A clear and paramount public interest in a particular behavioral research inquiry, in spite of a high cost in human privacy, can no doubt frequently be established. However, the recent emergence of behavioral science knowledge as a potential contribution to human welfare has yet to be matched with an explicitly recognized set of laws or codes or otherwise publicly expressed agreements on the value of different kinds of research. Thus, there are and will be many occasions in which conflict between the individual's claim to privacy and the larger community interest in research for the general good must be resolvedand the method of resolution must be an expression of community consensus.

This concept of consensus is not employed in any formal mechanistic way. In a sense, what is meant is that the issue of paramountcy as between private personality and a particular program of scientific research should not be left solely to the decision of the research investigator. There should be some strong element of community approbation; the delicate balancing of the colliding values involved should reflect more than a single point of view.

Community consensus can obviously be expressed in laws, judicial decisions, or political constitutions. But it demands no such formal manifestation, and can also be expressed in far more, subtle but equally pervasive ways. For example, consensus can be expressed in the values of our peers as they are articulated to us. Consensus can be formed through the stated views of our opinion leaders whether they be leaders in government or industry, in labor, the professions or the clergy. Consensus can also be reflected in the provisions of collective bargaining contracts between labor and management, in the executive orders or instructions issued by Presidents, cabinet officers, personnel directors, and administrators of all kinds.

Yet, most appropriate for scientific research-as it is for all the professions— is the expression of a consensus on values in a published and operative code of ethics. Such a code yields a triple return-it articulates the values involved, uplifts thereby the awareness and standards not only of the profession but the entire community, and can provide a means for disciplining transgressions within the profession.

Thus, in launching any behavioral research project, the investigator should first determine whether voluntary, informed consent, as well as anonymity, can be accommodated with the integrity of the research. If not, the investigator should then ascertain whether the community censensus approves the conduct of the research, under the proposed conditions, without the actual consent and anonymity of the subjects. As a minimum, this means the knowledeable concurrence of those responsible for both the research project (for example, the financing institution) and for the well being of the subject (as, for example, the administration of the college he attends). The history of public health and medicine in this country, and earlier in Europe, gives many illustrations of the establishment of just such a community consensus on the invasion of privacy for the general welfare.56

One may anticipate that, as behavioral science develops and its contributions to society increase, the democratic process may afford to it more occasions of publicly approved invasions of personal privacy.

VII. THE CONCEPT OF CONFIDENTIALITY

Whether private data are collected with consent, or without consent but with society's permission because of the perceived public interest involved, the minimal requirements of privacy seem to call for the retention of the private data in a manner that assures its maximum confidentiality consistent with the integrity of the research. Thus, the second privacy issue presented by behavioral research, as it is with all inroads on the private personality, is the issue of confidentiality.

One of the most important ways in which the concept of confidentiality in behavioral research can be served is to seek to design the research so that the responses of the persons providing the data can be anonymous; the design should avoid identifying any individual respondent with a particular response. While this should be possible in all opinion surveys, in many instances the nature of the research will require an ability to identify each respondent with the data elicited from him. This would of course be true in longitudinal studies—as of child growth and development-where respondents must be ex

56 See note 45 supra.

amined or interviewed a number of times, or in studies of several diverse sets of records which must be matched up to a particular individual.

If full anonymity is not possible in the research design," then there are several other safeguards which should be stressed to provide some degree of anonymity or confidentiality. The first, needing no more than a passing mention, is the integrity of the behavioral research scientist, which, along with his interest in science, must be assumed as a basic prerequisite. The integrity of the professional scientist will assure both his informants and society at large that he will be responsible and will maintain the confidence of any information given to him by identifiable informants. That there are occasional breaches of professional confidence at this level underscores the signficance of putting stress on the responsibility of the investigator both during his professional training and throughout his research career.

Another important safeguard for confidentiality can be provided through control techniques. For example, the identity of the respondent may be coded and separated from his response except for the code number. The code, in turn, may be made accessible only to a few of the most responsible officials, or perhaps, only on two signatures or by the use of double keys. Even as elementary a safeguard as a locked file can make for substantial improvement. Penalties within the profession may also be devised for any breach of the confidentiality which should be of the very essence of professionalism.

Another readily available step is the destruction of research data. At the very least, that part of the data which would identify any individual with any portion of it should be destroyed, and destroyed at the earliest moment it is possible to do so. Today, it is quite rare for an institution or an individual scientist to take what is now viewed as a radical step and destroy data which potentially has value over a longer time span. Indeed, behavioral scientists have strong incentives to retain all original research data.58 Such data can provide information of a longitudinal nature about the development of personality or organizations over time, the early childhood antecedents of career success, the degree of change in interest and attitude from one age to another, the effects of marriage upon personality characteristics and other fascinating problems. There are now great repositories of such data in the United States collected about individuals in schools, both secondary and college, and other institutional settings, which have been maintained because of this natural resistance of the research scientist to discard anything of such potential value. Nevertheless, the maintenance and use of this information for purposes other than that originally agreed to, and the threat to confidentiality inherent in its continued maintenance, strongly suggest that the proper course of the person or institution possessing such data is either to obtain the consent of the individual involved to its continued preservation, or to destroy the data, painful as the latter prospect may be.

It should be emphasized that neither the integrity of the scientist nor the technical safeguards of locks and codes can protect research data against a

57 It should be borne in mind that there are various degrees of anonymity in the gathering of research data, and it may be useful to distinguish between them in balancing the values of particular research with the costs in privacy that may be involved. Dr. Isidor Chein, Professor of Psychology at New York University's Graduate School of Arts and Science, in a letter to the authors making this point, identified, among the possible levels of anonymity, the following six: "(a) the particular subject is never identifiable, not even by the investigator or his agents; (b) the particular subject is temporarily identifiable, but his identity is never ascertained up to and including the point at which the data that he has provided are consolidated in some meaningful and interpretable form; (c) the particular subject is temporarily identifiable and his identity is known up to, but not including, the point at which the data that he has provided are consolidated in some meaningful and interpretable form; (d) the particular subject is temporarily identifiable and can be associated with data that are in themselves meaningful and interpretable, but his identity is not ascertained; (e) the identity of the particular subject is known in conjunction with meaningful and interpretable data, but his identifiability and identity are submerged in the treatment of the data from many subjects and his own data are never scrutinized from the point of view of interpreting or drawing any inferences about him or his behavior; and (f) the identity of the particular subject is known in conjunction with meaningful and interpretable data and these data are scrutinized from the point of view of interpreting some aspect of the individual or his behavior, but his identity is thereafter submerged in the collection of similar processes of interpretation for many subjects."

58 See, e.g., Johnson, Retain the Original Data!, 19 AMERICAN PSYCHOLOGIST 350-51 (1964). See also de Mille, Central Data Storage, 19 AMERICAN PSYCHOLOGIST 772-73 (1964). The prospect of the use of computers for central recording, storage and retrieval of research data in the behavioral sciences adds a troublesome new dimension to the protection of privacy. Computerized central storage of information would remove what surely has been one of the strongest allies of the claim to privacy-the inefficiency of man and the fallibility of his memory.

valid subpoena; such data are at present quite clearly subject to subpoena. In the last analysis, therefore, unless our laws are changed to accord a privileged status to privately given research information, confidentiality can be assured only by destruction of the data. The change in the law required to accord a priviledged status to research data can be accomplished by statue. Thus, by statute in eighteen states, a privilege has already been afforded to information received by a psychologist from his client. This statutory privilege does not, however, seem to extend to psychological research.60

61

59

62

While statutes may be desirable, they may not always be necessary. A privileged status has been afforded by the common law to communications between husband and wife, and attorney and client; privilege also inheres in constitutional doctrine-as in the privilege against self-incrimination. Thus, it is conceivable that privilege could be extended by the courts to other situations-perhaps in a persuasive case, where a research scientist was willing to resist a subpoena and risk imprisonment, in order to protect the private research data in his possession. While there is a role for the martyr both in science and in law, privilege should not be viewed as a status symbol for the scientist. It should, rather, be a protective shield for his informant. As the law now stands, however, it is apparent that the research scientist who probes in the realm of the private personality, without consent, bears a special and heavy responsibility to the subjects of his research. It is a responsibility for confidentiality which, at present, in the face of a subpoena he may find himself powerless to discharge.

Of crucial importance also to the protection of confidentiality is a sensitivity on the part of the scientist to the limited purpose for which the research data were originally obtained. It is generally accepted that research data should not be published by the investigator with identities of the individual subjects attached to the data, and there is no reason why this same ethical sense of the confidentiality, or the privacy, of the data cannot be extended to other forms of publication. Thus, it should be part of the responsibility of the research scientist not to make his research data, in which individuals are identifiable, available to others, whether such others be personnel directors, private detectives, police officers, journalists, government agents, or even other scientists.

Assuredly, one can visualize situations in which the release of research data for a use not initially contemplated would, because of the great public interest involved, be socially tolerable. But, just as certainly, it is possible to visualize situations in which it clearly would not. In the latter category, for example, obviously falls the sale of personal information to commercial organizations for subscription or mailing lists.

59 The eighteen states are: Alabama, ALA. CODE tit. 46, § 297 (36) (Supp. 1963); Arkansas, ARK. STAT. ANN. § 72-1516 (1957); California, CAL. BUS. & PROF. CODE § 2904; Colorado, COLO. REV. STAT. ANN. § 154-1-7(8) (1963); Delaware, DEL. CODE ANN. tit. 24, § 3534 (Supp. 1964); Georgia, GA. CODE ANN. § 84-3118 (1955); Idaho, IDAHO CODE ANN. § 54-2314 (Supp. 1965); Illinois, ILL. ANN. STAT. ch. 912, § 406 (Smith-Hurd Supp. 1964); Kentucky, KY. REV. STAT. ANN. § 319.111 (Supp. 1965); Michigan, MICH. COMP. LAWs § 338.1018 (Supp. 1961); Nevada, NEV. REV. STAT. § 48.085 (1963); New Hampshire, N.H. REV. STAT. ANN. § 330-A :19 (Supp. 1963); New Mexico, N.M. STAT. ANN. $67-30-17 (Supp. 1965); New York, N.Y. EDUC. LAW § 7611; Oregon, ORE. REV. STAT. 44.040 (1963); Tennessee, TENN. CODE ANN. § 63-1117 (1955); Utah, UTAH CODE ANN. 58-25-9 (1963); Washington, WASH. REV. CODE § 18.83.110 (1957).

60 A Montana statute does, however, seem to extend a limited privilege to certain types of behavioral research if conducted by a person teaching psychology in a school. The Montana statute reads as follows: "Any person engaged in teaching psychology in any school, or who acting as such is engaged in the study and observation of child mentality, shall not without the consent of the parent or guardian of such child being so taught or observed testify in any civil action as to any information so obtained." MONT. REV. CODES ANN. § 93-701-4 (6) (1964).

61 See generally 8 WIGMORE, EVIDENCE §§ 2332-41 (McNaughten rev. 1961).

62 See, e.g., Hurlburt v. Hurlburt, 128 N.Y. 420, 424, 28 N.E. 651, 652 (1891) (dictum). See also Louisell, Confidentiality, Conformity and Confusion: Privileges in Federal Court Today, 31 TUL., L. REV. 101 (1956). See generally 8 WIGMORE, op. cit. supra note 61, §§ 2290-2329. It is unlikely that testimonial privilege will be judicially extended to situations that do not fully satisfy Dean Wigmore's four conditions for the existence of a privilege: (1) the privileged communication must originate in a confidence that it will not be disclosed, (2) the element of confidentiality must be essential to the relationship of the parties to the communication, (3) the relationship is one which is to be assiduously fostered, and (4) the injury that would inure to the relationship by disclosure of the communication must be greater than the benefit to be gained from its contribution to the disposition of the litigation. Id. § 2285.

63 This, nevertheless, seems to be the situation in those eighteen states which accord the privilege only to licensed or registered psychologists. See Geiser & Rheingold, Psychology and the Legal Process: Testimonial Privileged Communications, 19 AMERICAN PSYCHOLOGIST 831 (1964).

In determining the proper limits to be placed on the availability of research data, a workable proposition may well be to confine such data to the particular research purpose for which permission was initially obtained, or to a reasonably equivalent purpose. At the least, such a proposition might be accepted as an operative rule in the absence of persuasive considerations to the contrary. Of course, it must be recognized that as an individual may consent to an initial privacy invasion, so may he waive a limitation of that consent to the original research purpose. Care must, however, be taken in such instances not to imply a waiver in situations where it may not have been intended.

As in other affairs, there is, unquestionably, a happy mean between excessive privacy and indecent exposure in behavioral research. One way to begin to establish such a mean is for the behavioral scientists themselves to demonstrate, by codes of ethics and research standards, their own acute sensitivity and concern for the problem. Psychologists have made a start on an enforceable code of ethical standards directed primarily to the client relationship. Other disciplines can learn from their example and all can extend such codes more broadly to behavioral research.

VIII. AN ETHICAL CODE

64

From the foregoing there emerges an outline of the contest between the values of privacy and those of behavioral research. The community is sensitive to both values. Our society will support, and indeed, will insist on, a decent accommodation between them. An accommodation which takes into account the ethical and legal obligations of the investigating scientist can be achieved without diminishing the effectiveness of the scientific inquiry. Scientists who are responsive to the claim of privacy will find themselves pressed to develop better and more rational research techniques. Their innate inventiveness can be expected to yield new and better research methods.

Not only will the behavioral scientists be inventive in accommodating the competing values of privacy and research, but in doing so they will be more sensitive to the complexities and nuances involved than either courts or legislatures. To be sure, however, judges and legislators do have a supportive role and can be expected to fill it either by correcting abuses or protecting the responsible investigator who operates in accordance with the ethical consensus of the community.

The supportive measures available to the law, several of which have already been mentioned, are numerous and varied. One is the extension of a privileged status to the confidential communication of private information to a behavioral scientist. Another is the provision of civil or criminal remedies for the breach of the right of privacy. A third is to assess and define the contexts in which, or the

65

64 See Ethical Standards of Psychologists, 18 AMERICAN PSYCHOLOGIST 56 (1963). 65 Remedies for the breach of this right are already available in many states:

(a) See the list of states which recognize a common-law right of privacy in Prosser, supra note 14, at 386-89.

(b) Oregon and Maryland have statutes which make eavesdropping, without the consent of all persons being overhead, a crime. Neither accords any exemption for behavioral research. Thus, in Orogen, it is unlawful to obtain any part of a conversation by an eavesdropping device "if all participants in the conversation are not specifically informed that their conversation is being obtained." ORE. REV. STAT. § 165.540 (1)(c) (17963). Violation of this Oregon statute is punishable by fine or imprisonment and renders the violator liable for damages in a civil suit. ORE. REV. STAT. §§ 30.780, 165.540 (6) (1963). In Maryland it is unlawful to use any device "to overhear or record any part of the conversation or words spoken to or by any person in private conversation without the knowledge or consent, expressed or implied, of that other person." MD. ANN. CODE art. 27, § 125A (a) (Supp. 1964).

(c) See the statutes in five other states which make eavesdropping unlawful without the consent of a party to the conversation-again without an exemption for scientific research: CAL. PEN. CODE § 653j; ILL. ANN. STAT. ch. 38, §§ 14-2, 14-4 (Smith-Hurd 1964); MASS. GEN. LAWS ANN. ch. 272, § 99 (Supp. 1964); NEV. REV. STAT. § 200.650 (1957); N.Y. PEN. LAW § 738.

(d) See also the comparable but more limited statutes in six other states: ARK. STAT. ANN. § 41-1426 (1964) (loitering for purposes of invading privacy); GA. CODE ANN. $26-2001 (1953) (peeping or similar acts tending to invade privacy); N.D. CENT. CODE § 12-42-05 (Supp. 1965) (using any mechanical or electronic device to overhear or record and to repeat with intent to vex or injure); OKLA. STAT. tit. 21, § 1202 (1941) (loitering with intent to overhear and repeat to vex or injure); S.C. CODE ANN. § 16-554 (1962) (peeping or similar acts tending to invade privacy); S.D. CODE, § 13.1425 (1939) loitering with intent to overhear and repeat to vex or injure).

(e) See RESTATEMENT (SECOND), TORTS § 286 (1965), which reflects the judicial acceptance of such statutory standards as a basis for civil liability.

68

conditions under which, the cost in privacy is either marginal or de minimis, or permissible, because outweighed by the positive gains perceived for society in particular research. A fourth measure is to preclude public officials or employees from disclosing confidential information acquired in the course of employment." A fifth approach is to develop “disciplinary proceedings" to enforce the claim to privacy against public officials in some form of mandamus or contempt," and against private professional persons through disbarment or loss of license. Still another possible supportive legal measure is to require registration for the possession of all privacy-invading devices. The alternatives are clearly varied. It should be noted, however, that the existing legislative attempts to prohibit eavesdropping by use of devices have been uniformly defective. The current statutes are either inadequate in scope or indiscriminate in application, or both. A precondition for the development of a proper balance between the values of privacy and those of behavioral research is the growth among behavioral scientists themselves, of a heightened sense of their own confidential professional relationship with their informants. One of the best ways of articulating and developing this heightened sense of the confidential professional relationship is through the development and observance of codes of ethics in which the claim to privacy is recognized.

Codes of ethics for the several disciplines of scholarship and research are sound and sensible, and such codes should be general rather than specific, simple rather than complex. A workable code of ethics should be subject to expansion, interpretation, and application in specific cases according to the distinctive character of the research situation.

In accord with this view, seven principles are suggested for inclusion in a general code of ethics for behavioral research:

One: There should be a recognition, and an affirmation, of the claim to private personality.

Two: There should be a positive commitment to respect private personality in the conduct of research.

Three: To the fullest extent possible, without prejudicing the validity of the research, the informed, and voluntary, consent of the respondents should be obtained.

Four: If consent is impossible without invalidating the research, then before the reseach is undertaken, the responsible officials of the institutions financing, administering and sponsoring the research should be satisfied that the social good in the proposed research outweighs the social value of the claim to privacy under the specific conditions of the proposed invasion. These officials in turn are responsible, and must be responsive, to the views of the larger community in which science and research must work.

Five: The identification of the individual respondent should be divorced as fully and as effectively as possible from the data furnished. Anonymity of the respondent to a behavioral research study, so far as possible, should be sought actively in the design and execution of the study as a fundamental characteristic of good research.

Six: The research data should be safeguarded in every feasible and reasonable way, and the identification of individual respondents with any portion of the data should be destroyed as soon as possible, consistent with the research objectives.

Seven: The research data obtained for one purpose should not thereafter be used for another without the consent of the individual involved or a clear and responsible assessment that the public interest in the newly proposed use of the data transcends any inherent privacy transgression.

66 See, c.g., Antitrust Civil Process Act § 4 (c), 76 Stat. 550 (1962), 15 U.S.C. § 1313(e) (1964); N.Y. EDUC. LAW § 1007; N.Y. LAB. LAW § 537; N.Y. PEN. LAW § 762; N.Y. PUB. OFFICERS LAW $ 74 (b).

67 The Swedish Ombudsman suggests another interesting possibility. See A State Statute to Create the Office of Ombudsman, 2 HARV. J. LEGIS. 213 (1965).

6s Maryland, by House Bill 1197, approved by the Governor on April S. 1965, added a new 125D to Article 27 of its Annotated Code and thereby became the first state to require "every person possessing any eavesdropping and/or wiretapping device" to register such device with the State Police. Unless registered it is unlawful to manufacture or possess any such device. It will be interesting to see how vigorously and effectively this new statute is enforced. Will it be applied, for example, as it would seem was intended, to the manufacturers of tape recorders or dictaphones? Or to the lawyers or scientists who use them?

77-577-67-15