toward farm mechanization resulting in "fewer and larger" farms, often through amalgamation of existing farms, and causing notable transformations in methods of farming operations and in farm employment.

Such rapid developments in population and agriculture indicated the need for bench marks at the 5-year period in order to provide a more accurate basis for annual estimates. Further, one of the chief values of the 1956 census, both for population and agriculture, is that it provided information for small areas (counties, municipalities, cities, towns, etc.), which cannot be obtained from intercensal estimates. Thus, the 1956 census has proved of value to Government departments at all levels for such purposes as determining "per capita" financial grants and subsidies, to research agencies and individuals in the study of many types of social and economic problems, and to business and industry for such uses as marketing research and the determination of sales potentials.

It was realized that if a complete census of Canada were to be undertaken in 1956, it must, of necessity, be of simplified design and restricted to basic essentials. Plans were made within a budget framework not to exceed $5 million, as compared with the 1951 census expenditure of some $8.3 million. Questionnaires were limited to basic inquiries on population and agriculture.

The 1966 census marked the second nationwide census taken on a 5-year basis. Much of the same factors which led to the first quinquennial census of Canada at the midpoint of the 1951-61 decade on June 1, 1956, influenced the decision that a modified mid-decade census was again a necessity 10 years later in 1966. These factors were related to the urgent needs in a period of rapid change in socio-economic conditions for small-area bench-mark data which are impossible to obtain from intercensal sample surveys. Moreover, pressures were extremely great from both Federal and provincial agencies of Government, as well as from numerous other users, to expand the restricted subject coverage of the 1956 population census to include in 1966 such topics as labor force, income, housing, etc. These pressures had to be resisted because of the serious lack of time and resources to plan and conduct another full-scale census within a 5-year period following the large decennial 1961 census.

Accordingly, a basic bench-mark census was conducted as of June 1, 1966, similar in scope to that of the previous quinquennial census in 1956. Despite the fact that the population questions were few (i.e., name, sex, age, marital status, relationship to head of household, structural type and tenure of dwelling), the 1966 census provided valuable up-to-date counts and distributions of population for all types and sizes of geographical areas, trends in the marital status, age and sex structure of the Canadian people, the size and composition of households and families, the homeowners versus tenant status of householders and the types of dwellings in which Canadians live.

The 1966 Census of Agriculture was more comparable in scope to the ful-scale 1961 census and included inquiries in farm areas and condition of land, acreages sown to the various types of crops, numbers and kinds of livestock and farm machinery, milk production, value of sales

of agricultural products, as well as certain items of farm expenditure and farm labor.

The passage of the new Statistics Act in 1971 now makes it a legal requirement for a national census to be held in Canada every 5 years. The act does not specify whether censuses in years ending in "6" should be "mini," "midi," or "maxi" in scope. Present indications are that the 1976 census may well be a midi census, as opposed to the mini censuses of 1956 and 1966, and the maxi censuses of 1961 and 1971.

Task forces have already begun to think about various aspects of the planning for the 1976 census. However, in view of the workloads of the 1971 census in the present period, these plans cannot be advanced beyond very preliminary considerations. One thing seems certain and that is the provincial statistical agencies are going to press harder than ever for occupational details of the labor force to be included. At this date, such subjects as education and migration also appear to be strong contenders for inclusion as well as the basic subjects of age, sex, marital status, and language. The list will likely grow longer as various departments of Government and other users are consulted over the coming next couple of years.

In the course of all Canadian censuses, decennial and mid-decade, all persons are required by law to reply to all questions. The regulation carries penalties of up to a $500 fine and/or 3 months in prison for a refusal to answer questions.

There was some question earlier about the costs of a mid-decade census. I can indicate to you the costs of recent censuses in Canada, but as a sort of rule of thumb, the mid-decade census tends to cost about half of what the preceding decennial census cost. We are expecting that the cost of the census in 1976 may be perhaps $20 million for a small-size census, up to as much as $40 million for a medium-sized

census.

The 1971 census will cost, we expect, about $50 million, so that the mid-decade census will probably be about half of that, depending very much of course on the number of questions that are asked.

APPROXIMATE COSTS OF DECENNIAL AND QUINQUENNIAL CENSUSES

The two restricted and basic quinquennial censuses of 1956 and 1966 have cost about half as much as each of their preceding decennial censuses. There is no way of knowing what 1976 will cost, but it could be anywhere from $20 million for a basic census to as much as $40 million for a medium-size census.

Mr. DUFFETT. Thank you, Mr. Chairman.

Mr. WILSON. Thank you, Mr. Duffett. Why is the cost so much greater per person in Canada than in the United States? Ours came to about $1 a head, and yours is about $2.50 a head.

Mr. DUFFETT. In the first place, our statistics cover a census of population and agriculture, and the agriculture census in a country like Canada is a substantial matter.

We also cover housing in our census. Perhaps you do in yours, too. But basically there is a certain, what you might call critical mass involved in census taking. You have to make very certain basic preparations, certain basic studies, certain basic arrangements for publicity. The same phenomenon occurs in connection with our monthly survey of the population, which is not very different from the American one. Proportionately it costs a good deal more than in the United States.

Mr. WILSON. I note from your statement that the quinquennial census program is established principally to measure the large increases in shifts of population. Were there any other conditions or circumstances which led to the establishment of the quinquennial census?

Mr. DUFFETT. The considerations that you mention, of large shifts in population and varying growths in different areas of the country concern the business community very much of course, and in market studies, business planning and what not, it is extremely important to have up-to-date information.

The character of the population, the sort of characteristics of the population that both government and business are interested in, do change a good deal and change very rapidly.

For example, the educational characteristics of the population change a great deal, and the need for training for particular occupations changes even more rapidly.

For instance, the Canadian Government is at the Federal level spending some $300 million a year on training, as opposed to education.

What is industrial training? It is of vital importance that this $300 million should be properly spent in order to train people for jobs that there are, and not for jobs that there are not. So that the changing composition of the job structure in the country requires up-to-date information.

This is an example of the sort of things I have in mind.

Mr. WILSON. You describe the penalties, and the fact that you have a mandatory census. Do you have any vocal groups or pressure groups of any significance that complain about the invasion of their privacy through the census programs in Canada?

Mr. DUFFETT. We have not up to now had any organized resistance to the census. The most organized resistance that occurred in 1966 was on the part of a number of what are called hot line operators, these are people who conduct conversations with citizens on the radio. One in particular found it to his advantage to try and stimulate opposition

to the census on the grounds that it was an invasion of privacy. This man was brought into court immediately, and that was the end of the difficulty. It did not spread.

There is no organized resistance to census taking in Canada as yet. Mr. WILSON. We anticipated that we would have some resistance this past year, but everything turned out pretty well. In spite of all the fears that we had, and the threats that had been made that there were going to be mass protests on the courthouse steps and so forth, our census enumerations worked out very smoothly, and there was a minimum of resistance to the census taking in our country.

Mr. DUFFETT. My colleagues in the Census Division and in my information division tell me that on balance, expressions of views of this kind perhaps do us more good than harm, in that they draw to the attention of the public that the census is being taken, that it is an important thing, and it tends to reinforce the desire on the part of all portions of the population to be counted.

Very often groups which are in other respects resent ful of the activities of government, nevertheless wish to be sure that they are included so that their importance can be recorded.

Mr. WILSON. Now in Canada you have a large country with some very sparsely populated areas, and you have many ethnic groups. Have you had any difficulties in getting what you regard as an accurate count of the population?

Mr. DUFFETT. The characteristics you mention do indeed cause certain special problems. For example, in 1971 census, we cannot use a mail-out system. We are using a mail-back system, but it is impossible to use a mail-out system because one cannot tell in advance in which language, English or French, people expect to be enumerated.

So it is necessary to have enumerators visit the households, note the number of people in the households, and supply questionnaires, and at the same time determine the language in which they wish to be enumerated.

Other aspects of our language problem in Canada are very important for census purposes. It is necessary to obtain information on the language which people spoke as their mother tongue and also the language which they currently speak, which is not necessarily the same thing.

The Canadian Government is making a determined effort to see that significant groups of Freenh- or English-speaking minorities are serviced by the Government in their own language and it is therefore necessary to find out the language which people normally use, and this will be one of the questions in the 1971 census.

Mr. WILSON. Are you required to print the forms in both languages? Mr. DUFFETT. Yes.

Mr. WILSON. It is just the two-French and English?

Mr. DUFFETT. That is correct. Identical forms printed in both languages, and all literature relating to the census, all advertising, appears in both English and French.

Mr. WILSON. And this is a standard form that goes to all parts of Canada? In other words, the same form would go to Alberta?

Mr. DUFFETT. That is true. Because in nearly all provinces of Canada there are small minority groups of French- or English-speaking people who expect to be enumerated in their own language. Needless to say, there are problems sometimes in finding the census enumerators who are capable of operating in both languages. There are similar problems, to a lesser degree, in other minority groups. For example, Toronto has a large area of Italian-speaking people, and special arrangements are being made there to provide interpreters to assist them in making out the forms. It is our feeling that the self-enumeration, which we are now using for the first time, will be a help in this respect because the people can gather around the kitchen table in the evening and fill out the forms with the assistance of some member of the family who is fluent in either English or French.

Mr. WILSON. Has the Government of Canada considered the conduct of a sample survey of population and housing at any time? Mr. DUFFETT.As an alternative to the census?

Mr. WILSON. Yes.

Mr. DUFFETT. No, these are not regarded as alternative approaches. We do use sample surveys to a substantial degree. We have a monthly sample survey of the labor force, extended to other characteristics of the population from time to time, education, household equipment, and so on.

But there comes a time when it is essential to have a full count both as a basis for good sample surveys which cannot otherwise be reliable, and in order to meet the needs for information on small areas.

In Canada, as I suspect in other countries, there is intense preoccupation with certain problem areas: areas of low income, areas of economic change, areas of social problems.

In order to obtain this information, it is necessary to do a full census. This is particularly the case with the areas of responsibility of the provinces and municipalities, who are vitally interested in smallarea detail.

Our principal problem is not only that people would not be satisfied with a sample, but that the demand for full-count small-area detail is such that we have difficulty in preventing the disclosure of information about individuals in providing data of sufficient detail.

I might mention by way of qualification that the 1971 census, as in the past, will have some of the questions asked on a one-third sample basis simply to reduce the burden.

Mr. WILSON. Mr. Duffett, I have a few other questions that will take a little more time, and I wonder if we might submit them to you. When you have time, could you send us back the answers?

Mr. DUFFETT. Yes; we will be glad to do that.

Mr. WILSON. Mr. Rousselot?

Mr. ROUSSELOT. Did I understand you to state that one of the reasons for higher cost in your enumeration program was that you thought that you included more data or requested more data on housing; is that correct?

63-434 0 - 72-3