many people the colony contained after the more than 50 years of its existence. He recorded all persons in the colony by name, age, occupation, conjugal status, and relationship to the head of the family. He also measured the wealth of industry and agriculture, the value of local timber and mineral resources, and the number of domestic animals, seigneuries, government buildings, and churches. When the data he collected indicated a need for a larger population, he requested that more people be sent from France. This resulted in the immigration of about 1,000 women ("filles du roi"), who were to become the wives of the men already in the colony. To help the new families, Jean Talon presented them with plants and animals to encourage agriculture. He also established industries such as shipbuilding and built the first brewery. Some say he sowed the seeds of permanence, and helped establish the country's roots.

Over the ensuing 200 years, data continued to be collected, albeit irregularly, on the population, as well as on other subjects, including agriculture, trade, mining, and manufacturing. It was under the British North America Act of 1867 (which in 1982 was renamed the Constitution Act, 1867), when the Dominion of Canada was created by the union of the provinces of Ontario, Quebec, Nova Scotia and New Brunswick, that statistics were formally assigned to the jurisdiction of the federal government. That same Act called for decennial censuses of the population. It was also under this act that representation of the provinces in the House of Commons was first adjusted to reflect changes in the population. The province of Quebec was given a fixed number of seats (65), which was the number held by Lower Canada in the Canadian Legislature prior to Confederation. The other provinces were each assigned a number of seats based on their population as determined by the Census. The Dominion's Census Act, passed in 1870, laid the foundation for the 1871 Census—the first nation-wide census conducted after Confederation. One-hundred years later, in 1971, commemorative postage stamps would celebrate this significant census milestone. The groundwork for a general system of Canadian statistics had been laid by Joseph Charles Taché, appointed in 1864 by D'Arcy McGee as the Province of Canada's first permanent Deputy Minister of Agriculture with special responsibility for statistics. Taché therefore became the Commissioner of the first census of the new nation of Canada.

Statistics is an ancient calling

Jean Talon, Canada's first official statistician
Jean Talon, Canada's first official statistician

Jean Talon's census of 1666 places Canada among the first countries to take a census in "modern times." The word "census" has its roots in Roman times: it is derived from the Latin term "censere," which means "to estimate." Used by ancient civilizations for military or taxation purposes, censuses were also conducted in the eighteenth century in Sweden, Denmark, Austria and a number of Italian states.

The first census conducted in the United States dates back to 1790, and was notable for its primary purpose, which was to provide a basis for the apportionment of Members of the House of Representatives among the several States. Representation by population in the Federal House of Commons would not be introduced into Canadian legislation until Confederation, in 1867.

In the United Kingdom, the first census took place in 1801. Canada's 1921 Census Report makes reference to how, in the British House of Commons in 1753, a member stated that his constituents "looked on the proposal as ominous, and feared lest some public misfortune or an epidemical distemper should follow the numbering."

Fear of the census was not unknown to Canada, either. One of the reasons the Census of 1851 of Upper and Lower Canada was not a great success was that people regarded census-taking with suspicion. Joseph Charles Taché, the Secretary of the Board of Registration and Statistics at the time, indicated in his report that "…a very general feeling was found to prevail throughout the Colony, that the Census had some direct or indirect reference to taxation – and in this belief the Enumerators were received most ungraciously, and the information sought was, not only partially, but, in some cases, altogether withheld." Instructions to enumerators for subsequent censuses clearly stated that a census was not taken for the purposes of taxation.

The first Year Book

After Confederation, other official statistics developed. For example, the first compilation of Canadian statistics was performed in 1867 with "The Year Book and Almanac of British North America for 1867 – being an annual register of political, vital, and trade statistics, tariffs, excise and stamp duties; and all public events of interest in Upper and Lower Canada; New Brunswick; Nova Scotia; Newfoundland; Prince Edward Island; and the West India Islands."

From 1867 to 1879, the book was published by the Department of Finance, and for many years was the only vehicle to pull together the vast body of statistical information on the country and its inhabitants. The early books were produced commercially, and came complete with advertisements for steamship tickets, sewing machines, pianos, and life insurance. The book was officially made a Crown publication in 1879 under the Department of Agriculture, and advertisements were no longer accepted. The Year Book and Almanac was the progenitor of the Statistical Abstract and Record of Canada, which began in 1886, published by the Department of Agriculture, which had been given the power to regulate the collection and publication of general statistics under the Census and Statistics Act of 1879. The first French edition was published in 1887. The publication was later to be renamed the familiar Canada Year Book, in 1905, which became a staple in many Canadian homes, especially before the age of digital information. With a few respites, this compilation of statistics was produced for 145 years, including an electronic version starting in 2003. It ceased publication in all forms in 2012. A total of 111 editions were produced over those 145 years, the imparity a result of cost, organizational changes, or even the availability of Parliament in the early years. For many years, the Canada Year Book was the only vehicle pulling together the body of knowledge collected through the national statistical system, an official record chronicling over a century of Canada's economic, social, and legislative progress. However, it ultimately came to an end in recognition of the growing need of data users for up-to-date and online information, which the agency would make freely available starting in 2012.

Statistics were also beginning to be assembled as "by-products of administration," including insurance statistics from the Insurance Branch of the Department of Finance in 1875, data which were subsequently pushed back to 1869. Other early statistics of an administrative nature included criminal, mortuary, railway, postal, merchant shipping, trade, and immigration statistics.

The West India Islands

The West India Islands referred to the former British island colonies in the Caribbean, most of which gained their independence in the late 20th century. In 1854, the United States and Great Britain (on behalf of its British North American colonies) had signed a treaty that eliminated customs tariffs. Trade did consequently increase between the two countries, but American politicians were later pressured by the protectionist fervour sweeping the United States and demanded the treaty be abrogated. It expired in 1866 and was not renewed. The 1867 Year Book and Almanac includes a section on the trade of the West Indies, indicating that, "when it became evident that the United States were indeed about to abrogate the Reciprocity Treaty under which the main staples of Canadian produce had been freely admitted into their markets, it was decided that fresh outlets should be sought out for us abroad." A delegation from what was soon to become the Dominion of Canada went to investigate the Trade of the West Indies on a man-of-war steamer placed at their disposal by the Admiralty. The delegation produced a report, included in the 1867 Year Book and Almanac, on the nature and extent of West Indies production, trade, tariffs, and prices, along with suggestions on how to increase trade with British North America.

Creation of the Dominion Bureau of Statistics

The creation of a permanent Census and Statistics Office took place in 1905 under the Ministry of Agriculture and was another milestone in Canada's statistical history as it meant that the experience gained in the taking of a census was preserved for use in the next. Most other statistics were being produced in a decentralized fashion by many departments of the Canadian government, with next to no standards or comparability. The driving force behind the establishment of the permanent office was Archibald Blue, who was appointed special census commissioner in 1900 and who conducted the censuses of 1901 and 1911.

At this time, Sir Robert L. Borden was Prime Minister, and the Right Honourable Sir George Eulas Foster, the Minister of Trade and Commerce, was the Minister responsible for Statistics Canada. In 1912, Foster incorporated the Census and Statistics Office into his department, and shortly thereafter formed a Departmental Commission on Official Statistics to determine the best way of creating "a comprehensive system of general statistics adequate to the necessities of the country and in keeping with the demands of the time." The Commission would confirm the fragmentary and disparate nature of official statistics in the country and recommended a central statistical office.

Among the members on the Commission was Robert Hamilton Coats, a Toronto journalist, who started his Ottawa career in 1902 with the new Department of Labour under Deputy Minister William Lyon Mackenzie King. At the turn of the century, the Dominion Department of Labour had been created to collect information on labour conditions and wages, which it published in the Labour Gazette. R.H. Coats would become editor of the Labour Gazette with significant expertise in the area of wages and cost of living statistics. He had made his name largely through his work on a report in wholesale prices from 1890 to 1909, which was instrumental in demonstrating his masterly grasp of the technical issues involved in statistics. After his work on the Commission, he returned to the Department of Labour and was soon appointed to another board of enquiry, into the cost of living. He achieved notoriety by refusing to sign the report produced, on the grounds that he did not agree with its anecdotal nature, and produced his own report on the side. Quite collegially, both reports would be published.

In June 1915, R.H. Coats was appointed as the Dominion Statistician and Controller of the Census at the Department of Trade and Commerce to carry out the recommendations of the Commission and lay the groundwork for a centralized statistical system. He drafted An Act respecting the Dominion Bureau of Statistics, which Foster introduced as Bill 32 in the House of Commons on April 4, 1918. The bill was passed shortly thereafter, and received royal assent on May 24, 1918, creating the Dominion Bureau of Statistics and a centralized and coordinated national statistical system, with R.H. Coats as its very first Dominion Statistician. He would hold the position for a total of 27 years, actively encouraging innovation and development in the collection and compilation of data.

Structurally, the Statistics Act of 1918 was a consolidation of previous statistical legislation: the Census Act, the General Statistics Act, the Railway Statistics Act, the Criminal Statistics Act, and the statistical sections of other legislation. Specific sections dealt with the decennial and quinquennial censuses of population and agriculture, an annual census of industry (including Mines, Fisheries, Forestry, Manufactures and Construction), statistics of trade and commerce, transportation statistics, and criminal and general statistics. It also included sections on collaboration with other government departments and with the provinces. Most importantly, the Statistics Act established a central statistical system with broad authority to collect administrative and survey data for statistical purposes.

The unified and centralized system that resulted has been the underpinning of the Canadian national statistical system ever since, and it is therefore based on the creation of the Dominion Bureau of Statistics in 1918 that Statistics Canada celebrates its 100th birthday in 2018.

In the twitter of the moment

One cannot help but be humbled by the eloquent language of our predecessors. The following excerpt is from a 1936 address by R.H. Coats, speaking of his appointment as the seventh president of the Canadian Political Science Association, and as the first statistician in that role:

"Even in Canada, a statistical association was born, gasped and died before our own got on its feet. Thus, in all delicacy, speaking for my statistical confrères, we feel that after six political economists for president all in a row, it is fitting we were called up higher. In the twitter of the moment, we might even wonder if we are Cinderella, who from her ashes was once invited into the parlour. She came, you remember, in all meekness, with nothing of the kitchen fire in her eyes, yet I feel sure in some consciousness of merit, for it was Cinderella who finally married the prince."

Statistics Canada gets its name

Walter Elliott Duffett was Dominion Statistician when, fifty-three years later, a new Statistics Act was passed and proclaimed on May 1, 1971. Like R.H. Coats, Walter Duffett had come to Statistics Canada from the Department of Labour. He became Canada's first Chief Statistician when the Dominion Bureau of Statistics was reborn as Statistics Canada, and served in that role from 1957 to 1972.

Walter E. Duffett

Walter Duffett was an internationally well-respected economist. He was born in Toronto in 1910, although he attended public and high school in Galt, Ontario. He pursued a bachelor of commerce degree at the University of Toronto and a master of science degree in economics at the London School of Economics. After working as a security analyst in the Investment Department of Sun Life in Montréal and in London, Walter Duffett joined the Wartime Prices and Trade Board in 1942. He transferred to the Research Department of the Bank of Canada in 1944, and 10 years later was appointed Director of the Economics and Research Branch at the Department of Labour. He was appointed Dominion Statistician in 1957, and held that title until his retirement, in May 1972. He was technically Canada's first "Chief Statistician" when the Dominion Bureau of Statistics was renamed Statistics Canada, although he preferred to call himself "Government Statistician." After his retirement, he became Vice-President of the Conference Board of Canada. He was elected President of the Inter-American Statistical Institute in 1980.

Canada's population clock, 1969

The "racy" new Statistics Canada

Although the Dominion Bureau of Statistics officially became Statistics Canada earlier in the year, implementation of the new name was delayed until August, when the census would be out of the field, to avoid confusion in the public's mind as the 1971 Census documents bore the previous name. The first publication to bear the new name was the Statistics Canada Daily, launched on August 3, 1971. At the time, a number of alternatives to the new name were proposed. In fact, the public wrote in to the Minister responsible for Statistics Canada with suggestions. The alternatives proposed included the "Canadian Bureau of Statistics" and the "Statistics Office of Canada." Some people thought that the new name "Statistics Canada" was "a racy title" and a bit "flashy" and apparently there was internal concern over the inelegant potential contraction of "Statistics Canada" to "StatCan." Arrangements were made to change the name on the population clock in the lobby of the Main Building of the Dominion Bureau of Statistics. This clock had been unveiled by Prime Minister Pierre Trudeau in 1968, to commemorate the 50th anniversary of Statistics Canada. The clock, at that time, recorded a population of 21 million people in Canada.

The revised Statistics Act aimed to achieve more meaningful federal–provincial cooperation in statistical matters and included provisions for avoiding duplication in the collection of data through joint data collection and data sharing with the provinces. On the data user side, the Statistics Act included modifications to confidentiality provisions to allow researchers access to non-identifiable microdata. It also provided for increased access to administrative data, including income tax returns and other data held by federal government departments, to avoid duplication in the reporting of information. In response to growing concern over burden on business respondents, Duffett reported to a House of Commons Finance Committee on some of the positive outcomes of the new act, including that "the respondent is our best friend and we do not want to kill the goose that lays the golden eggs."

Statistics Canada's population clock

Statistics Canada's population clock was seven feet tall and four-sided and had been constructed as part of the Census program and to highlight the novelty of computer technology at the time. It had a digital counter across the top, and additional counters across the four faces, which counted births, deaths, and immigration. Statistics Canada's demographers had calculated the frequency of these occurrences and the clock was programmed accordingly. It was extremely popular at the Bureau, and with the public, who would pass through the lobby of the Main Building just to see it. The fate of the clock is somewhat uncertain, but it was likely dismantled after only a few years of operation.

A department in its own right

Walter Duffett was Dominion Statistician when the Bureau was recognized as a department in its own right, no longer accountable to the deputy head of another department. The 1960 Royal Commission on Government Organization (the Glassco Commission) was appointed to recommend changes in the interest of increased efficiency, economy, and an overall improved public service.

One of the most important symbolic outcomes of the Glassco Commission was that it formalized the independence by convention that had been achieved step by step since the time of R.H. Coats, to whom objectivity was of great importance. The formalization of independence to ensure ongoing objectivity and impartiality is an ongoing preoccupation of national statistical offices. In January 1965, by Order in Council, the Bureau was designated as a department of the federal government, with the Dominion Statistician as Deputy Head.

A key outcome from their report was the importance of the objectivity of statistical operations. This included the recommendation that the Dominion Statistician should not hold office "during pleasure" but instead "during good behaviour and removable only for cause." This recommendation resurfaced over half a century later in Bill C-36, An Act to amend the Statistics Act, which proposed the reinforcement of Statistics Canada's independence through various legislative amendments. Similarly, the Glassco Commission recommended the formation of an advisory council, comprising representatives of the principal users of statistics and other public bodies, to meet periodically with the Dominion Statistician on statistical matters, and to report annually to the responsible minister. At the time, the Bureau was reluctant to create an additional advisory committee, with an existing network of 45 to 50 subject matter and other committees, including those of a federal–provincial nature. It was not until two decades later that the National Statistics Council was formed, and it was not until 2017, with the introduction of Bill C-36, that the possibility of an advisory group undertaking an annual report on the state of the statistical system was brought forward. In fact, the recommendation for an advisory statistical commission was first brought forward unsuccessfully by R.H. Coats in 1918 and again a number of times in later years.

A major expansion in a time of great change

At the end of 1918/1919, there were a total of 123 staff at the Bureau, initially recruited from the Department of Trade and Commerce and the Department of Agriculture. In the early seventies, Statistics Canada's permanent and casual workforce numbered between 3,500 and 4,000. The agency hired 300 to 400 summer students each year, and had a difficult time finding and retaining enough staff to meet its growing needs, especially at professional and senior levels. For example, in the 1969/1970 fiscal year, there were about 1,700 new hires (of which 1,100 were full-time). This high number of new hires was likely partially attributable to the fact that the previous year had seen a government-wide recruitment freeze and the fact that the agency had recently re-organized. The re-organization that had just taken place, in January 1967, was an important step for a rapidly growing organization. The previous arrangement had a large number of divisions reporting directly to the Dominion Statistician. Such a structure was not sustainable as the size and complexity of the organization grew. Over a number of years, the Bureau had been considering grouping the divisions into branches, but the availability of professional staff was a big obstacle. By 1966/1967, it was felt that recruitment as well as the development of existing officers could meet the needs of the new re-organization, and the change was implemented. As well, as it does for each census, the agency staffed up for that of 1971, with about 50,000 temporary employees ready and in place for Census Day, which at that time was June 1.

Naturally, changes at the agency stemmed largely from what was happening in Canadian society at the time, and the 1970s was certainly a time of great change for Canada and for the world. Pierre Elliott Trudeau was Prime Minister; the country officially became a bilingual nation; the separatist movement was active in Quebec; the Vietnam War was under way; and world oil prices would soon quadruple in the 1973 oil crisis. Overall consumer prices would rise by almost 8% in 1973 compared to a year earlier, from price increases in gasoline, electricity, and food. These price increases were the first step to almost a decade of stagflation, where consumer price inflation was high and economic growth was slow.

The population of Canada was flourishing as a result of the post-war baby boom and increased immigration, and the labour force was expanding significantly, largely as a result of increased female participation. To keep up with the statistical needs of the country, the agency experienced the greatest growth in size, complexity and budget in its history of operation, with the budget growing by a factor of 7 and the number of staff more than doubling over the time Duffett was in office, from 1,449 in 1957 to 3,545 in 1972. Pressures for growth were not only the need for more precise and detailed information, but also the growing importance of educational, cultural and other social spheres of knowledge.

Employees punching cards for the 1931 Census
Computing colossus, the IBM 705 III, 1969

The separatist movement and the October Crisis

The Front de libération du Québec (FLQ) was a separatist and Marxist-Leninist paramilitary group formed in the early 1960s, known as the violent wing of the Quebec sovereigntist movement. It was responsible for a number of attacks throughout the 1960s, including bank hold-ups, kidnappings, and bombings. The most notable bombing carried out by the group was that of the Montréal Stock Exchange in 1969. In 1970, after about a decade of growing support for the FLQ, the October Crisis occurred in Quebec. Members of the FLQ kidnapped the provincial Minister of Labour, Pierre Laporte, and British diplomat James Cross. In response, Prime Minister Pierre Trudeau invoked the War Measures Act, in its only peacetime use, a controversial move that suspended civil liberties by granting the police wide-ranging powers of arrest and internment. Troops were also sent to patrol vulnerable areas and guard prominent individuals thought to be at risk. The FLQ called for a socialist insurrection against Anglo imperialism, the overthrow of the Quebec government, and the independence of Quebec from Canada. Public support grew among left-leaning Canadians, until mid-October, when Pierre Laporte was murdered. In early December, after he had spent about two months in captivity, Cross's release was negotiated. The list of demands included, among other things, the release of 23 political prisoners, currency in gold, the publication of the FLQ manifesto, a flight to Cuba or Algeria, and the cessation of all police search activities. The five known kidnappers were granted safe passage to Cuba by the Government of Canada after approval by Fidel Castro. Those responsible for Laporte's murder were arrested and charged. By January 1971, the troops were withdrawn from Quebec. The events accelerated the use of electoral means to greater autonomy, as opposed to violence, and garnered support for the sovereigntist political party, Parti Québécois, which would form the provincial government in 1976 and set the stage for a referendum on Quebec sovereignty four years later.

The age of automation

This was also the age of automation as the agency began harnessing computing power for its operations. While mechanical tabulation equipment was experimented with as early as 1891, computer tabulations were first carried out at the Bureau for the 1961 Census, using magnetic tape and an IBM 705 Mark III. For the 1971 Census, processing began with microfilming each questionnaire, and subsequently transferring the data to magnetic tape to feed into an IBM 360/65. At that time, the agency was also piloting facsimile transmission between the regional offices and Ottawa.

Chief Statistician Walter Duffett guided the Bureau through 15 years of fast-paced expansion and automation with the support of three assistant chief statisticians (ACSs). The agency had grown from a staff of 1,800 to about 5,000 by the time he retired. In fact, his habit of coming to visit each new employee to ask whether he or she were comfortably installed in Ottawa was never pursued by any subsequent Chief Statistician—the agency had just become too large. At the time of his retirement, H.L. Allen was the ACS assigned to "General Assignments" and assisted with management of the Bureau, Simon A. Goldberg was assigned to Integration and Development, and Lorne E. Rowebottom headed the Socio-Economic Statistics Branch. The remaining five branches (Administration, Economic Accounts, Financial Statistics, and Economic Statistics) were all administered by directors general. When Directors General positions were first instated at the agency, there was some hesitancy about the title, as it was thought to have "empire-building" connotations. For a time before the new position title stuck, some documents referred to the position, instead, as that of "Super Director."

There were eight regional offices at the time—Vancouver, Edmonton, Winnipeg, Toronto, Ottawa, Montréal, Halifax, and St. John's—responsible for enumeration and initial processing of the censuses of population, housing, and agriculture, the monthly Labour Force Survey, the monthly pricing of commodities and services for the Consumer Price Index, the Job Vacancy Survey and other special projects. They collected a variety of reports from local business establishments and provided assistance to users as well as local reference libraries as at this time there was no Internet to disseminate information.

The agency's resident inventors

Resident inventor, Fernand Bélisle, designing mechanical tabulators

The 1911 Census was the first to be processed almost entirely by mechanical tabulation—a temporary staff of 160 clerks worked with 70 card-punching machines and 20 tabulators. The first tabulator was invented by Herman Hollerith in the United States and first used for the 1890 U.S. Census. Hollerith's company would later become International Business Machines (IBM) Corporation. In Canada, the Dominion Bureau of Statistics experimented with electrical tabulation for the first time the following year, for the 1891 Census, for the processing of its population data. To help put this in historical context, electric power generation was relatively new—in fact, street lighting was first installed in the larger Canadian cities in the early 1880s. Electrical tabulation machines allowed population statistics to be compiled more quickly and at a lower cost, using perforated cards. Each enumerated person (of which there were about 4.8 million) had a card. The cards were first punched with a pantographic punch, each hole corresponding to one data point. "These cards are then passed through the electrical tabulator, which, by ingenious contrivances, records the answers on a number of dials." In fact, metal pins would pass through the holes on the punched cards to make contact with metal plates to complete an electrical circuit and register a count. It must have been a noisy room, as one of the census bulletins remarks that "Each time the circuit closing device is brought down upon the card, the hand of one of the dials moves and a bell rings to tell the operator that the religion has been registered." The data release bulletin remarks that "one of the advantages of this system is the accuracy with which the statistics are compiled. The instrument is not lazy. It is not made heedless by the state of the atmosphere in the room. It is not dishonest. It will not register a Presbyterian as a Roman Catholic, nor an Armenian as a Calvinist, from any of these causes. It is absolutely impartial." It took 40 days to do the first run of the cards through the machine to compile the religion data.

In the United States, the rental costs of the machines from Hollerith grew too high, and the Census Machine Shop set out to develop tabulation machinery of its own in the early 1900s. A newspaper account credits Charles W. Spicer with the invention of the tabulator used for the 1910 U.S. Census. After it was shown to be successful, the Canadian government subsequently bought the use of the patent and commissioned the manufacture of three of its machines by a company in Toronto. Later machines at the Bureau would be designed and built by resident inventors A.E. Thornton and F. Bélisle.

Fernand Bélisle's full name was Georges Etienne Fernand Bélisle, born in 1889 in the municipality of Wotton, Quebec. In 1923, R.H. Coats praised his work in a memo: "Mr. Bélisle is particularly good on machine tabulation. While in our employ, he invented an electric gang-punch machine which has been patented and which we are using at the Bureau. He has been in charge for some time of our horizontal sorting machines and has greatly increased their output; in fact, he has brought certain machines which handled only about 12,000 cards per day to a point where they are turning out nearly as many cards per hour." Mr. Bélisle is renowned as the developer of the Pantograph machine, used for punching cards since 1911, as the inventor of the electric gang-punch machine used in the 1931 Census, and as the inventor of the compressed-air sorter-counter machine used in the 1941 Census. In fact, one of the Bureau's six machines was loaned to the Government of Jamaica for the tabulation of its census of January 4, 1943.

Simon Abraham Goldberg

Dr. Simon A. Goldberg, father of national income accounting
Dr. Simon A. Goldberg, father of national income accounting

If Duffett was the stable force of the agency, Dr. Goldberg was the sparkplug for innovation. Simon Goldberg firmly believed that sampling was the path to the future, although this was a concept that divided the agency, which had historically been a census organization. He was an avid promoter of analysis and methodology. In fact, methodology at Statistics Canada owes a handsome debt to Simon Goldberg and his protégé, Ivan Fellegi, who were largely responsible for creating the agency's first methodology unit from the ground up.

Goldberg had immigrated to Canada as a child from Poland in 1927. He graduated from McGill in 1939, and then obtained master's degrees in economics from McGill and Harvard. After spending some time in the air force, he joined the Dominion Bureau of Statistics in the newly created Research and Development Section, which was set up to coordinate work in economic statistics and develop the income and expenditure accounts. Later, Goldberg would take a leave of absence to complete his doctorate at Harvard. He worked at the Dominion Bureau of Statistics from 1945 to 1972.

He was promoted in 1954 to Assistant Dominion Statistician, charged with the task of integration of statistics at the agency. He was a strong advocate of interdisciplinary teams and was among the first to recognize the potential impact of the computer at Statistics Canada. He had an innate ability to plan and separate the practical from the impossible, and was instrumental in the establishment of a planning function at the agency.

He was and still is regarded highly at the agency as the father of the Canadian National Accounts, through his vision for an integrated system, rather than a fragmented collection of indicators. He designed and co-authored Canada's first comprehensive publication of the National Accounts, and it was under his direction that Canada became the second country in the world to publish quarterly national accounts and the first country to develop data on income distribution by size. He is said to be one of the most important people to the history of the agency after Robert H. Coats, not only for his role in elaborating the national accounts, but also for building the methodology capacity of the agency along with other infrastructure elements, such as classification systems. Two of his important initiatives were that he created a research section in support of the national accounts and that he pushed for the agency to be a sample survey organization instead of a census organization—to provide for more timely information for decision-makers. To enable this switch to sample surveys, he had the idea of fostering administrative data to create a business register. He was also largely responsible for the legal access to tax returns granted to the agency for the first time in 1971.

He was a firm believer in the importance of research and development and of decision by consensus, and was immensely popular at the agency, with many senior managers seeking his counsel. David Worton, in his history of the agency, would indicate that, in the later Duffett years, the agency was run more or less as a partnership, with many senior staff taking their cues from Dr. Goldberg. Dr. Fellegi is quoted as saying "He had a very powerful individuality. He just swept away obstacles in his path by sheer power of intellect. He really had very little formal line authority in the Bureau, except to make proposals. None of the Assistant Dominion Statisticians formally reported to him. De facto, everybody did."

"In many ways, the modern face of Statistics Canada was drawn by Simon Goldberg, Deputy Dominion Statistician in the 1960s. He alone realized how critical it was to develop, maintain and use the framework of the national accounts to organize basic economic statistics; how to improve their quality by constant confrontation between measurements of the same variables available from different sources; and how to use the framework of the national accounts as the training ground for young statisticians and economists." – J. Ryten, Statistics Canada Historical Compendium

Simon Goldberg left the agency in November 1972 on a 3-year leave of absence to take up the position of Director of the United Nations Statistical Office. He would serve there until 1979, before taking on a number of consultation positions.

Mr. Simon Goldberg passed away in 1985 at the age of 85, and the main conference room at the agency was named in his honour. He is described in a SCAN special issue as "one of those rare individuals in whom there is both a consuming desire for the advancement of a science and a heart-warming interest in the colleagues engaged in that work." Just prior to his death, he was nominated for the Order of Canada, but never received it, as it is not awarded posthumously.

Conversion to bilingualism

When the first Official Languages Act was passed, in 1969, providing for equal status of English and French in the federal government, the Bureau followed suit. The Administration Branch included a bilingualism unit, which measured progress towards bilingualism objectives by reporting on language capacities, training of staff, and the availability of publications in both official languages. The agency was also taking part in a federal government initiative where units would be designated as "French Language Units" on an experimental basis. The creation of such units was part of the government's public service bilingualism program, designed to serve the public more effectively in either language, and to reflect the equal status of both languages. The agency chose four units—the two regional offices of Montréal and Ottawa, the Student Information Section of the Education Division, and the Building and Housing Permits Section of the Business Finance Division. Conversion of regular periodic reports to a bilingual format was under way; by 1971, 89% of the 620 publications were available in both French and English, up from 40% in 1970. This is not to say that bilingual documents had not been produced earlier—in fact, R.H. Coats had a translation staff for bilingual publications such as census reports, The Daily Bulletin and the Canada Year Book. The 1970s were an important time for the growth of bilingualism, and especially for the support of Francophones working for the public service in Ottawa. It was also in the early 1970s that Place du Portage was being built to house various federal departments, as the government increased its presence in Quebec.

Equal opportunities for women

In 1970, the Royal Commission on the Status of Women published a groundbreaking report with recommendations to the federal government to ensure equal opportunities for men and women in all aspects of Canadian society, including in the civil service. The recommendations included updating the legislative system and addressing such critical issues for women as poverty, family law, the Indian Act and the need for a federal representative for women. In 1970, the Royal Commission created the Office for the Advancement of Women and encouraged other departments to find ways to foster opportunities for the advancement of female employees. In 1971, the first Minister responsible for the Status of Women was appointed; Status of Women Canada became a departmental agency of the federal government five years later. Statistics Canada created its Equal Opportunities for Women Committee in 1973, which reported directly to the Chief Statistician. For a short time, there was also publication of an in-house paper called Action, which publicized the activities of the committee. Federal government implementation of the Royal Commission's recommendations was an important step toward greater equality for women in the public service.

A sign of the times

An example of the times was the Miss DBS Beauty Contest sponsored by the staff association. No married women could take part, and each contestant was judged based on "personal appearance which includes the clothes she wears, her poise and her personality." It was a popular contest, with 43 contestants, for example, in 1952. The winner would represent the Bureau in a competition against other departments of government, vying for the title of "Miss Civil Service." Such contests took place in the civil service up until 1973.

Innovation at work

The civil service had an incentive award program, which was first implemented at the agency in 1953. Cash or merchandise awards were offered for employee suggestions for improvements in any aspect of the civil service, under what was called the "Civil Service Suggestion Award Plan." In 1971, one such suggestion was for a "Census cardboard desk," which was subsequently processed for patenting under the Public Servants' Inventions Act. Each collapsible desk weighed about 8 pounds and cost $1.75 for the 1971 Census, compared to $39 for the usual wooden tables. The Statistics Canada administrative officer whose invention it was, Bill Butler, worked on his idea after seeing an American cardboard table which cost $8.50 but was too small for census purposes. He was awarded $5,000 for his invention, which saved taxpayers well over $500,000 in two censuses alone. The patent expired 20 years later, in 1992, although cardboard desks continue to be used today for census support units. When the temporary offices are closed down, the desks are folded up and recycled. Cardboard desks were also used by the U.S. Census Bureau in temporary field offices for their 1990 and 2000 censuses. While the desks are treated with fire retardant chemicals, apparently one U.S. office discovered a fatal flaw—they were no match against a malfunctioning fire sprinkler system!

Employees working in crowded office during the wartime period in the 1940s
Main computer centre, 1969

Employee news gets an overhaul

A new magazine for agency employees was introduced in 1972, called Scan. Its predecessors included the inaugural DBS News, which first appeared in the summer of 1931 (unfortunately no copies have survived the years); Green Island News, produced in 1946 while the Bureau was located on (surprise!) Green Island (now home to the John G. Diefenbaker Building, which served as Ottawa's city hall from 1958 to 2000; now mostly home to Global Affairs Canada employees); DBS Pasture-ized News first appeared in 1953, although the "DBS" was later dropped from the name; Stat Monthly followed from the end of 1954, followed by DBS Staff News from 1957 to 1959, and finally after 13 years of no staff newspaper, Scan was published from 1972 to 2000. Meanwhile today's staff newspaper @StatCan was first published in April 1997, and is still going strong 21 years later.

Notable milestones in the statistical program

The early seventies was a pivotal time for the application of statistical methodology to the Bureau's programs. Although mechanical tabulation had begun at the turn of the century at the Bureau, computer tabulation had just been realized in the early 1960s, and its marriage with methodology resulted in great synergies. In the preceding years, subject matter divisions were relatively siloed and more or less entities unto themselves. Each division designed its own questionnaires, mailed them out, received, coded and tabulated them. However, as automation and the use of administrative data grew, so too did the need for a more integrated approach to survey development, collection, and processing.

A new Methodology and Systems Branch was formed in 1970, which allowed for better integration of the major disciplines involved in the development of survey and census designs and their automation. The branch was created to increase the reliability of statistics, improve timeliness, reduce compilation costs, facilitate the extensive use of data by establishing machine-readable databases, and develop software to create integrated databases.

The paper "A Theory for Record Linkage," by Dr. Fellegi and Alan Sunter, had recently been published in 1969 in the Journal of the American Statistical Association, marking a major milestone in the development of record linkage, and paving the way for more efficient use of administrative and other databases in the construction of registers and the compilation and analysis of statistical data.

The year 1971 marked the completion of the first Canadian Occupational Classification and Dictionary of Classifications, which had been a massive 5-year research and development project carried out jointly with the Canada Department of Manpower and Immigration. Previously, occupational classification systems had been produced internally for census use and were subsequently used throughout the agency. However, this was the first standard occupational classification prepared for use in Canada.

The agency was expanding the availability of small-area data, as interest in subprovincial data grew as a result of greater involvement of the provinces in economic and social planning. Research began on the development of a system to provide census data at the level of small areas of geography, resulting in a geocoding system called the Geographically Referenced Data Storage and Retrieval (GRDSR) System and a storage, retrieval and tabulation system called Statpak.

It was also during these years that the feasibility of an address register was being discussed. A register of urban addresses would be established as a by-product of the 1971 Census, and was proposed as a sampling frame for the Labour Force Survey and other household surveys. Such a register had the potential to reduce field costs for future censuses and could be used for studies related to housing stock. Similarly, a central register of farms based on the 1971 Census of Agriculture was also in development, and was first used as the sampling frame for the 1972 December survey of crops and livestock.

These were also the formative years for the Central List of companies and establishments. The Central List started as a basic vehicle to standardize the classification of companies and their subunits. A major project in 1969/1970 expanded the list to cover all businesses in Canada and to enable classification by industry and point of employment. At the end of 1971, this list comprised 550,000 business names, each coded by industry and geography. A year later, the list would prove extremely difficult to maintain, and was phased out as the Business Register was phased in. The Business Register would become operational in 1972, and would initially cover all employers in Canada. In 1974/1975, the Register would be extended to include all corporations, including those who were not employers.

The Socio-economic Statistics Branch was responsible for the censuses of population, housing and agriculture, and vital and judicial statistics, as well as those on education, agriculture, health, family income, assets and liabilities. The branch was also responsible for the newly minted Field Division in 1970/1971, with responsibility for all of the Bureau's field survey activities. The division had previously been the Special Surveys Division, but was reorganized to expand its functions from data collection for a few large and critical surveys such as the Labour Force Survey to a larger survey base. The branch also had responsibility for the regional offices, provincial liaison, and consultative services. These would not stay long in the branch—as the Bureau would soon undergo the largest re-organization in its history, in 1972/1973.

The Labour Force Survey at the time employed 750 part-time interviewers who interviewed 30,000 households. The monthly Labour Force bulletin detailed employment and unemployment for Canada and by the five regions of Canada (which at the time were Atlantic, Quebec, Ontario, Prairies, and BC), and in 1971 started to publish data by province. Supplementary questions and sometimes special questionnaires were included with the Labour Force Survey to collect information on subjects such as smoking habits, immigration, paid absence from work, rental costs, and movement of persons within Canada. Other social surveys were conducted separately, such as the National Food Expenditure Survey and Family Expenditure Survey, but used the Labour Force Survey sampling frame. In 1970, the Labour Force Survey was moved to the Labour Division from Special Surveys, and plans were being laid for quarterly and annual publications to supplement monthly data. The survey was also the subject of a telephone pilot study, which was designed to determine the effect of telephone interviewing on the quality of the Labour Force Survey data, on the non-response rate, and on the cost of collection. The pilot project would prove that the method was feasible, less costly, and produced higher response rates. As a result, telephone interviewing was introduced for 40% of the sample of 35,000 households interviewed each month.

In late February of 1972, Duffett invited Dr. Sylvia Ostry, then-chair of the Economic Council of Canada, to participate in discussions about the Labour Force Survey, which at the time was undergoing a major review. There was increasing demand for additional information on the labour market in Canada, further to increasing levels of unemployment. A project team was established in 1971 to develop and implement major changes, involving virtually all aspects of the survey, and extensive user consultations were carried out. While the redesigned survey would not be implemented until 1975, there were significant changes made to the survey in 1972. Analysis began to be based on seasonally adjusted data in February; data by sub-provincial area were first released in April; the data first became available on CANSIM; and industry estimates were classified according to the new 1970 Standard Industrial Classification.

The first Job Vacancy Survey, introduced in 1971, provided information on occupational job vacancies by occupational group and province. These data were the first official measure of labour demand in Canada, with output to be used for decision making on training, vocational guidance, and general labour market analysis. Annual averages of minimum hiring rates associated with job vacancies were constructed, with the objective of developing hiring rates by occupation, industry, province, and census metropolitan area. There were, however, "Help-Wanted Indexes" developed in the federal Department of Finance, which extend back to 1962, and were based on help-wanted advertising space in selected daily newspapers across the country. This index was taken over by Statistics Canada from the Department of Finance in 1973/1974.

The onset of the National Accounts

Mr. Duffett, when he was at the Bank of Canada, contributed to the development of a conceptual framework for the first set of National Accounts for Canada, along with many other key players, including Simon Goldberg, Claude Isbister, and Agatha Chapman. In the 1930s and early 1940s, there was no single widely accepted way to define and measure national income. Key milestones to their eventual development include the work of John Maynard Keynes in the United Kingdom, who used macroeconomic theory in 1936 to justify government intervention in the business cycle. In the United States, economist Simon Kuznets would become the first to develop a conceptual framework for U.S. national income and product accounts. He was later awarded the Nobel Prize in economics, in 1971. In Canada, R.H. Coats set up a Committee on National Income Statistics in 1939, but the Bureau did not make much headway until the establishment of a central research and development staff in 1944 to support the post-war reconstruction efforts under the direction of George Luxton and his assistant director, Agatha Chapman, who was on loan from the Bank of Canada.

In the fall of 1944, experts from the United States (Milton Gilbert), the United Kingdom (Richard Stone), and Canada (George Luxton) held discussions in Washington on the development of a theoretical framework for national income statistics, referred to as the Tripartite Discussions, which resulted in agreement on the main essentials of the system. Luxton died at the age of 31 in 1945, and Claude Isbister was appointed to take over as director of the Research and Development Staff, with Chapman in charge of developing the accounts. In the summer of 1945, the dominion / provincial conference on reconstruction was held, at which Isbister and Chapman presented estimates of gross national product from 1938 to 1944 and income and outlay accounts by province. These were readily accepted and became the foundation for the calculation of equalization payments. Isbister and Chapman also attended the Princeton Conference in December 1945, which was the first meeting of the sub-committee on national income statistics of the League of Nations committee of statistical experts. This was the first attempt at international standardization.

The first annual estimates of balanced income and expenditure accounts in Canada were published in April 1946. These data would support a new federal program of fiscal equalization for the provinces as well as the goals established in a 1945 government White Paper on Employment and Income by Professor W.A. Mackintosh, formerly of Queen's University in Kingston, Ontario. He had been called to Ottawa during the war to help manage the war effort, working for Minister C.D. Howe, responsible for the Department of Reconstruction and Supply. He felt it was important to reassure the country in a time where uncertainty and confusion were beginning to develop about the transition to a peace-time economy that the difficult economic conditions that arose prior to the war would not redevelop. His paper referred to stable levels of employment and income and higher standards of living as an aim of government policy.

Simon Goldberg, Canada's father of the National Accounts, had the vision to adapt the White Paper into the core of the agency's economic programs, designing the country's first quarterly national accounts. The approximately 50 staff of the research and development unit were tasked to take the economy apart, classify the parts, attach measurement to them, and fit them together into a conceptual framework of the national accounts system. It would take many years to develop, with the first quarterly national accounts appearing in 1954 and the first major reference manual appearing in 1958.

The beginning of the Cold War in Canada

Economist Agatha Chapman

Agatha Chapman was born in England in 1907, and immigrated to Canada with her family. Her great-grandfather was Sir Charles Tupper, a Father of Confederation and one-time prime minister. She earned a master's degree in economics from the University of Toronto, which awarded her the prestigious Maurice Cody Fellowship. She also served as the president of the University College Women's Literary Society. She began working as an economist at Sun Life Assurance in Montréal and attended meetings of the McGill chapter of the Student Christian Movement, an active movement focusing its attention on topics such as poverty, disarmament and fascism.

Meanwhile, in September 1945, Igor Gouzenko, a young cipher clerk posted to the Embassy of the USSR in Ottawa, defected with secret documents revealing the existence of a Soviet spy ring in Canada, particularly within the federal public service. The Canadian government placed Gouzenko and his family under protection and this news was kept secret while Prime Minister King and his American and British counterparts discussed the matter. Prime Minister King set up a Royal Commission (the Kellock-Taschereau Commission, headed by judges Robert Taschereau and Roy Lindsay Kellock) to investigate. The matter became public in February 1946 when an American columnist began leaking rumours of the issue. The Commission recommended the arrest of 13 people, two of whom were friends of Agatha Chapman's. They were rounded up by police and taken to the RCMP barracks in Rockcliffe to be interrogated. The Commission ultimately reported that the spy ring had been operating since 1924 inside several government departments in Canada, in the British High Commission in Ottawa, and in the joint Canadian-British atomic research project. The Gouzenko affair marked the beginning of the Cold War for Canadians and exposed the country's vulnerability in security matters, ultimately resulting in the restructuring of the national security system.

Agatha Chapman testified before the Commission in March 1946. That June, she was identified by the Commission as a "spy" and a communist "cell leader" who had aided the transmission of secret information to the Soviet Union. At her hearing, she indicated being a member of study groups discussing socialist and Marxist literature—such meetings were frequented by people of interest to the committee. While she worked at the Dominion Bureau of Statistics, her de facto employer was still the Bank of Canada and, although she had not yet been charged, the report had become public. As a result, she was put on leave with pay pending resolution of the issue. In an effort to deal with the associated stigma (of being named but not charged), in August she asked to be put on trial to clear her name and restore her reputation. The trial was interrupted after about 4.5 hours, with the accusations withdrawn, as the Crown failed to produce any persuasive evidence that she was a Soviet agent. The defence never even had the chance to present its case.

Her life, despite the clearing of her name, would never return to normal. The Bank informed her that it would not reinstate her in her research post, and she was ostracized from the Canadian civil service. Claude Isbister wrote to Richard Stone in Cambridge indicating her clean slate, her considerable talents, stating that it was "unfortunate that she was not to be allowed to return to the government service." She was quickly hired at Cambridge as a research economist by Mr. Stone, and left Canada in March 1947. After three productive years at Cambridge, she returned to Canada in the 1950s, and formed a research consultancy in Montréal. She tragically committed suicide in 1963. It was suggested that she had never recovered from the trauma of the Gouzenko affair.

Employees at the agency may recognize her name as the title of one of the agency's awards—the Agatha Chapman Innovation Award, which recognizes those who have demonstrated a commitment to innovation and dedication to the development of noteworthy ideas.

The Census program

The first census taken by the newly established Dominion Bureau of Statistics post-1918 was the 1921 Census, which had five separate schedules or questionnaires covering population; agriculture, animals and animal products; fruits not on farms; and manufacturing and trading establishments; as well as a supplemental questionnaire for persons who were blind and deaf.

The 1971 Census saw several major improvements and innovations. It was the first census during which self-enumeration was adopted in all but three percent of the population located in remote areas (examples given included the "Newfoundland outports" and the Northwest Territories) and institutions. This was a major change in census collection methods, as every census up to that point had been conducted by interview. Self-enumeration minimized enumerator error and improved data quality, while addressing privacy concerns and respondent burden. Hand in hand with self-enumeration, telephone assistance was first introduced to help Canadians complete their questionnaires.

Sampling also became an integral part of the 1971 Census, with most questions asked of 1 in 3 households. The "short" form, comprising basic population and housing questions, was distributed to two-thirds of Canadian households, while the "long" form contained additional housing and socio-economic questions. The first use of sampling in a Canadian Census was for the 1941 Census of Housing, which collected a descriptive record of every 10th occupied dwelling in the Dominion. The 1941 Census followed a decade of depression and almost two years of war, and the associated social and economic effects were unprecedented. The rapid expansion in agricultural and factory production during the war had brought about widespread movements in population, which created acute housing shortages. There was also concern about the state of housing. After the First World War, a Royal Commission had been appointed to study Canadian social and industrial problems and had reported on "the scarcity of houses and the poor quality of some of those which did exist." It was recognized that the report had not been based on comprehensive statistics. Therefore, the 1941 Census of Housing was deemed important to the recovery after the war.

The 1971 Census saw the introduction of geo-coding to allow custom tailoring of areas of interest. Another important innovation was the reverse record check, which sought to determine how many people were missed by the census. It was also the first census to carry out an agriculture–population linkage, combining data from the Census of Agriculture with data from the Census of Population and Housing. As well as manually coding write-in answers for the population and housing questionnaires, regional processing offices hand-matched Census of Population and Housing and Census of Agriculture questionnaires to prepare for the later computerized linkage in Ottawa. The result was a rich socio-economic information base facilitating new analytical possibilities.

The 1971 Census witnessed protests by women's liberation proponents over the designation of the husband as the head of the household. In 1971, the head of the household was defined as "the husband if both husband and wife are present [or] the parent (regardless of age or dependency) if living with unmarried children." Five years later, the definition changed to refer to either husband or wife; by 1981, the reference to a head of the household had been dropped altogether, in favour of a somewhat less hierarchical "Person 1."

Flashback to early censuses

Early censuses were labour intensive enumerations. In explaining the time required, the first bulletin of the 1891 Census indicates a staff of enumerators of 4,300 who "had to traverse the immense area of Canada by every imaginable method of locomotion. A steamer with enumerators on board went in and out the deep indents of the Pacific Coast line as far as Alaska, thence to Queen Charlotte's Islands, to enumerate the people. Pack-horses were required in the mountain regions of the same province to carry the enumerators and their portfolios through the valleys which run among the hills of the Rockies. Dog-trains were a necessity in Saskatchewan. To obtain the population on the northern slope of the "Height of land" in Ontario and Quebec, a canoe expedition started from the head waters of the Lievre River to go by lake and river and portages to Albany River, at James' Bay. Camping outfit and canoes were needed to enable the enumerators to take the population in the Nipissing district just beyond the "Height of Land." The enumerators in Manitoba had, now to foot it, now to go by buckboard, and now by boat, and in one instance the man, losing himself, had to save his own life by slaying and eating his horse. Many townships in Algoma had to be taken by slow and toilsome pedestrianism. For the north shore of the Gulf of St. Lawrence a schooner had to be chartered, the enumerators put on board and dropped at different points till the Straits of Belle Isle were reached, from which point the schooner was directed to the Isle of Anticosti, the Census of which having been taken she returned to the Straits and sailed along the coast picking up the enumerators and returning to the mouth of the Saguenay River. Delay is therefore inevitable. No time limit can be given the enumerator."

The germination of sampling

In the wake of the 1930s Depression and the upheaval of the Second World War, the use of sampling helped the Bureau to meet an unprecedented demand for statistics. Its use to produce data more quickly and more cheaply, with less questionnaire burden was invaluable to the Bureau and to the growing information needs of the country. The first major survey to use sampling was the Labour Force Survey in 1945, largely as a result of the efforts of Nathan Keyfitz, one of the most eminent authorities in the field of demography and one of Statistics Canada's most distinguished alumni, who worked at the Bureau from 1936 to 1959. In 1948, the Statistics Act was amended to authorize the collection of statistics by means of sampling.

On a clear day you can count forever

The advent of self-enumeration for the 1971 Census gave rise to the slogan "Count Yourself In" in an effort to increase public awareness of the importance of the census and support self-enumeration. A film ("On a Clear Day You can Count Forever") was prepared, likely a play-on-words to the popular Barbra Streisand song which had just been released—"On a Clear Day You Can See Forever." A commemorative stamp was also issued on Census Day, and special census cancellation dies were used in two hundred post offices across the country. Another public publicity project involved the participation of one million students in the 1971 School Census, a voluntary project in which school children were familiarized with the census, which, by extension, served to familiarize their parents with self-enumeration.

The international scene

Statistics Canada's international contributions began in the 1920s with R.H. Coats' participation in the First Conference of Empire Statisticians, at which he expressed strong advocacy for statistical centralization. The agency continued to contribute in the 1930s to the statistical work of the League of Nations (which would later be replaced by the United Nations) and the International Labour Organization. In 1935, the British Government enquired whether the Government of Canada would allow secondment of an officer from the Dominion Bureau of Statistics to go to Palestine for three years as head of a proposed Bureau of Statistics. Sedley Anthony Cudmore would take up the post (and would later become Dominion Statistician after the retirement of R.H. Coats). By the 1940s, after a lull during the war, the provision of technical assistance internationally became a common practice.

In the 1970s, Statistics Canada continued to take an interest in the statistical needs of developing countries, including support for the processing of the 1970 Census in the West Indies. In cooperation with the Canadian International Development Agency, the agency's former IBM S/360-30 was relocated to Kingston, Jamaica, at the Computer Centre in the University of the West Indies, for its census processing.

The Bureau also helped with preparations for the 7th Conference of Commonwealth Statisticians, held in India in the autumn of 1970, especially as the previous conference had been held in Ottawa in 1966. The 1966 conference in Ottawa was a two-week affair, held from September 19 to 30, and was attended by 22 delegates from 15 countries, including Walter Duffett, Simon Goldberg, and Lorne Rowebottom from Canada, as well as observers from the Commonwealth Secretariat-General, the Republic of Ireland, and the United Nations Statistical Office, and specialists from the Bureau and from the West Indies. The conference was officially opened by Mr. Mitchell Sharp, then-Minister of Finance, who, when he was Deputy Minister of Trade and Commerce, had recommended Walter Duffett to his Minister the Right Honourable C.D. Howe for the job of Dominion Statistician. The first Commonwealth Conferences had been held in London in 1920, Ottawa in 1935, Canberra in 1951, London in 1956, and Wellington in 1960. This 6th conference was targeted at promoting contacts and collaboration between senior officials in each Commonwealth country as, over the years, the growth of independence from the United Kingdom had increased the number of countries attending.

A more active role in sharing information

The agency's most recognized data bank goes by the moniker CANSIM, which stands for the Canadian Socio-Economic Information Management system. The Information Division was responsible for the implementation of the first phase of CANSIM. The system was launched in 1969, first programmed in FORTRAN and COBALT, with data fed into the system using punch cards. It was originally developed in the United States at Southern Methodist University in Texas, and was brought to Canada in 1965 by the Economic Council of Canada, and subsequently turned over to the Dominion Bureau of Statistics in 1966. The first data included in the databank were those from the National Accounts. In 1968, the system contained 2,500 time series, requiring "gigantic memory requirements" of 100 kilobytes. In 2017, the more than 72 million time series took up more than 300 gigabytes of storage space. Electronic access to CANSIM was introduced in 1972, first to employees of the federal government through remote terminals and, then, a year later, to the general public through secondary distributors. It would not be repatriated to Statistics Canada until 1984.

Employee working with a Vari-Typer machine, circa 1950
Employees working in print shop, circa 1960

A growing focus on timeliness

In the early seventies, the agency also moved toward a more active, rather than responsive, publishing role, and there was a significant drive to improve the timeliness of publication, with initial focus on statistics of employment and payrolls, imports and exports, retail trade, and industrial production. This stemmed largely from a 1968 task force on government information, which conducted a user survey, whose respondents criticized the timeliness, usefulness, clarity, and accessibility of the Bureau's data. The Bureau-wide drive to effect gains in timeliness starting in the late 1960s would precipitate analysis of the causes of delays and means of eliminating them. Of course, longer-run plans would include conversion to automatic data processing. Recall that computing was in its infancy yet, and there were no desktop computers or email.Computation of results involved using machines to punch cards, lugging around heavy boxes of card, and using card-readers to submit data to the computer, which would then spit out results on bulky paper printouts. Communication was by telephone or type-written memos, and dissemination of information to the public was only by means of paper publications, which were warehoused and ubiquitous. In fact, the national accounts were compiled in pencil, in enormous, well-thumbed ledger books. The issue of timeliness was of particular importance to the accounts, where the quarterly accounts were being released with a 90-day lag, which users found to be much too slow, and a good deal slower than the lag in the United States for the same data. Steps were taken to reduce the time required for the compilation of the accounts to 60 days, where it remains today.

The introduction of a policy on official release

Until the early 1970s, the agency was still struggling with the effects of having grown in size very quickly. While Duffett had made organizational changes to strengthen top management, there were still varied approaches in different areas of the agency to things like the pre-publication release of material. Pre-release material was provided by hand, by mail, and by coded telegram to various departments, regional offices, provincial governments, and even companies and industry associations. There were many long-standing arrangements, which sought for example to improve timeliness or improve relations with business respondents by promoting cooperation in exchange for information. Pre-release policies were not clearly described nor disseminated throughout the agency, and, in July 1971, the Executive Committee expressed dissatisfaction about the pre-publication release of material and recommended the development of a policy on the release of data. Discussions ensued throughout the fall. However, in October and November of 1971, Labour Force Survey data were leaked to the Montréal Gazette. At this time, Statistics Canada and the Department of Manpower and Immigration were collaborating on joint press releases for the Labour Force Survey, with the Department of Manpower contributing to the analysis of the data. Both departments carried out investigations as to the possible source of the data leak. Duffett asked the Department of Finance to carry out an investigation as well, noting in his request letter that "the reporter in question indicated to my staff that his source of information was a good one and that he proposed to continue to use it." In one instance, a Department of Finance officer spoke to the Press at 7:50, thinking the release was at 7:30, although the official release was at 8 am. To curb the data leaks, a policy on official release was developed, which included systematic use of The Daily for all releases. The Chief Statistician would also need to approve all pre-releases of information. The joint press release of the Labour Force Survey was dropped, and Statistics Canada fully took on the analysis of the data. This was not because the Department of Manpower and Immigration was responsible, but merely to cut down on the number of people privy to the pre-release data. The Labour Force Survey data was also released two days earlier, and pre-release was discontinued to all parties except the departments of Finance, Labour, and Manpower and Immigration. The source of the data leaks was never discovered.

The changing of the guard

Employees in front of the Main Building at Tunney's Pasture in Ottawa, 1966

In the late 1960s and early 1970s, the agency was trying to establish a planning system to set priorities and guide long-term planning. There was a central planning group established under Dr. Simon Goldberg. The planning did not seem to be advancing very quickly, and the Executive Committee kept sending the planning group back to the drawing table. The agency knew it needed to plan better, as did the Privy Council Office and the government at the time, but it just did not seem to be able to get there quickly enough.

In February of 1972, Simon Goldberg's paper "Towards the development of a comprehensive medium term plan for Statistics Canada" had been presented to the second meeting of the Interdepartmental Advisory Committee on Statistics. The paper was intended to provide a framework and background for the Committee's deliberations, and to serve as general guidelines and perspective for the planned 1973/1974 program reviews. The Committee had been set up in 1971 to seek the objective views of other federal government departments with respect to the content of the statistical program and the agency's service capacity. The Committee was short-lived, as it did not quite have the objectivity that was intended—it quickly morphed into a lobby for promoting departmental interests.

There was an added complication: the Bureau had grown so quickly and had lost much experienced personnel in preceding years as a result of the previous differences in pay afforded by the Civil Service Commission between departments involved in the policy-making process and those not. Recommendations to the centre for promotion of personnel by the Bureau had previously involved prolonged, time-consuming and frustrating negotiations. The differing pay structure meant that the Bureau could not recruit or retain experienced staff, who would leave for other departments. While the Glassco Commission report would recommend an end to the pay discrepancy for Bureau employees, the effects of this loss of experienced staff was likely still being felt in the early 1970s.

In January 1970, a Cabinet Committee was discussing the topics to be included in the 1971 Census and requested a memo on how priorities were determined in meeting the government's statistical requirements, and how the needs of departments were ranked to ensure the Bureau's priorities corresponded with those of the government as a whole. They also requested information on the extent to which duplication existed in the collection of information and how duplication could be avoided. A report was prepared by Duffett and submitted to the committee by Minister Pépin, which noted some of the difficulties the agency encountered in determining the long-run needs of the government and of government departments, and made a number of suggestions. One suggestion was the establishment of an advisory committee of senior government officials chaired by the Privy Council Office (PCO). It recommended that the agency seek to improve and extend interdepartmental consultation procedures, that the PCO make available to the Chief Statistician documents that reflected the implementation of government priorities, and that both parties consult at regular intervals. It would not be until three years later that, under Dr. Ostry, the PCO would first share relevant Cabinet documents to help improve the agency's perception of emerging informational requirements for policy formulation and evaluation and help it to be responsive to federal needs. Duffett had tremendous difficulty gaining cooperation from the PCO, and his letters to the PCO and to Minister Pépin indicated growing frustration that he could not fully answer the request of the committee without their input and cooperation. The PCO explained its delay in responding by indicating that there were fundamental questions as to the role and function of the agency that they were working through first. Duffett felt that there was a lack of confidence in his judgement and signalled his desire to retire. He officially retired on June 30, 1972, and went on to become vice-president of the Conference Board of Canada. A few years later, he took up the post of vice-president and later, president, of the Inter-American Statistical Institute. He was also the founding editor of the Canadian Business Review, which was launched in 1974. He passed away in 1982.

Upon Duffett's retirement, Minister Pépin, the Minister responsible for Statistics Canada, gave a speech at his retirement party in May at the National Arts Centre. He spoke of his direct and thorough approach, noting that he was a precise, quietly purposeful and persuasive man. "We in Cabinet recognize his achievements as many and substantial. We have seen the Bureau grow in public esteem under his leadership and in international stature. We have seen Statistics Canada grow into a far stronger and much more productive organization than it was when Walter took it over 15 years ago."

Date modified: