MEMORANDUM FOR THE RECORD

From: Benny L. Kass, assistant counsel

Subject: Results of subcommittee questionnaire.

On August 11, 1966, Senator Edward V. Long, chairman of the Senate Subcommittee on Administrative Practice and Procedure, sent a detailed questionnaire to all departments and agencies of the Federal Government to determine the amount, nature, and use of information which Government agencies currently maintain on individuals. Upon receipt of the agency answers, the subcommittee arranged with the General Accounting Office to borrow the services of one of their supervisory accountants, Mr. John H. Carney. Mr. Carney performed a most valuable function by collating and compiling all of the information. A memorandum summarizing Mr. Carney's work follows (the views expressed in this memorandum are the views of Mr. Carney and do not necessarily represent either those of the General Accounting Office or of the Senate Subcommittee on Administrative Practice and Procedure):

MEMORANDUM

To: Benny Kass, assistant counsel, Subcommittee on Administrative Practice and Procedure.

From: John H. Carney, Supervisory Accountant, General Accounting Office.

Subject: Summary of work performed in the analysis of Senate questionnaire.

My objective in undertaking this project was to review and analyze in detail the completed Senate questionnaires returned by the various Federal departments; independent agencies; and selected boards, committees and commissions and designed to obtain information currently maintained about identifiable individuals in the files of Federal agencies in order to evaluate this information and determine whether or not excessive data is required and obtained from individuals.

Since the rapid advancement in computer technology has created the definite possibility for a National Data Center, there is an evergrowing concern that the individual citizen's privacy can and will be invaded. By the mere push of a button, it will be entirely possible to coordinate information from various sources on a given individual and obtain a "cradle to the grave" report about him. Naturally, having the ability to coordinate this information has unlimited statistical merit, but a happy medium between these two computer potentials must be reached.

In light of the above, I received and analyzed the questionnaire returns and the following general observations should be made:

(1) Government contractors, consultants, and their employees are required to submit such information as the birthplace of parents, income, legal, investigational, and health data as a condition to securing a potential job and in some cases, no confidentiality is extended on this data.

(2) The Selective Service System requires all registrants under the Universal Military Training and Service Act to supply legal and in

vestigational information such as police records, security or other investigative reports, and any involvement in civil and criminal court action. All additional information such as religious or financial information is only required from those registrants seeking a deferment for such reason.

(3) Certain agencies require their top or key employees to furnish certain financial information on a "Statement of Employment and Financial Interest".

(4) Forms 57, credit reports, and agency personnel files contain an excessive amount of information about an individual such as the birthplace of his parents, his income and assets, and certain health information such as a personality inventory and alcoholism or drug addiction traits.

(5) Most medical and police forms require the individual to list his race. This would be permissible if it is used for statistical purposes only.

(6) Security clearance tend to require the individual to list most of the information cited in the questionnaire. However, since the individuals securing a clearance will have access to classified records, it is essential that this detailed information be obtained.

(7) Applicants for the Government's various grants and fellowships are required to usually divulge their police record, security or other investigative reports, and any involvement in civil or criminal court action. This information appears to be excessive and nonessential as long as all academic requirements are fulfilled.

It appears that the majority of Government forms require either nonessential or too detailed information from the individual citizen. However, for the following two reasons, the results achieved through a thorough examination of Government forms could not justify the work that would have to be performed, unless it could be statistically programed for computer use:

(1) The various forms are too voluminous;

(2) One would have to have a good working knowledge of both the agency and the purpose of each form in order to evaluate it adequately.

Also, in concluding, some mention should be made on the use of the confidentiality with respect to the Government's handling of data obtained from the individual. In many instances, the individual is required either on a mandatory or conditional basis to complete certain forms; however, the limits of confidentiality have been either nonexistent or a pledge not to disclose outside the Government has been given. This pledge cannot be too meaningful since the Government is quite a large organization.

Based on the above conclusions, I recommend that this project be directed to general trends in Government data collection methods (agency forms) rather than become "bogged down" in specific agency forms. Also, it should be suggested to the various Federal agencies that they take a hard look at their statutes or other agency regulations pertaining to confidentiality standards in order to determine that the maximum protection is being afforded the individual with respect to the personal data which he has surrendered.

LETTER FROM THE BUREAU OF THE CENSUS

Hon. EDWARD V. LONG,

U.S. DEPARTMENT OF COMMERCE,

U.S. Senate, Washington, D.C.

BUREAU OF THE CENSUS, Washington, D.C., June 1, 1967.

DEAR SENATOR LONG: The Bureau of the Census has completed, at your request, on a reimbursable basis, the processing of the returns obtained by the Subcommittee on Administrative Practices and Procedures when it took the inventory of information on individuals in the records of Federal agencies. The instructions for filling out the subcommittee questionnaire provide for each respondent sets of categories into which he classified his records. (In processing the results, the Bureau of the Census added one set.) These categories then became the basis for numeric codes which were assigned to the

responses.

The categories and the numeric codes are shown in detail in appendix A. Also shown is the column on the inventory punchcard in which is located the punch for a given piece of information.

Several tables have been constructed from these punchcards and they are attached. If there are questions about the meaning or interpretation of the tables, they may be resolved by reference to appendix A, or to the instructions which accompanied the subcommittee request for information to the agencies participating in preparing this inventory.

In interpreting the accompanying tables, it should be noted that three different kinds of numbers were tabulated. In tables 1A, 2A, and 3A, the basic count is of entire files of records. There were 1,755 different files of records, all told. In table 4A the count is of items in a record. A single file might have from one to 43 items. Overall, the 1,755 different records contained 14,673 items, an average of 8.4 items per record. Tables 1B, 2B, 3B, and 4B all provide a count of people covered. In effect this means a count of the number of individual person-records in each file. There were about 3,111,500,000 individual person-records in total. In table 4B the number of personrecords in a file is tabulated each time a file item of information is indicated. Thus, 2,926,280,000 person-records in files in which names are shown, 2,061,110,000 in files in which age is shown, and so on. Obviously, many persons are counted more than once. Some of these files contain data for persons no longer living.

A number of cautions should be applied in interpreting the tables. First, the classifications were made by the reporting agencies. Second, the agencies may have left out some files which include names of persons. Third, some agencies apparently included files which did not have names of persons, contrary to instructions (these are shown in table 4A where we see that only 1,664 of the 1,755 files have names). Fourth, the number of records in each file was given as a round estimate in most cases, and sometimes was estimated by taking 10 times the

81-314-67- -2

9

yearly accession rate. There is also the possibility that coding errors may exist, as well as data processing errors.

We also noted that at least one agency did not report at all, and there may be other exceptions.

In general, we would interpret the accompanying tables as inventory estimates (not precise counts) of the volume and type of information in the records of Federal agencies.

Sincerely yours,

ROBERT F. DRURY, Acting Director, Bureau of the Census.

DEPARTMENT OF COMMERCE

BUREAU OF THE CENSUS

APPENDIX A1

Punchcard layout, items, code categories, and definitions used in processing the inventory of information on individuals in the records of Federal agencies

(The U.S. Senate Subcommittee on Administrative Practice and Procedure took the Inventory and the Bureau of the Census processed it. The punchcards described in this layout were used for processing.)