gories of questions required to be answered under penalty of the law in the decennial census of population, unemployment, and housing.

The Federal Statistics Users' Conference is an association comprising 167 organizations generally classified as business firms, labor unions and nonprofit research groups who have a common interest in obtaining adequate, timely, and reliable information from Federal statistical programs. Its membership is highly diversified and it is unique in that almost every segment of the economy that uses Federal statistics is represented in the Conference.

The principal activity of some of our member firms is economic and market research. Most of the other members engage in economic and market research, although this is not the principal purpose of their firm or organization. In both instances, a number of them conduct research through the use of surveys, both local and national in scope.

A national census is a recognized need of all nations, and mandatory reporting is accepted as a basic requirement. The decennial census of population is our most important statistical activity, having its origins in the Constitution of the United States and having been conducted in each decennium beginning with the Census of 1790 conducted under the direction of Thomas Jefferson, the Secretary of State.

As our nation has grown, there has been a great increase in the magnitude and complexity of economic and social problems. A great deal of additional and new information is now required to assist and guide decision-makers in finding solutions to these problems.

Despite a greatly enlarged statistical base, a regular reading of the Congressional Record reveals many instances where Senators and Representatives have pointed up the lack of adequate or sufficient statistical information to guide the Congress and the Administration in making important policy decisions.

Congressional Committee hearings on various social, economic, health, welfare and other legislative programs draw extensively on census statistics to define problems and to guide legislators in making decisions. These hearings often indicate our statistical gaps and needs.

The Federal government is recognized as the principal user of our statistical output that is essential for determining and carrying out public policy and programs. For this reason, the Federal Government's needs enjoy first priority so far as changes, improvements or expansion of our statistical base is concerned. In our opinion, the Bureau of the Census has scrupulously adhered to this principle in carrying out its many programs.

Just as governmental needs have increased, so have those of an expanding number of statistics users outside of government. The needs of the latter are receiving greater attention in the shaping of satistical programs, but even more consideration could and should be given to certain of these needs for the following reasons: Census data provide a foundation for the making of many vital economic and policy decisions by private enterprise that have an impact on the total economy and thereby affect the general welfare. The myriad of decisions made daily affect the level and direction of change in our economy. It is in the public and national interest that the information upon which these decisions are based be adequate, reliable and complete.

Those acquainted with census history know that for many years it was the Congress that had the responsibility for determining census inquiries. Then, some decades ago, Congress transferred that responsibility to the Executive Branch. We believe this was a wise decision because the planning and conducting of a census is an executive, administrative function involving many technical details. Experience has shown that the nation is best served when the censuses are devised by those who are experienced in census taking and in survey techniques; by those who devote full time to studying and determining which questions are most important and most feasible, and by those who are skilled in knowing the types of inquiries which are acceptable and those which are not. The Congress, however, and properly so, has the responsibility for overviewing our statistical programs and making recommendations.

When the responsibility for determining census inquiries was first transferred to the Executive Branch there was an obvious invitation and potential for an increase in respondent burden and for “bureaucratic prying" into private affairs. However, this did not come about. Current censuses are less burdensome in many respects and numerous unnecessary, prying questions are no longer included in the census. The census does not ask, as it did in the past, such questions as

whether a person is diseased, requiring details as to nature and degree; whether a person is defective in mind, sight, or speech; whether crippled, maimed or deformed; or whether a convict or a pauper.

As questions have been eliminated, others have been added to meet new needs, and also to satisfy a greatly expanded use of census data. New questions have not been included at the whim of bureaucrats, or by pressures of special interest groups, or because of any interest in more statistics merely for statistics sake. Rather, census inquiries have been dictated by the emerging needs of our society; and those needs have been carefully screened through a comprehensive structure of advisory groups.

Neither are more statistics, per se, an objective of the Federal Statistics Users' Conference. One of our primary functions, as spelled out in our By-Laws, is "To coordinate efforts of Federal statistics users in developing Federal statistical programs of optimum usefulness at minimum expense.”

Census questions are based on established and proved needs. They are subjected to intensive scrutiny by many interested, responsible groups. As this Subcommittee knows, it is the general practice of Federal agencies that collect and produce statistics to examine and evaluate past needs against current needs and to eliminate data or statistical series that are no longer required or essential. We know that the Census Bureau would not continue to collect data that are not of broad, general use nor in the public or national interest. We also know that the Census Bureau has strongly resisted the addition of new questions where the needs and uses have not been amply demonstrated or proved. Furthermore, we know that the statistical agencies of the Federal government have limited resources and rather than waste those resources on the collection and production of unessential information, they would much rather use them for improving the content and quality of the essential data they do collect.

The general practice of regular examination and evaluation of original needs against current needs for statistical data often reveals that the data are being put to many new uses which were not contemplated at the time they were first collected.

Currently, there are many organizations, mechanisms, and avenues available to evaluate the types of questions included, the degree to which they are essential, the reporting burden on respondents, and to recommend or actually place limitations on them. Among these are the Census Bureau itself through its own system of program evaluation and its recognition of its public responsibility; the various advisory committees to the Bureau of the Census; inter-agency committees of the government and particularly the Federal Council on the Decennial Census established by the Office of Statistical Standards in 1965 for the purpose of providing a mechanism for some 40 Federal agencies to express their interests in data collection and tabulation in the 1970 Census of Population and Housing; the Office of Statistical Standards of the Bureau of the Budget and the Advisory Council on Federal Reports; the House Subcommittee on Census and Statistics; the Subcommittee on Economic Statistics of the Joint Economic Committee; the appropriations subcommittees of the House and Senate: and various outside groups and organizations, including the Federal Statistics Users' Conference.

We believe this constitutes an excellent system for overseeing the work and plans of the Census Bureau with regard to the decennial censuses as well as for all our statistical programs. Furthermore, Mr. Chairman, we commend this Subcommittee and its parent Committee for the excellent job they have done, and are doing, in carrying out the responsibility of overviewing our statistical

programs.

While H.R. 10952 is concerned with limiting the number of mandatory questions included in the decennial censuses, its author, Congressman Betts, has outlined a program, in a statement in the Congressional Record, for the complete elimination of certain questions in the decennial census or for placing some of them on a more limited sample basis.

The question of limiting the number of inquiries in the decennial census and the limiting of some to a sample basis is not new and has been considered many times. The problem was seriously considered by the Intensive Review Committee of the Department of Commerce in 1953-1954. The mission of that Committee was to review, appraise, and make recommendations on programs of the Bureau of the Census. It gave careful consideration to the needs and uses made of census data. The Committee Report was issued in February, 1954. Parts of that Report, I respectfully suggest, are pertinent to, and significant for, an evaluation of some of the issues and problems that are implicit in the bill under consideration by the

Subcommittee. I would particularly refer you to page 23 of the Report of that Committee.

Since that Report was issued, a considerable amount of study and consideration has been given to this question, and others, by many interested, responsible groups regarding the 1960 Census and plans for the 1970 Census. As a result, we believe that the Census has been kept within reasonable limits and confined to the most essential questions that are required on a mandatory basis. We believe also that it is clearly understood which inquiries are needed on a full count basis and those which can yield valid, reliable data on a sample basis. The intensive study and work of these various groups should not be discounted, discredited nor discarded. If there is some question about their judgment, then we would recommend that their work be carefully reviewed and evaluated.

Certainly we appreciate the sincerity of the press, the public, and some Congressmen who have expressed concern and alarm about some parts of the 1970 Census of Population, Unemployment and Housing. However, their fears exist because they are unacquainted with the historical development and evolution of the Federal statistical system, with its operation, and the means already available to deal with the very problems which concern them.

It is important that there be a clear understanding of the basic needs and uses for decennial census data. The data provide a substantially complete count or enumeration of population, and an indentification of certain basis characteristics of such population, in given areas for a given time. These data are absolutely essential for use as benchmarks in making sample surveys, projections, forecasts, and various types of special estimates. Starting from these total, as reference points or benchmarks, the analyst may, by sampling or estimating methods, derive measures for an area or category not explicitly shown in the full count; or he may make projections into the future or prepare estimates for dates falling between past censuses.

Because of the many sample surveys being conducted by government agencies, these decennial census data are absolutely essential for benchmark purposes. They are also essential to private industry. The uses which our members make of census data were specifically spelled out in testimony of the Conference before the House Post Office and Civil Service Committee at its hearings last year on Plans for the 1970 Census.

To indicate more specifically the needs, uses, and value placed upon decennial census data by some of our members that are engaged in certain types of research activity, I would like to have included in the record at the end of my statement, with your permission, Mr. Chairman, copies of letters of three FSUC members written in response to an inquiry from Congressman Betts regarding techniques employed in gathering data by private research organizations and by the Census Bureau.

The basic question involved in the bill under consideration is that of mandatory vs. voluntary response. In 1790, the Congress, in its wisdom, decided that census data should be collected on a mandatory basis. In the past 177 years, the Congress has placed no limitation on that authority. In fact, in certain instances, it has broadened it. This was certainly not done without considered study and judgment. Furthermore, we believe that the Census Bureau has not abused its authority to require mandatory response to its inquiries.

We believe that data collected on a mandatory basis are more complete, valid and reliable than those collected on a voluntary basis. On the other hand, we recognize that it would be impractical, unnecessary, uneconomical, and not in the public interest, to collect, on a mandatory basis, all the data now being assembled by the Federal government. An extremely high percentage of information is now being collected by government agencies on a voluntary basis because sufficient valid results are obtained for the purpose for which the data are intended. However, this achievement has been made possible only because of the availability of a limited, but essential amount of data collected by the Census Bureau on a mandatory basis so as to provide required benchmarks.

Without these benchmarks, or if there should be a serious curtailment of much useful benchmark data to tie to those that have been collected over the years, many Federal government departments and agencies would be severely handicapped in their conduct of sample surveys with the result that their statistical output and projections and forecasts would be of questionable value and quality. The loss or curtailment of current benchmark data would be a severe handicap to those engaged in economic and market research and analysis and deprive management of essential data needed to guide them in overall decision making.

It would be a great loss to state, county and municipal governments that employ the whole gamut of data yielded by the Census of Population and Housing.

The implications of the bill before you are widespread, the ramifications many, but the central issue is clear. In our opinion, passage of this bill would be the most damaging step Congress could take regarding our statistical system and it would reverse the intent and policy of the Congress which has been repeatedly expressed over the long history of the Census.

This bill, if enacted into law, would seriously disrupt the conduct of the census; it would jeopardize and diminish the value of statistical data that are critically needed particularly with regard to many new legislative measures connected with the Great Society programs.

The purpose of this bill, however well intended, is self-defeating because it would directly result in greater costs, inefficiency and a lessening in the effectivness of programs of Federal, State and local governments. In the long run it could result in greater burden for respondents.

Data collection has become a necessary part of our system of government and involves two responsibilities. The first is that of the citizenry to provide certain information to its government that is essential as an aid in making wise public policy decisions and in carrying out governmental programs. The second is that of the government to confine inquiries to those which are considered most essential, which impose a minimum of reporting burden, with a guarantee and respect for the confidentiality of individual data provided. Inherent in our entire statistical program is the recognition of this governmental responsibility, and we doubt if our statistical system could have reached its present level and state development if scrupulous attention had not been given to this responsibility. Many means have been provided for ensuring that the government is meeting this responsibility. We would certainly welcome and support any proposals or recommendations that would further assist the government in this regard. The present bill, in our opinion, does not fit this category.

Thank you, Mr. Chairman, for the opportunity to present the views of the Conference.

HOUSE OF REPRESENTATIVES,
Washington, D.C., June 19, 1968.

Hon. WILLIAM J. GREEN,

Chairman, Census and Statistics Subcommittee,
Washington, D.C.

DEAR MR. CHAIRMAN: I would appreciate it if you would include in the record of the Subcommittee's hearings a copy of my speech delivered before the House, June 18, 1968, relating to the Census.

With kindest personal regards, I am,
Sincerely yours,

ARNOLD OLSEN, Member of Congress.

SENSE ABOUT THE CENSUS

Mr. OLSEN. Mr. Speaker, the time has come, in fact it is overdue, for someone to speak up and set forth the facts about the 1970 Censuses of Population and Housing. For some time now, we have been hearing and reading too much misinformation, distortion, and exaggeration about these censuses, and I cannot in good conscience let these distortions stand.

As a member of the Subcommittee on Census and Statistics for eight years, I have had a unique opportunity to study at firsthand the work of the Census Bureau, so I feel I can speak with knowledgeable authority. It is a professional agency of the United States Government that is recognized throughout the world as having the most advanced statistical techniques. In fact, the United States Bureau of the Census has, in part, become an international university for training in the techniques of census and survey taking, with 1,300 nationals from 90 foreign countries having been trained by the Bureau in the last 22 years. Supplementing this effort, the Census Bureau, at the request of the governments concerned, has sent 300 of its technicians to 70 foreign countries to assist them in the development of their statistical programs. In addition, hundreds of nationals from practically every country in the world have visited the Bureau for consultations, conferences, and seminars.

I am proud to be a member of this great legislative body which has given bipartisan support to the censuses throughout the history of our Nation, making

it possible for the United States Bureau of the Census to develop the most advanced statistical programs in the world.

But all of their scientific competence would be of questionable practical application were it not matched with equal concern in dealing with people. Census employees visit homes, businesses, and industries and it is because of the confidence the Census Bureau enjoys among our peoples, as much as it expertise in statistical and computer science, that the Census Bureau is able to collect and process quickly the facts we need to have about ourselves: our social and economic characteristics, the houses we live in, the State and local governments that protect us, the industries that employ us, the businesses that serve us— facts that make possible intelligent decisions not only by the Congress and the Administration, but by our State and local governments, our businesses, and our industries.

So it is with amazement that I hear voices raised objecting to vital aspects of the 1970 decennial census as it is proposed to be conducted by the Bureau. These voices are crying out that the census is unduly coercive, an invasion of privacy, burdened with nonessential questions, and that results will be chaotic if the census bureaucrats are allowed to continue their nefarious scheming.

What, specifically, is it that is worrying these people? For one thing, they say that many of the questions have no relation to the constitutional purpose of the census, and that the number and personal nature of such inquiries have increased almost unnoticed and unobstructed.

CONGRESS BROADENED SCOPE OF CENSUS

Let us review a little history here. From the very beginning, the census was not confined to merely a count of the number of persons. The very First Congress decided not to limit the 1790 census to the bare essentials needed for apportionment stipulated in the Constitution. Here is what the U.S. marshals were instructed by the Congress to do:

"*** to cause the number of inhabitants within their respective districts to be taken; omitting in such enumeration Indians not taxed, and distinguishing free persons, including those bound to service for a term of years, from all others; distinguishing also the sexes and colours of persons, and the free males sixteen years and upward from those under that age ***” The additional information was desired for the purpose of ascertaining the military strength of the country.

So it was that the value of the census machinery as a means of obtaining essential and accurate information about the Nation beyond the purpose set for in the Constitution was recognized by the First Congress. And it was the subsequent Congresses, working with distinguished leaders of the various Administrations, that further expanded and modified the scope of the census to meet the growing and changing needs of the Nation. James Madison, Thomas Jefferson, John Quincy Adams, and Martin Van Buren were among those instrumental in expanding the scope of the early censuses.

The Congress enacted a separate authorization for each census up to 1930, before turning the authority for determining questions and procedures over to the Census Bureau. The Census Bureau has not, since that time, gone beyond the precedents established by the Congress; on the contrary, in spite of the greater needs of modern society for statistical data, they have succeeded, by judicious pruning and the use of sampling, in making the census trimmer, easier, and better.

In looking over the lists of questions asked in early censuses, I find in the census of 1820 the first indication of the need to know characteristics of the population other than age, sex, free or slave. In that census Congress asked whether each person was engaged in agriculture, commerce, or manufactures, and, if foreign born, whether naturalized.

A decade later came the first recognition of health and social problems. At every household, the census taker asked if there were persons who were deaf, dumb, or blind. Then in 1840, Congress became concerned with mental illness, and the crude question: "Is there anyone living here who is insane or idiotic?" was asked of every household. That question remained in the census questionnaire each decade through 1890, by specific direction of the Congress.

It was in 1840 that the first questions about education were asked. How many people over 20 could read or write? How many schools were there, and how many scholars?