1. Preamble: Limited Confidentiality at SFU *

2. The Policy Statement’s Ethics Review Criteria *

3. The Values Underlying Confidentiality Are Not Absolute – But Is This A Reason To Require A Priori Limitation of Confidentiality? *

4. Designing Research Protocols in a Way That Maximizes the Rights and Interests of Research Subjects and Respects the Law *

5. What if the Supreme Court of Canada Were to Order Disclosure? *

Appendix A: Summary of US cases from the Law and Contemporary Problems Special Issue (1996) On Court Ordered Disclosure, where Confidentiality of Participants was at Issue *

Implications of the Tri-Council Policy Statement 1

September 1998 saw release of the long-anticipated Tri-Council Policy Statement entitled Ethical Conduct for Research Involving Humans (hereafter referred to as the " Statement" or "Policy Statement"). The 90-page Statement is a complex document that gives researchers much to consider. Any two or three-line summary of its contents is doomed to distort the complex balancing of considerations the Policy Statement both reflects and requires.

Problems with how to balance various aspects of the Policy Statement’s principles and its numerous exemption clauses will no doubt surface as institutions wrestle with the problem of bringing their own procedures into compliance by September 1, 1999. These problems will be all the more difficult because the Tri-Council has not devised a mechanism for providing institutions with assistance in interpreting the Policy Statement, even though the three granting councils have reserved the right to withhold funds from institutions that do not comply! And although its authors say they "welcome comment and discussion, and commit to regular updates" of the Policy Statement (p. i.1), they do not provide an address for receiving such feedback! It is in this light [or darkness?] that the SFU Research Ethics Policy Revision Task Force (and comparable bodies in other universities) must devise a new policy that complies with the Policy Statement. We can anticipate that many debates will arise about how the Policy Statement should be interpreted. Indeed, the purpose of this, our second submission to the Task Force, is to contribute to that debate by challenging recent interpretations of the way the SFU Policy Statement deals with conflicts between research ethics and the law.

The issue that concerns us most is the policy of "limited confidentiality" that was instituted by SFU's VP-Research and University Research Ethics Review Committee (URERC) in response to the Russel Ogden case. Its manifestation in practice is that, since 1995, SFU researchers obtaining information from subjects2 about activities that may be in violation of criminal or civil law have been required to limit confidentiality by informing their subjects that:

The imposition of this statement has been the subject of controversy at SFU for the past eighteen months,4 and has had a considerable impact on research in the School of Criminology. In The History of Limited Confidentiality at SFU,5 our first submission to the SFU Ethics Task Force, we described the litany of procedural errors and policy violations surrounding the inception and imposition of the limited confidentiality consent statement, and our reasons for believing that its use is unethical. In the current submission, we set those procedural problems aside, and consider the place of confidentiality, and the a priori limitation of confidentiality, in the Tri-Council Policy Statement.

We begin by considering whether the imposed a priori limitation of confidentiality as practised at SFU is consistent with the research ethics laid down in the Policy Statement. We conclude that it is not. In contrast, the two principal architects of the limited confidentiality consent statement – Bruce Clayman, former Chair of the URERC, and Adam Horvath, the current Chair – have argued that SFU-style limited confidentiality is consistent with the principles laid out in the Policy Statement. In a September 24, 1998 article in the Simon Fraser News, for example, VP-Research Bruce Clayman stated that:

Similarly, in a letter to us dated October 8, 1998, Adam Horvath stated that:

When it comes to court-ordered disclosure of confidential research information and protecting confidentiality to the extent possible within the law, we flatly disagree with these conclusions. We believe that the SFU limited confidentiality consent statement will no longer be permissible once the Policy Statement takes effect.7 Furthermore, we would suggest that the strategy of picking out a few lines from the Policy Statement as Dr. Clayman has done or offering only a conclusion without any statement of rationale as Dr. Horvath has done, is anathema to the approach the Tri-Council advocates:

Research confidentiality is included among the Policy Statement’s core principles:

The Policy Statement says that researchers are "honour bound" to maintain confidentiality "to the extent possible within the law" (p. 3.2). As academic research subjects are not protected by a statutory privilege, they will have to turn to the Wigmore test,8 the common law test recognized by the Supreme Court of Canada as the appropriate measure for establishing privilege of communication on a case-by-case basis (R. v. Gruenke, 1991).9 As far as we can ascertain, this test is the only legal device currently available to academics in Canada to protect confidential research information, particularly the anonymity of subjects, against court-ordered disclosure. For reasons that we have explained in The History of Limited Confidentiality at SFU (sections 4 and 5) and as we elaborate below, the SFU limited confidentiality consent statement compromises a researcher’s ability to use the Wigmore test to protect confidential information "to the extent possible within the law." In this light, we suggest that it is not possible to reconcile the limited confidentiality consent statement with the principles elaborated in the Tri-Council Policy Statement.

In this paper, we outline an approach to confidentiality that reconciles the ethical obligations the Policy Statement espouses with other important obligations such as the provision of informed consent. Our position with respect to research confidentiality is straightforward – we believe researchers are ethically obliged to do their utmost to preserve it, and that to do otherwise makes their interactions with subjects exploitative and unethical. We explain how these protections are most effectively provided in a way that is consistent with the principles laid out in the Policy Statement.

The submission is divided into five sections, including this preamble. In the second section, we consider structural aspects of the ethics review process and the Policy Statement’s criteria for ethics review. In section three, we examine arguments about limiting confidentiality, and show why the SFU limited confidentiality consent statement does not comply with the ethical principles articulated in the Policy Statement. Also we draw a distinction between a priori limitation of confidentiality and violating a guarantee of unlimited confidentiality in the interest of preventing prospective serious harms, such as murder, and between mandatory reporting requirements and court ordered disclosure. The fourth section considers how to use the Policy Statement in conjunction with the Wigmore test to protect confidential information to the extent possible within the law. In the final section, we consider the options open to researchers should the Supreme Court of Canada order disclosure of the identity of research participants. Throughout the document we draw on the experience of US researchers with court ordered disclosure to indicate the serious problems that could occur if researchers were to use the SFU limited confidentiality statement south of the border, and how these same considerations might well apply in Canadian courts as well.

To set the stage for our discussion of the two alternatives that have been debated at SFU over the past fifteen months, we describe the Policy Statement’s criteria for research ethics review, and the requirement that Research Ethics Boards must operate at arms length from the parent institution’s University’s administration.

Much of the concern that gave rise to the striking of the Tri-Council issued from concerns about bio-genetic research. The prospect of big money from pharmaceutical windfalls, genome patents and the like gave rise to concern about the extent to which universities and university researchers caught up in an entrepreneurial spirit had, perhaps, forgotten their allegiances and responsibilities to society in general, and to their research subjects in particular. In this regard, Dr. Michael McDonald, a member of the Tri-Council Working Group that created the 1996 and 1997 drafts of the prospective Tri-Council Code, offered the following reflections in the process of commenting on the 1998 Policy Statement:

• Increasing private sector dollars pouring particularly into medical research, much of this in the private sector
• Attendant pressures on REBs to issue quick and favourable verdicts on research proposals10

Likewise, the Policy Statement observes that University administrators may feel tempted or pressured to approve projects in order to maintain co-operative relations with private companies, and/or when the possibility of large rewards through patents, research funding, endowments etc. loom ahead:

These issues are no less relevant to the social sciences, where such activities as consultative relationships, evaluation research, and the development of standardized tests for assessment and classification of research subjects have been financially fruitful for both university researchers and universities. At the core of these concerns is the issue of conflicts of interest, i.e., where a financially concerned researcher and/or research institution is placed in the position of weighing the interests of its sponsors with those of research subjects.

In the social sciences, we see fiscal interests cutting both ways. On the one hand, a university might wish to "fast track" some projects because of the financial benefits to be obtained. On the other hand, there are also projects and responsibilities a university might wish to thwart or impede, because the ethical responsibilities it bestows may expose the university to what it deems to be risk. One example would be research where protection of subjects' rights and interests to confidentiality is seen as a potential liability to be avoided, not as a positive opportunity to defend academic freedom. The SFU administration brought exactly this perspective to bear when it decided to abandon Russel Ogden’s research subjects in court (see Blomley and Davis, "Russel Ogden Decision Review," 1998).11 Also it appears to have been one of the main guiding forces, if not the main force, behind the imposition of limited confidentiality (see Lowman and Palys, The History of Limited Confidentiality at SFU, section 3).

The Policy Statement is unequivocal in its requirement that these kinds of institutional conflicts of interest should be avoided and, in this sense, the document can be seen as a way of protecting research subjects and researchers from the institutions that have power over them. Similarly, researchers should avoid conflicts of interest.

The Policy Statement reminds researchers that research subjects – people who we approach and who volunteer their participation – are our responsibility, and that we must not engage in any relationship with their keepers, caregivers, employers, teachers, or other authorities, that distort where our research allegiances must ethically lie. For example:

And elsewhere:

The central issue here is conflict of interest, and, accordingly, the changes the Tri-Council requires will begin with the organizational relationship between those involved in the ethics review process and the university administration. Currently the SFU ethics policy requires that the VP-Research or his designate sit as Chair of the Ethics Committee and appoint its members. If the VP-Research Chairs the Committee, the VP-Academic adjudicates appeals of the URERC’s negative decisions. If the VP-Research designates the Chair s/he adjudicates appeals. In contrast, the Policy Statement requires that Research Ethics Boards (REBs),

Under this new arrangement, University administrators will no longer be permitted to have anything to do with ethics review or its administration.

2.3 Membership of the Ethics Review Committee12

With respect to the disciplinary composition of REBs, the Policy Statement notes that researchers on the committee, "should have both the training and the expertise to make sound judgements on the ethics of research proposals involving human subjects." Also, it says that, "The terms of REB appointments should be arranged to balance the need to maintain continuity with the need to ensure diversity of opinion…" (p.1.3). We applaud this recognition of the need for disciplinary and epistemological diversity in the appointment of REB members, and in the consideration of proposals for ethics review.

An enduring problem at SFU over the last decade has been the lack of disciplinary and epistemological heterogeneity in appointments to the URERC. SFU’s current ethics policy requires that members "be drawn primarily from Departments or Faculties whose members frequently conduct research with subjects." For the ten years prior to August 1998, eight out of thirteen such departments and faculties at SFU were never represented on the committee.13 And apart from one member who, in 1997, resigned in protest of the Committee’s actions, no practitioner of qualitative research traditions was ever appointed to the committee. This situation mirrored the problems that were revealed in the Tri-Council’s first policy draft (released in 1996). Although there may well be ethical principles that transcend disciplinary boundaries, the way they are realized in different research situations makes it highly problematic when those who have no experience with certain research contexts create rules to govern that research. An applied rule that makes sense in one disciplinary context may create huge injustice in another. Similarly, solutions that make sense in one disciplinary context may make no sense in another.14

At SFU two significant problems have arisen from the URERC's limited disciplinary representation. First, by imposing an a priori limitation of confidentiality without representation from many of those most directly affected by it, the Committee went beyond strictly "ethical" considerations to violate the academic freedom of other researchers. Second, the imposition of limited confidentiality was both formulaic and inconsistent. It was "formulaic" to the extent that only researchers collecting information from subjects that may be in violation of criminal or civil law were required to limit confidentiality. It was inconsistent insofar as the American experience with court-ordered disclosure – to the extent we can use this as a harbinger of what might occur in Canada – shows that such cases comprise only a small proportion of research areas in which such issues arose. The modal scenario is one involving litigants who compete over access to data gained in a clinical, experimental or consultative context, a scenario the SFU URERC has not addressed.

The Policy Statement addresses these difficulties. In addition to the requirement that Research Ethics Boards (REBs) must reflect a diversity of views, the Policy Statement directs REBs to avoid "formulaic" approaches to ethics review, and "to avoid imposing one disciplinary perspective on others." (p.i.2). Epistemological imperialism at SFU and elsewhere is to be a thing of the past.

Criticism of the earliest Draft Code of Ethics produced by the Tri-Council Working Group (1996) focussed in large part on its heavy-handedness. The Tri-Council Policy Statement has settled for a lighter touch. Intervention in the research process is to be measured, and minimized. Toward this end, the Tri-Council advocates a "proportionate approach," in which the level of risk to subjects is assessed, and the degree of warning and intervention required by Ethics Committees is proportionate to that risk.

In cases of minimal risk, the degree of REB intervention is expected to be minimal. "Minimal risk" is to be defined from the perspective of research subjects:

This approach highlights the differences between the Policy Statement and the approach maintained by the current SFU Ethics Committee, one example of which is a proposal the two of us submitted for ethics review. The research we proposed involves studying managers of escort agencies and other venues of off-street prostitution. What are the confidentiality risks associated with such a project?

From the perspective of subjects, our interest in their worlds poses no greater risk than they run in the course of their daily lives, i.e., a situation consistent with the Policy Statement's definition of "minimal risk." With respect to the likelihood of us being subpoenaed to testify, we again conclude the risk is clearly minimal; in Canada, there has only ever been one case in which a researcher was subpoenaed and asked for confidential information. If, however, we were to be subpoenaed, what is the risk a court would compel us to share confidential information? Obviously the risk is minimal. As far as we know, there has never been a case in Canada in which a researcher has been ordered to reveal confidential information. And, when we invoked the Wigmore criteria as a legal explanation in order to protect the confidentiality of our respondents, the idea that a court might value the conviction of a prostitute or pimp more than the research enterprise is highly unlikely. And even if the Supreme Court of Canada were to order us to reveal confidential information, given the guarantee of confidentiality we must give to satisfy the first criterion of the Wigmore test, we see no ethical course but to defy the order and accept the consequences. Each one of these possibilities is highly unlikely, and the joint probability of them all occurring in sequence would seem incalculably low. If we take the Wigmore test seriously, particularly the first criterion, the risk to subjects is minuscule.

According to the principles laid out in the Policy Statement, it would be alarmist and excessive to require a researcher to utter warnings about the possibility of subpoena in our situation. Indeed, in another document we submitted to the Ethics Committee, entitled "Implications of the Tri-Council Policy Statement for Balancing the Interests of Informed Consent and Confidentiality"15 we described how the Tri-Council waiver provisions would apply to our research (this argument is reviewed section in 2.5 below). It is thus instructive to use the Policy Statement to evaluate the URERC’s reasons for rejecting two of our ethics applications. These applications were rejected after 8+ and 11+ months respectively because of problems created by the Committee’s requirement that we mention the minute risk of subpoena in our informed consent statement. We are now in the process of appealing these decisions.16 By causing this lengthy delay the URERC has compromised our ability to conduct this research, and may have destroyed one or both of these projects altogether (we have lost several key contacts over the past year).

Given the URERC’s requirement that we use an informed consent statement that could deny our right to legal representation by the university17 and would undermine our ability to protect research subjects in court (see section 4.4.i, below),18 we can only surmise that the URERC's concern had as much or more to do with protecting the university and researchers from liability than it did with protecting research subjects from harm. Similarly, reflecting on the experience of US researchers who have received subpoenas, Michael Traynor, a US lawyer, observes:

According to Traynor’s analysis, then, while limited confidentiality may protect researchers and universities from liability, it is likely to expose research participants to harm.

In following a proportional approach, the Policy Statement’s informed consent alteration provisions (Article 2.1(c)) explicitly state that the interests of informed consent should be balanced against "the needs for research:"

According to Article 2.1(c) of the Policy Statement, an REB can, "approve a consent procedure which does not include, or which alters, some or all of the elements of informed consent … provided that the REB finds and documents that:

1. The research involves no more than minimal risks to the subjects;
2. The waiver or alteration is unlikely to adversely effect the rights and welfare of the subjects;
3. The research could not be practicably carried out without the waiver alteration;
4. Whenever possible and appropriate, the subjects will be provided with additional pertinent information after participation; and
5. The waivered or altered consent does not involve a therapeutic intervention."

Under the Policy Statement’s standards for full disclosure, a Research Ethics Board should not formulaically require that a certain class of researchers mention the minute risk of court-ordered disclosure of confidential information. Aside from situations where mandatory laws might apply (see section 3.1.ii, below), which for us pose mostly different problems, this argument can be applied to much or all of the research where the SFU "limited confidentiality" consent statement is currently required. For example, considering the five consent alteration criteria in relation to our research on prostitution reveals the following:

Criterion (v) does not apply because our research does not involve any therapeutic intervention. Also, criterion (iv) does not apply because we do not plan to deceive anyone.

Our prostitution research meets criterion (iii) because it cannot be practicably carried out using the current SFU consent form template. Our prospective subjects interpret SFU’s limited confidentiality statement as meaning that the University expects us to divulge information if a court so orders, without contesting the order in court. Under these circumstances, we would not expect subjects to talk honestly about their undetected criminal activity. Indeed, if we limit confidentiality this way, most of our contacts will not talk to us at all. Nor would we want to talk to our research subjects in these circumstances: to gather data from voluntary subjects without taking their rights and interests to heart is exploitative and unethical. Additionally, under a regime of a priori limitation of confidentiality such as the one embodied in the SFU consent form template, our research results would not be valid.

As to criterion (i), we have outlined above the various ways in which our research is clearly "minimal risk" from the perspective of subjects. We are confident the courts will support our decision to pledge unlimited confidentiality. The Supreme Court of Canada has ruled (in Regina v. Gruenke) that the Wigmore test provides appropriate criteria for establishing privilege on a case-by-case basis. The first criterion of the test requires that, to be privileged, a communication must arise in confidence that it will not be disclosed. Having made this undertaking, it would be unethical to violate it. Also, violating it would have a devastating impact on the integrity of the research enterprise. The courts cannot have it both ways. Consequently, our bottom line is that we are prepared to accept the sanctions if the Supreme Court of Canada were to disagree with our research-ethical position. Our position is consistent with the Policy Statement, which says, "legal and ethical approaches to issues may lead to different conclusions." Researchers who choose this course are not "above the law," for, as the Statement says, "ethical approaches cannot pre-empt the application of law"(p.i.8).

As to criterion (ii), would the omission of a statement about risk of court-ordered disclosure adversely affect the welfare and rights of our subjects? No. It would do just the opposite. It would enhance a researcher’s ability to use the Wigmore test to protect their right to privacy, confidentiality and anonymity.

Given the "consent alteration" provisions of the Policy Statement, its concept of "minimal risk" and its "proportional approach" to ethics review, we suggest it would be unreasonable for any Research Ethics Board to require researchers to use the limited confidentiality consent statement that the URERC has foisted on SFU researchers. According to the principles of a proportionate approach, the SFU Ethics Committee cannot justify holding up our research for a year.

Before we started our debate with SFU's ethics administration more than a year ago, we took the University’s treatment of Russel Ogden to reflect an institutional conflict of interest. We took it for granted that all researchers would see research confidentiality as something worth vigorously defending. Our protracted debate with the VP Research and the URERC has made us think otherwise. Over the past year we have discovered that several URERC members19 apparently would have no qualms about handing over confidential information to a court, without a fight, as long as they had warned research subjects that a court might require them to do so.20

The US literature regarding court-ordered disclosure (reviewed below) shows that academic freedom and research confidentiality can be challenged in any research area. We now understand that some researchers would not be prepared to go to court and fight for the rights and interests of their research subjects. Consequently, we suggest that all researchers, not just persons collecting information from subjects about criminal and civil law violations, should be required to declare their and the University’s intention to resist court-ordered disclosure. Only if subjects are given this information can they make properly informed consent when deciding whom to trust. By not requiring that all researchers declare their intentions, the URERC has prevented the vast majority of SFU research subjects from obtaining the information they need to make properly informed consent.

Drs. Horvath and Clayman have argued that confidentiality is not absolute.21 For example, in a recent article in the Georgia Straight,22 author Charlie Smith quotes Dr. Clayman as saying:

Similarly, in his article posted on the university's web page,24 Dr. Horvath states:

Obviously confidentiality is not absolute, and we have never argued that it is. No matter how apparently innocuous the research context, it is possible to imagine a circumstance in which we come across something so heinous that a higher ethic would compel us to violate a pledge of confidentiality. In an interview a terrorist reveals that a deadly shipment of anthrax is about to be smuggled into the country. A gang member states that he has acquired a gun and intends to kill a person who is about to testify against him. Whether such scenarios are likely is another matter, but, in theory, we accept the possibility that something along these lines might occur.

Such a scenario is actually one of three that can occur that might challenge the confidentiality we would guarantee. These include: (a) heinous discovery – we find that someone is about to blow up Saskatoon, or kill someone tomorrow, or whatever; there is no law that requires reporting in this case, but there are clear moral obligations to do something to stop the event from happening; (b) mandatory reporting – in the course of our research, we come across a situation that is addressed by mandatory reporting laws, e.g., we are doing research involving children and come across a case of suspected child abuse; and (c) subpoena – a third party subpoenas our data and seeks information that, if supplied, would violate the pledge of confidentiality we have made to our subjects.

The range of options we have in such situations is to some degree constrained by facts of life. These include: (a) we can't anticipate every possible event that might occur in any given research situation; (b) there are mandatory reporting laws that we must deal with when they arise; and (c) because we are not Statistics Canada researchers, we do not have a class privilege defined by statute, but can claim it only on a case-by-case basis.

The distinction between "class" and "case-by-case" privilege is not necessarily relevant in terms of challenges to confidentiality – even persons protected by a class privilege can come across heinous situations where there are strong moral injunctions to prevent prospective harm. And presumably they, too, are subject to mandatory reporting requirements – or are they? With respect to subpoenas, neither a Statistics Canada researcher nor a university researcher holds "absolute" privilege (although those with class privilege do have a distinct advantage over university researchers who must invoke a claim of privilege on a case-by-case basis); the key distinction is where the onus of proof lies. In the case of a class privilege, the onus is on the party wanting the information to establish that it should be exempted from the class. When privilege is established on a case-by-case basis, the onus is on the party claiming the exemption to establish that the communication is deserving of privilege. In any event, the central point here is that even class privilege cannot, in theory, be considered absolute.

It is in this context that researchers must decide what how to approach heinous discovery, mandatory reporting laws and subpoenas. Also, in the process, they must comply with the principles laid out in the Tri-Council Policy Statement which says that researchers must protect confidential information to the "extent possible within the law," the minimum standard the Tri-Council has set as being ethically acceptable. It is the level that researchers are "honour-bound to protect" (Policy Statement p. 3.2).

Two distinct approaches have been proposed to deal with this set of situations and the constraints surrounding them. The approach offered by Drs. Clayman, Horvath and the URERC is to impose an a priori limitation on confidentiality whenever a researcher indicates that s/he may hear about information that is or may be in violation of criminal or civil law. We offer a different approach, one that resolves the unnecessary ethical and legal problems created by the SFU limited confidentiality consent statement.

These two approaches are considered in the next two sections. We examine both approaches in light of the ethical criteria laid out in the Policy Statement.

The alternative provided to us by the VP-Research and the Ethics Committee involves the a priori limitation of confidentiality. Superficially, the ethics of this choice makes sense. If confidentiality is not absolute, then the requirements of informed consent tell us that prospective subjects must be told as much when we solicit their participation. Let us take this as a starting point.

SFU researchers who obtain information from subjects on activities that "are or may be in violation of criminal or civil law" (ethics application screening question number 9) are required to inform subjects that the researcher may be "required" to divulge this information to a court or other legal body.

There are three problems with this statement and its application:

1. The limitation of confidentiality mentioned here involves court-ordered disclosure, not mandatory reporting or heinous discovery. However, mandatory reporting laws and heinous discovery situations are always invoked to justify use of the statement.
2. The statement contains a contradiction that potentially exposes research subjects to serious harm. The statement says that confidentiality is promised "to the full extent permitted by law," but then says, "it is possible that, as a result of legal action, the researcher may be required to divulge information obtained in the course of this research to a court or other legal body." This makes it appear as if a court order is the limit of the legal protection of confidentiality. It is not. Worse, it makes it impossible for the researcher-subject communication to pass the first criterion of the Wigmore test because, as we explain below (see, section 4.4.i), the second part of the statement prevents researchers from living up to the promise in the first part. The SFU limited confidentiality statement is thus inherently unethical.
3. The experience of court-ordered disclosure in the US suggests that, according to its own logic, the URERC should force all researchers to use the limited confidentiality consent statement, not just those collecting information from subjects about activities that may violate criminal or civil law.

We examine these various problems in more detail next.

The conflation of mandatory reporting requirements and court-ordered disclosure involves a change of subject from the concerns that ostensibly drove the URERC to limit confidentiality in the first place. When we have pointed out the problems with limited confidentiality with respect to "court-ordered disclosure," the defenders of the limited confidentiality consent statement change the subject to "mandatory reporting" laws or "heinous discovery" situations to explain why confidentiality should be limited. In Canada, mandatory reporting laws pertain to the discovery of child and elder abuse, sexually transmitted diseases, and possibly intent to murder.25 We believe these two different situations call for different responses with respect to the confidentiality issues involved.

If mandatory reporting is the concern, we wonder why the URERC restricts its imposition of the limited confidentiality consent statement to only those situations where court-ordered disclosure is perceived to be a risk. The URERC does not systematically warn subjects about mandatory reporting requirements, because it does not systematically require researchers to limit confidentiality in situations where such requirements might possibly be triggered. For example, one of the questions on the research ethics application screening form (Form 1) asks, "Will children be involved as subjects in your research?" Those who answer "yes" are told they must incorporate a "consent form for the parent/guardian" (Form 3). But nowhere on Form 3 is there any warning to subjects or their parents/guardians that there are laws requiring researchers to inform legal authorities in the event that abuse is discovered. The definition of abuse in the British Columbia Family and Child Services Act is so broad that just about anyone working with children might have reason to believe they have detected abuse. Consequently, we do not understand why, according to its own logic about informing subjects of every conceivable risk, the URERC does not require all researchers working with children to mention this risk.26

Similarly, the Committee does not include a screening question that asks whether researchers might learn about sexually transmitted diseases, even though mandatory reporting requirements exist in this domain as well.

Instead, the Committee insists that the limited confidentiality consent statement be used in situations where researchers collect information about violations of criminal or civil laws, only two or three of which, depending on the province, have any mandatory reporting requirements associated with them. The main legal complication that has prompted use of the limited confidentiality consent statement is thus risk of court-ordered disclosure, not risk of mandatory reporting.

Reliable and valid research that asks subjects to divulge information that might harm them requires unlimited confidentiality or a methodology that makes it absolutely impossible to identify subjects. In many types of research, it is not possible to do the latter. We are concerned that mandatory reporting laws preclude the possibility of conducting certain kinds of qualitative and longitudinal research where subjects may be identified. Because limited confidentiality would compromise the reliability and validity of this kind of research, three options are open to the researchers doing it:

1. Not doing the research, since doing so would expose research participants to harm;
2. Asking appropriate legal authorities for "privacy certificates" or statutory privilege, and doing the research only if these concessions can be won; and
3. Doing the research with the full intention of ignoring the law.

We suggest that (c) is not a viable ethical option. To avoid the legal and ethical conflicts created by mandatory reporting requirements, we strongly suggest that representations be made to appropriate legislative authorities to exempt academic researchers from mandatory reporting requirements. We suggest that the Tri-Council is the research authority most appropriately situated to initiate that campaign and for all universities and CAUT to support it.

We believe that the URERC's limited confidentiality consent statement is unethical because it offers a barren promise, and creates its own deception. Researcher-subject privilege is not explicitly recognized in Canadian law for anyone other than Statistics Canada researchers, who have been granted privilege in section 18 of the Statistics Act (see section 4.4.iii, below). Instead, to be considered privileged, research information must pass all four Wigmore criteria (see footnote above, and section 4 below). The Wigmore criteria describe the extent to which confidentiality can be protected legally, which in Russel Ogden’s case turned out to be "absolutely." Because privilege is decided on a case-by-case basis, Ogden’s victory does not guarantee that the next researcher will have the same success. However, his victory does establish the standard all researchers must aim for.

Stating that the researcher might be "required" to divulge information to a court makes it impossible to pass the first criterion of the Wigmore test, which requires that a communication must arise in a confidence that it will not be disclosed.27 Because of the limited confidentiality statement warning about court-ordered disclosure, the court can simply say, "the research subject was told that a court might order the divulging of information, and it is ordering it now, so hand over the information." Indeed, this is exactly what happened in Atlantic Sugar, Ltd. v the United States of America, where respondents to an International Trade Commission questionnaire were told that the information would not be disclosed "except as required by law." As it turned out, the law (in the form of the US Customs Court) required it, and because confidentiality had been limited, the information was no longer considered to be confidential.28

The limited confidentiality statement thus prevents the researcher from living up to the promise that information will be kept confidential "to the full extent permitted by law." By virtually guaranteeing that they cannot resort to the Wigmore test for protection, researchers who accept the URERC's limited confidentiality consent statement effectively give themselves and the University a license to hand over confidential information should a court ask for it. In the process, the URERC will have deceived research subjects by (mis-)informing them that they will protect confidentiality "to the full extent permitted by law," when their choice of wording makes it impossible to do exactly that. Indeed, the policy of limited confidentiality effectively ends confidentiality by making it impossible to protect confidential research information from court-ordered disclosure.

The Policy Statement specifically eschews formulaic approaches to research ethics decision-making (p.i.9). We wonder, therefore, what members of the Tri-Council would say about the formula the URERC has used in deciding which researchers it should force to use the limited confidentiality consent statement.

The URERC’s formulaic approach to limited confidentiality can be seen in its inclusion of a "screening" question (question 9) on Form 1 of the ethics application "checklist." Question 9 asks:

Only those who answer "yes" are required to use the limited confidentiality consent statement. If the URERC is sincere in its desire to ensure that research subjects are able to give informed consent, we do not understand why the Committee would discriminate this way against researchers collecting information about activities that might violate criminal or civil law. For one thing, most researchers cannot possibly hope to know all the various civil laws that might apply to the activities they are studying. Furthermore, an examination of cases in the United States suggests that the URERC formula is foolhardy at best. The US evidence shows research on law breaking represents only a small minority of the cases in which confidentiality was threatened by the possibility of court-ordered disclosure.

Although any generalization from the American to the Canadian situation should be done cautiously, we see the US situation as the best available crystal ball if one's interest is in speculating on where protection against the possibility of a subpoena might be most needed. Also, given that some SFU researchers conduct research in the US, it is worth asking how well the limited confidentiality consent statement would fare under US law. Cecil and Wetherington's edited volume on court-ordered disclosure29 is a useful compendium in this regard and describes the following cases:30

1. An organizational researcher did interviews with employees at a utility company. Long after the research was completed, the company was sued for breach of contract, and sought access to the names of interviewees in order to determine who said what about the company in the confidential interviews.
2. A manufacturing organization took part in an international trade survey that was presumably confidential. Legal authorities later sought the organization's responses to the survey in order to determine what sorts of business practices they might have admitted to.
3. A multinational chemical company was sued for the effects allegedly associated with one of its products, and they knew that research produced by an independent lab at a well-known university would be used against them in court. Although the university researchers had nothing to do with the suit, the chemical company subpoenaed their data, and asked for names of subjects, ostensibly to verify the authenticity of the data.
4. A medical researcher created a confidential database that included virtually every case of a particular disease over a 50-year period. Supplementary data were gathered from the persons in the database. Later, some persons who had the disease came to believe that a drug they had taken had caused the disease, and sued the pharmaceutical company that manufactured that drug. The pharmaceutical company subpoenaed the medical researcher for access to the database, including the names of respondents, assuming that information about some of the litigants was contained in the database.
5. A graduate student did a thesis on "the sociology of the American restaurant" that involved participant observation. In the middle of his project, the restaurant burned down under suspicious circumstances. Legal authorities subpoenaed the student's field notes in the hope they contained information that would help them discover the identity of the arsonist.
6. A medical research institute maintained a database that included disease as well as other "personal" information obtained from women who used a particular form of contraception. A manufacturing company was being sued for effects associated with its contraceptive devices, and subpoenaed the researchers in order to glean information about the data base and find out the names of the respondents.
7. A university professor had done research on the propensity of a particular vehicle to roll over when cornering. The manufacturer of the vehicle was sued over exactly that problem, and subpoenaed the researcher's data, including the names of subjects who provided information.
8. A medical researcher gathered information that related to the link between smoking and cancer. A tobacco company subpoenaed his data, including the names of his subjects.
9. A university researcher did field research about the integration of women officers into a police department. A government department subpoenaed her records, including names of subjects, in order to determine what the women police officers were saying about the department.
10. A medical researcher did a survey of 3- to 6-year old children to see whether they were able to recognize "Joe Camel," a "cool" mascot/logo for a cigarette company. It turned out that they could. The company subpoenaed the researcher's data, including the names, phone numbers and addresses of the children, arguing that it wanted to assess whether the researchers had followed appropriately unbiased procedures.
11. An anthropologist was studying Alaskan First Nations communities at the time the Exxon Valdez oil spill occurred. In the middle of battling several multi-billion dollar lawsuits, Exxon subpoenaed the anthropologist's survey data and field notes, in which subjects (some of them litigants) were identified.
12. A sociology graduate student was doing research on animal rights activists, when a destructive raid occurred on the university's laboratory animal facilities. Legal authorities subpoenaed the student in order to determine what information he had regarding their chief suspect.

We know of only two other US cases beyond those cited in the Cecil and Wetherington volume:

13. A sociology graduate student was doing observational field research regarding police interrogation processes. One of the persons being interrogated later sued the police, and subpoenaed the researcher's notes, presumably in order to authenticate his story.31
14. Two organizational researchers conducted interview research with Netscape, a major computing firm. Netscape's major competitor, Microsoft, was later charged with predatory business practices. Part of the charges involved the claim that Microsoft had gradually increased it's market share at the expense of Netscape, the employees of which were involved in the research. Microsoft lawyers subpoenaed the researchers’ interview notes, including names of respondents, in the hope it would assist them is substantiating their argument that Microsoft gained market share not because of predatory business practices, but because of Netscape's poor business decisions.32

These cases reveal that the target of a subpoena is unpredictable; in the US, researchers in any discipline may be subpoenaed. The cases above include research conducted in criminology, sociology, anthropology, women's studies, kinesiology, psychology, medicine, computing science, economics and business administration. The implication is that, if the URERC wishes to warn all prospective subjects who might be affected by a subpoena, the only reasonable way to do so would be to warn every prospective subject in every research project involving humans. We suggest that if SFU-style limited confidentiality were to be imposed on all researchers, research confidentiality would no longer be possible. We sincerely hope that the SFU Research Ethics Policy Revision Task Force will not be responsible for taking such a momentous step.

One of the main objections to unlimited confidentiality is that there are circumstances when a researcher should disclose information about serious prospective harm, or "heinous discovery," as we referred to it above. For example, hearing "threats to someone's life" is a situation in which Dr. Horvath indicated he could imagine limiting confidentiality. But is this a reason to create an a priori limitation of confidentiality? For example, should a researcher limit confidentiality by saying, "I guarantee I will maintain confidentiality unless you tell me that you are going to kill someone, in which case I will feel ethically bound to inform the appropriate authorities"?

From the perspective of maintaining confidentiality using the Wigmore test, this statement would presumably satisfy criterion 1, because it specifies only a singular circumstance in which confidentiality would not be maintained. But consider the irony of limiting confidentiality a priori this way: it would likely result in the death of the intended victim, because the informed consent statement is effectively an admonition that, "If you're going to kill someone tomorrow, don't tell me." Having limited confidentiality this way, it is unlikely that the researcher would discover the subject’s intention to murder the third party. Ironically, then, the specific a priori limitation of confidentiality to account for the reporting of serious prospective harm would produce its own apparently unethical resolution, the death of the victim.

Consequently, we believe there is a key difference between violating a pledge of confidentiality and limiting it from the outset. Limiting confidentiality a priori does nothing to prevent the event, and creates the unethical situation of retaining one's ethical purity by donning blinkers that prevent one from seeing someone else's misfortune. Therefore, we cannot see how limiting confidentiality to avoid heinous discovery would be to anyone’s ethical benefit.

Similarly, as we demonstrate in section 4.i below, the SFU limited confidentiality consent statement is to no one’s ethical benefit either, as it virtually guarantees failure to maintain confidentiality to the extent possible within the law because researchers shoot themselves in the legal foot when they use it.

The URERC's limited confidentiality consent statement violates the Policy Statement’s requirements in at least eleven different respects:33

1. The URERC has argued if there is any possible risk associated with participation in research, no matter how remote the risk, the subject must be informed. By taking "possible" to be synonymous with "conceivable" – which it deigns to be the case whenever a researcher answers "yes" to the question on Form 1 that asks, "Does information to be obtained from subjects include information on activities that are or may be violation of criminal or civil law?" – the Committee has imposed a rote formula for ethics decision-making, something the Policy Statement specifically eschews (see Policy Statement, p. i.9; for further discussion, see History of Limited Confidentiality at SFU, section 4)
2. The Policy Statement specifically "seeks to avoid imposing one disciplinary perspective on others" (Policy Statement, p. i.2), and notes that ethical principles "should be applied in the context of the nature of the research and of the ethical norms and practices of the relevant research discipline" (Policy Statement, p. i.9). By imposing its limited confidentiality consent statement on researchers, the URERC ignores variation in disciplinary norms and standards; for example, it is contrary to the norms/standards of criminology.34 These were the standards observed by Russel Ogden, for example; valuable research such as his would no longer be possible under the URERC's regime of limited confidentiality.
3. By undermining researchers' ability to invoke the Wigmore criteria, the use of a limited confidentiality consent statement thwarts the researcher's and university's ability to protect research confidentiality and the rights and interests of research subjects "to the extent possible within the law," as the Tri-Council asserts we are "honour bound" to do, with the university's support (see Policy Statement, p.3.2). Limiting confidentiality a priori is tantamount to pleading nolo contendre (no contest).
4. By limiting confidentiality, and thereby making the defence of research confidentiality virtually impossible, researchers are transformed into prospective agents of the state. Why would anyone talk to a lawyer if they thought there was a possibility the lawyer could be subpoenaed and would readily become a witness for the Crown? Similarly, why would any prospective research subject discuss his or her involvement in embarrassing, deviant, or criminal activity, if s/he thought that the researcher would supply embarrassing or incriminating information to legal authorities on request?

And once it became clear that researchers were prepared to subjugate ethical obligations to legal ones, surely the number and magnitude of those requests would increase. In this scenario, certain kinds of researchers – and particularly criminologists – would potentially become little more than informants for the state, thereby defying the Policy Statement’s injunction that researchers must "avoid being put in a position of becoming informants for authorities or leaders of organizations" (p.2.4). Furthermore, the Policy Statement says that, "It is … unethical for researchers to engage in covert activities for intelligence, police or military purposes under the guise of university research" (p.1.12). With these injunctions in mind, we suggest that criminologists should steadfastly refuse to limit confidentiality a priori, lest they effectively become little more than agents of the state. If criminologists in particular use the SFU limited confidentiality consent statement, we will undermine our ability to protect confidential research information, invite the authorities to pursue it and, in the process, undermine the integrity of the entire research enterprise.35

5. The imposition of limited confidentiality subordinates all other ethical considerations to the requirements of informed consent, instead of attempting to achieve a sensitive balance incorporating multiple principles. As the Tri-Council states, "For meaningful and effective application, the … ethical principles must operate neither in the abstract, nor in isolation from one another" (see Policy Statement, p. i.9). By paying attention only to informed consent, the policy of limited confidentiality creates a situation where its other promises – e.g., "to protect confidentiality to the full extent permitted by law" – are empty and shallow. This caveat emptor approach to ethics, where we are somehow being ethical as long as we warn prospective respondents that we may act unethically, is contrary to what the Tri-Council Policy Statement would appear to stand for.
6. The Policy Statement says that, "The consent of participants shall not be conditional upon, or include any statement to the effect that, by consenting, subjects waive any legal rights" (p. 2.6). By agreeing to participate in research using the SFU limited confidentiality consent statement, participants effectively waive their right to claim privilege using the Wigmore criteria, which is a legal right.
7. Because prospective research subjects are not informed that they are effectively being asked to waive the right to invoke the Wigmore criteria to claim privilege, subjects are not given the information they need to make informed consent.
8. A policy in which confidentiality is limited fails to live up to the Policy Statement's requirement that "researchers and REBs must strive to understand the views of the potential or actual research subjects" (see Policy Statement, p. i.7). For example, in the realm of criminological research, strong moral injunctions against "ratting" exist among both lawbreakers and the persons whose job it is to catch, prosecute and manage them (e.g., police, prison guards).
9. By imposing limits to confidentiality, the Clayman-Horvath policy in effect at SFU fails to treat subjects with the dignity and respect that the Tri-Council affirms as its "cardinal" ethical principle. Indeed, the idea of gathering data from people, and then doing nothing more than say, "Gee, … I told you I might have to divulge the information to a court" if the researcher is subpoenaed and asked to provide confidential information, is exploitative. As the Tri-Council asserts, "Part of our core moral objection would concern using another human solely as a means toward even legitimate ends" (see Policy Statement, p. i.4).
10. The refusal by an ethics committee to approve research by those who decline to limit confidentiality a priori is a violation of academic freedom. We cannot do our research using a device that exposes our research subjects to harm. The Policy Statement recognizes the importance of academic freedom to the research enterprise:

k) Elsewhere, the Tri-Council Policy Statement notes that:

Using cases from the United States we have problematized the URERC’s formulaic approach to limiting confidentiality, and suggested that, to be consistent with its own logic, the Committee would have had to require all researchers to use the limited confidentiality statement relating to court-ordered disclosure. But we have also suggested that had the URERC done this, it would have effectively ended confidentiality for all SFU researchers, not just criminologists and a few other researchers who collect information from subjects about activities that violate criminal and civil laws.

We are not prepared to stand by and watch the demise of confidentiality, because we believe that it is unethical for researchers to expose research subjects to harm. Also, we believe that without confidentiality much research would become superficial, unreliable, and invalid. Confidentiality is a core ethical principle, not only in the Policy Statement but also in every other code of ethics we have ever seen. As the Policy Statement says:

We accept that it is up to us – university researchers supported by their university administrations – to provide and protect confidentiality using the most effective legal means at our disposal, because without confidentiality, many types of research could not be conducted. In keeping with the Policy Statement admonition that, "though ethical approaches cannot pre-empt the application of law, they may well affect its future development" (p.i.8) we suggest that researchers use a three-pronged approach to resolve potential conflicts between ethics and law:

1. Lobby legislatures for statutory protection of researcher-subject privilege.
2. In lieu of a statutory privilege, campaign the appropriate legal authorities to establish the equivalent of United States "privacy certificates" so researchers can study activities that are subject to mandatory reporting requirements and reduce the risk of court-ordered disclosure.
3. In the mean time, researchers should design research protocols with the Wigmore test in mind. This would involve guaranteeing unlimited confidentiality to all research subjects, and making it clear that the researcher and the institution with which they are associated will vigorously defend research confidentiality and academic freedom against challenges by a court or any other public authority.

We consider each of these strategies in turn.

Our main effort should be lobbying through the appropriate channels that Parliament establish a statutory privilege for university research that has been approved through appropriate university channels (i.e., passes ethics review) and is done in a manner consistent with the ethical principles that govern the researcher's discipline. At the moment, statutory privilege is granted only to Statistics Canada researchers in Section 18 of the Statistics Act. Although it does not yet extend to university researchers, we appreciate the existence of that Act because it shows Parliament's awareness of the integral role that confidentiality plays in the generation of reliable and valid data from which society as a whole benefits.

One could argue that the granting of researcher-subject privilege in statute would amount to formalizing customary practice.36 In a personal opinion solicited by the URERC, Dr. Bernard Dickens37 asserts that, in the absence of legal precedent, Canadian courts will not be sympathetic to researcher-subject privilege. Also he says that the Ogden case does not set a precedent because it was won in Coroner’s Court. Perhaps this lack of cases supports our view rather than weakening it. The fact that there is only one case in Canadian history in which researcher confidentiality has ever been challenged, and that no court in Canada, as far as we know, has ever ordered a Canadian researcher to divulge confidential information, suggests that researcher-subject privilege is already customary practice. Evidence from the US also might assist a campaign to develop statutory protection for accredited Canadian academic researchers.

Two types of US evidence could be persuasive: a) the results of court cases where third parties have pursued confidential research information, particularly the identities of research subjects; and b) instances where the assertion of privilege and deference to it have resulted in authorities abandoning attempts to obtain information.

An examination of US cases indicates that the courts have almost always protected the anonymity of research subjects. When they have not, it is always for the most unusual of reasons in circumstances that should have been avoided (see section 4.5 below). Also, we know of five instances, including one in Canada, where the claim of privilege was sufficient to dissuade third parties from even asserting in court their claim to disclosure:

1. "An economist conducted a study of unemployment patterns among ghetto youth. Some 600 youths were interviewed. A final report was released indicating that one consequence of high unemployment was that many youthful ghetto residents engage in petty crime. A local police commission subpoenaed the researcher and a research assistant and demanded the public disclosure of subject identities. The economist refused to provide this information on the grounds that anonymity had been promised. Pressures ceased shortly thereafter."38
2. "A sexual research institute received a genuine (unstaged) motion picture of a sado-masochistic episode. The supplier of the film inquired whether the institute was interested in securing a copy of the film for research purposes. The director of the institute returned the original and requested a copy. The supplier was subsequently arrested. The director of the sex institute was then subpoenaed to provide all information pertaining to the film. When the director refused, a bench warrant was issued for his arrest. The police never arrested the director, however, and proceedings were later terminated."39
3. "A sociologist conducted a study of police socialization patterns through participant-observation by attending a police academy and by becoming a patrolman. After completing the research and leaving the police department, the sociologist was subpoenaed by a court investigating an alleged act of police brutality witnessed by the sociologist. The subpoena ordered the disclosure of all research files pertaining to the incident and the testimony of the sociologist. The researcher refused on the grounds that a scholarly right existed not to disclose this information. The assertion was not challenged; pressures ceased shortly thereafter."40
4. The FBI threatened to subpoena highly sensitive material on sex research held by the Kinsey Institute at Indiana University. "Since the Institute had pledged confidentiality to many of its research subjects, its officers insisted it could not comply with the FBI demand and would, if necessary, 'accept the consequences of such defiance.' In the end, the FBI backed down, and a legal resolution was not required."41
5. A researcher at the University of Toronto had engaged in long term field research with a major Canadian police department. When a book based on this research was published, it revealed that some police officers had violated suspect rights. The police department approached the researcher for the names of the police officers who were engaged in these illegal activities and threatened legal action if the researcher did not comply. The university hired a well-known lawyer, who wrote to the police department, ordering them to cease and desist. The police withdrew.42

No doubt there are other examples where authorities of one sort or another have pressured researchers for confidential information, and where the resolve of researchers and their universities or research institutes has prevented any court action from occurring in the first place. We suggest these examples are important to the extent they demonstrate that legal and other authorities, when their requests for confidential information are challenged, have deferred to an assertion of academic privilege, and, by doing so, further reified its existence. The problem, of course, is that such examples rarely are published in the literature. We invite any readers of this document who know of such instances to draw them to our attention.

An alternative possibility that has been instituted in the United States has been the "Privacy Certificate." Although we know little about them thus far, they are a part of the Code of Federal Regulations, and apparently began when, in the 1970s, the US Congress "empowered the Secretary of the Department of Health, Education and Welfare to:

More recently – and certainly by 1997, although we do not yet know when exactly these revisions occurred – the provisions of the Code (in section 46.101) had broadened to include,44

Also by 1997, the list of departments/agencies signing on had come to include the Department of Justice, and the National Institute of Justice. Research eligible for privacy certification need not be funded by any of these agencies; any research that has gone through appropriate institutional ethics review (e.g., in a university) may be certified. Further, the statute creating privacy certificates "has been judicially reviewed and held to be constitutional."46

Information obtained from the Johns Hopkins University information web-site indicates that

The receipt of certification allows one to confidently guarantee that:

Although the few things we have read about privacy certificates thus far appear that they may well allay any fears that researchers and university administrators might have about offering unlimited guarantees of confidentiality, our preliminary reaction is to view them with caution. We have no data, for example, on what an application actually involves, in what proportion and according to what criteria applications are approved, and whether there are any other strings attached that the researcher signs on to in the course of obtaining such a certificate. To the extent that the approval process is anything other than pro forma subsequent to receiving ethics approval from one's university, our greatest worry is that it puts academic freedom in the hands of an external bureaucracy that could potentially be in a conflict of interest in determining which applications will be approved. The Johns Hopkins information web site, for example, notes that, "the investigator must request a certificate of confidentiality from the appropriate official." But does this mean that a researcher who wishes to engage in research that may be critical of the policies of the US Department of Justice will be trying to gain a privacy certificate from the US Department of Justice?

In Canada, we know of no such equivalent, and assume that researchers who would like to acquire the equivalent of a privacy certificate would need to approach their Provincial Ministry of the Attorney General, since the administration of justice is a Provincial responsibility. At the same time, there would be limits to what such protections could entail. Although the Attorney General could presumably guarantee that no agent of the Crown (e.g., Crown attorneys, police) would be able to initiate any action to acquire confidential information, the US provisions extend much further, in areas that are not within the direct control of the equivalent of the Provincial Attorney General (e.g., to municipal and administrative departments). In any event, we see the use of privacy certificates as a secondary solution to a problem that is better and more directly solved by the recognition of researcher-subject privilege in statute, although they may well offer a useful solution until that day comes, if indeed it ever does.

In lieu of a statutory privilege or the development of the equivalent of US privacy certificates, the only legal mechanism currently available to protect confidential research information is the Wigmore test. Unfortunately, the Policy Statement does not mention the significance of the test for establishing privilege on a case-by-case basis. We regard this as an extremely unfortunate oversight.

To set the stage for our discussion of the test’s four criteria, and how researchers might apply them, we return to the situation at SFU. Recently the University administration made clear its commitment to assert academic freedom if and when a court or anyone else tries to obtain confidential research information. Doing this will mean that researchers need to design their research protocols with the Wigmore criteria in mind.

The vigorous defence of confidentiality by the university research community safeguards many interests. Most directly, it protects the rights and interests of research subjects. Those individuals volunteer to participate in research, usually at our invitation; they typically receive nothing but a thank you or token reward for having done so; and because of the aura of integrity that still surrounds the University, they trust that their rights and interests will be safeguarded. Just as the client, not the lawyer, holds lawyer-client privilege, we would argue that researcher-subject privilege is the subject’s, not the researcher’s privilege. However, because the subject cannot assert the privilege, the onus is on the researcher to assert it. Asking prospective subjects to give up any legal protections we can muster in return for their participation – which the SFU limited confidentiality statement effectively requires subjects to do – is directly contrary to the Policy Statement, which asserts:

Our ethical obligation is to safeguard those rights.

By vigorously safeguarding subject rights, the research enterprise itself is enriched because the safeguarding of research confidentiality maximizes academic freedom, which the Policy Statement (p.i.8) views as fundamental to the pursuit of knowledge. The confluence of confidentiality and academic freedom was recognized by Drs. Nicholas Blomley and Steven Davis, whose independent Russel Ogden Decision Review49 was accepted in its entirety by SFU President Jack Blaney, and its recommendations implemented (see footnote ). Blomley and Davis recognized that Russel Ogden’s use of the Wigmore criteria not only protected his research subjects from harm but also asserted academic freedom against court-ordered disclosure. Further, they recognized that the university is responsible for protecting both academic freedom and research subjects:

More generally, they state:

Indeed, if universities do not defend academic freedom, who will?

Although Simon Fraser University's behaviour in the Russel Ogden case will remain a stain on the university's record, we hope President Blaney's recent acceptance of the Blomley-Davis report and its recommendations signals the university’s willingness to assume national leadership in the protection of academic freedom. The university's acceptance of Blomley and Davis' third recommendation – that graduate students be guaranteed legal support when their academic freedom is challenged (e.g., in the form of a subpoena for confidential information) – was accompanied by the following statement:50

The wording of this guarantee parallels a clause in the SFU Framework Agreement to the effect that the university is obliged to support faculty members in court in situations where, acting in good faith and in the execution of their duties, they run into a legal challenge.

Although we are pleased the university adopted this position without the Tri-Council requiring it to do so, it is also clear that any repetitions of the Russel Ogden debacle would have been unethical when measured against the Tri-Council's Policy Statement:

If researchers are to live up to these requirements, they are obliged, at minimum, to assert researcher-subject privilege and defend research confidentiality until there are no courts left in which to assert it, and universities are ethically obliged to support them. The Policy Statement is particularly important because it recognizes that the minimum ethical requirement of researchers and the university is to protect confidentiality, and thereby academic freedom, "to the extent possible within the law." This can only mean fighting court-ordered disclosure of confidential research information all the way to the Supreme Court of Canada, if necessary.

It thus behoves persons responsible for designing and maintaining ethical standards to understand the legal means that are available to academics to protect confidential information, and require that researchers design their research protocols to maximize the likelihood of success.

So what is the law of the land with respect to protecting confidential research information? As there is no statutory protection for researchers other than those at Statistics Canada, we have to turn to the common law and, in particular, the Wigmore criteria, the same test Russel Ogden used when he successfully asserted researcher-subject privilege in Vancouver Coroner's Court. Not only has the Supreme Court of Canada recognized the Wigmore criteria as the appropriate test for examining any case-by-case claim to privilege (R. v. Gruenke, 67 C.C.C. (3d), 1991), but also it has asserted that the door is still open to establish class privileges through common law. For example, Sopinka, Lederman and Bryant (1992) note that:

Later, they quote Lamer C.J.C., writing for the Supreme Court in R. v. Gruenke, who expresses a similar view regarding the Wigmore criteria and the establishment of privilege:

These statements by the Supreme Court, in conjunction with Russel Ogden’s success at invoking the Wigmore criteria in Coroner’s Court, lead us to reaffirm that Canadian researchers would be remiss if they did not design their research protocols and consent statements with the Wigmore criteria in mind. Research Ethics Boards will be negligent if they do not require researchers to take this approach.

Talking about the US experience with court-ordered disclosure, Michael Traynor offers very much the same advice:

Although we appreciate the manner in which the Policy Statement unambiguously expresses the value it attaches to confidentiality, we believe it is remiss in not explicitly drawing Canadian researchers' attention to the Wigmore test. It is only by designing their research in a manner that leaves them prepared to best assert and defend researcher-subject privilege that researchers can meet the Tri-Council's minimal requirement that they protect confidentiality "to the extent possible within the law" (Policy Statement, p.3.2). Future revisions of the Policy Statement should incorporate this advice.

In the next two sections, we (a) present an analysis of the Wigmore criteria and their applicability to the defence of researcher-subject privilege; and (b) consider that strategy in light of the Tri-Council Policy Statement.

In his defence, Russel Ogden successfully argued his communications passed the Wigmore test. This test summarizes the criteria traditionally used in common law to establish, "a privilege against disclosure of communications between persons standing in a given relation,"52 and thereby recognize an exemption from the obligation all citizens normally have to testify. Indeed, Ogden’s case is the most explicit application of the Wigmore criteria to the research enterprise we have found. The criteria require that:

It is important to recognize that the Wigmore criteria54 are always considered in the context of a "given relation[ship]", e.g., lawyer-client, police-informant, priest-penitent, therapist-patient, or, in this case, researcher-subject. The basis on which privilege is recognized lies less in the content of the communication itself than in the need to recognize the privilege in order to maintain the integrity of a socially valued relation. Wigmore approvingly cites an 1851 decision by Wigram to the effect that, "The rule does not rest simply upon the confidence reposed; … it seems to rest, not upon the confidence itself, but upon the necessity of carrying it out."55

Canadian legal commentary on the Wigmore test56 indicates there needs to be clear evidence on each criterion in order for communications in a given relation to be recognized as privileged. In this regard, it is instructive to consider each of the four Wigmore criteria in the context of the researcher-subject relationship, and in light of the Tri-Council Policy Statement.

In his original treatise, Wigmore (1905) discussed the notion of privileged communication in the context of many different kinds of relations. The first criterion asserts that the subjects in the relation must be clear that the communication in question is a "confidential" one. From Wigmore’s perspective, this would seem to require two things. First, the privilege can exist only to the extent that the communication in question is integral to the relationship. For example, when a person discusses a prospective legal defence with a lawyer, the discussion is within the confines of the lawyer-client relationship and hence can be privileged. But when the same persons informally discuss hobbies of common interest, the discussion is not a discussion between a "client" and a "lawyer" – one could have such a conversation with anyone – and hence is not privileged on the basis of a lawyer-client relation.

Second, both people in the communication must share the understanding that the communication in question was uttered in confidence since, as Wigmore notes, "The moment confidence ceases, privilege ceases."57 According to Wigmore, this did not necessarily require an explicit assertion of confidentiality with every new utterance, and, in some ways, the redundancy of the reminder is a measure of the extent to which confidentiality is understood to be an integral part of the relation. Wigmore’s formulation allowed for the existence of un-stated and yet common understandings of confidentiality.

The Supreme Court’s reasoning in Regina v. Gruenke reveals vital information about how the Wigmore criteria might relate to researcher-subject interactions, and the monumental strategic legal blunder contained in the SFU limited confidentiality consent statement. In the Gruenke case, the Crown sought to introduce as evidence incriminating statements an accused murderer had made to a counsellor and her pastor. The Supreme Court concluded the statements were not privileged because they did not satisfy the first criterion of the Wigmore test, i.e., the communications did not originate in a confidence they would not be disclosed: "The pastor and the counsellor were unclear as to any expectation of confidentiality," and "there was no evidence that the accused made her admissions to them in the confident belief that they would be disclosed to no one else." (Gruenke, p.307)

The implication of Gruenke for researchers is that they should begin their research with an explicit statement that asserts the research interaction is in confidence, and should probably record the assertion, or ensure it is part of one's written statement for securing informed consent, in the event the matter ends up in court. In her article discussing the Gruenke ruling, although she does not explicitly consider the research context, lawyer Mary Marshall (1992) advises:58 "You should begin your discussion with the statement ‘This must remain absolutely confidential.’" Traynor (1996), a US lawyer, makes much the same point: if there is any danger of confusion, be explicit that the communication is confidential. Without the expectation of confidentiality, the raison d’être of the Wigmore privilege is missing.

It is for this reason that the SFU limited confidentiality statement virtually assures defeat, just as it did in Atlantic Sugar, Ltd. v. The United States of America (see section 4.5.i, below). By telling subjects that a court might require the information to be divulged, the statement creates the raison d'être for the court to make such an order. With the limited confidentiality statement in hand, a judge would be able to say, "You warned the participant that a court might order you to divulge the information, and they nevertheless provided the information. This court is now ordering that you hand over the information, because the communication did not arise in the expectation that it would be withheld in these circumstances."

It is with respect to criterion 1 that the Tri-Council Policy Statement is the weakest and most ambiguous, a situation we believe must be rectified in future revisions. Notwithstanding all the positive and directive statements included in the Policy Statement, there is one section of the document which says that, "researchers should indicate to research subjects the extent of confidentiality that can be promised, and hence should be aware of the relevant law." (p. 3.2) The sentence is noteworthy insofar as it is the one sentence in the document that SFU's VP-Research could seize upon, and did, to make his assertion that the Policy Statement was consistent with the limited confidentiality consent statement that he was responsible for introducing at SFU. Accordingly, the sentence deserves close scrutiny.

For three reasons we believe the VP-Research’s interpretation is misleading. To begin with, if the Tri-Council intended that confidentiality should be limited beyond mandatory reporting requirements, then in view of the legal requirements imposed by the Supreme Court in Gruenke (see section 3.1.iii and 4.4.i, below), there effectively can be no meaningful "confidentiality." If that is the case, all the other sections of the Policy Statement expressing the Tri-Council's reverence for this ethical principle are hollow. We presume the Tri-Council did not intend its ethical assertions as a sham, and hence that it did not intend the many pages it devoted to articulating the importance of confidentiality to research to be undermined by the single sentence that appears on p.3.2.

Our second reason for disagreeing with Dr. Clayman's assertion that the Policy Statement supports use of the SFU limited confidentiality consent statement is that the law of the land does not limit confidentiality in any way, save for a very limited number of mandatory reporting laws. In fact, quite the opposite is the case. The one case we have in this country where researcher-subject privilege was challenged involved Russel Ogden’s research, and the Coroner agreed that Ogden had done the proper thing in guaranteeing "absolute confidentiality" to his research subjects. Judge Daniel Steinberg of the Provincial Court of British Columbia, in his obiter dictum in Russel Ogden v. Simon Fraser University, made much the same point:

We see Judge Steinberg's statement as recognition that, without such guarantees, internationally acclaimed and socially valuable research such as Ogden's could never be done. Given all that the Policy Statement says about the many social benefits that accrue from the existence of such research (e.g., see p.i.4 of the Policy Statement, and section 4.4.ii-iv, below), and the integral importance of research confidentiality to the conduct of that research (e.g., see p.3.1 of the Policy Statement and the section 4.4.ii, below regarding Wigmore criterion 2), we simply cannot believe that the Tri-Council would intend, by the sentence on p.3.2, to destroy the possibility of such research being done again.

Finally, with respect to the possibility of court-ordered disclosure, our awareness of "the relevant law" is that we must make an unequivocal guarantee of confidentiality in order to fully comply with the Tri-Council's admonition to protect the rights of research subjects to confidentiality to "the extent possible within the law." The Policy Statement unambiguously states that:

How else can we live up to that injunction other than through the Wigmore criteria? If we extrapolate from the Supreme Court of Canada’s reasoning in R. v. Gruenke, to pass the Wigmore test, we must have a clear understanding, which the research subject must share, that the communications between us are being uttered in complete confidence. Given that the Policy Statement requires researchers to be aware of all the relevant law and protect the rights and interests of research subjects to confidentiality, then again we say, surely researchers should design their research protocols with the Wigmore test in mind. An unequivocal guarantee of confidentiality is the only solution that meets all the principles of the Policy Statement and is consistent with the law. The SFU practice of limiting confidentiality a priori via the limited confidentiality consent statement prevents researchers from maintaining confidentiality to the extent possible within the law. According to the criteria laid out in the Policy Statement, then, the SFU limited confidentiality consent statement is unethical.

Of course, beginning with a guarantee of unlimited confidentiality is not in and of itself sufficient to establish that a communication should be considered privileged. The communication has to pass the three other criteria as well. The second criterion asserts there must be evidence that confidentiality is so important to the continued existence of the relationship that it would cease to exist, or be eviscerated, without it. Indeed, this prerequisite for the granting of privilege draws our attention to the fact it is the relationship even more than a given communication per se that is being protected by the recognition of privilege. As Wigmore (1905) asserts, the object of granting privilege is "to protect the perfect working of a special relation, wherever confidence is a necessary feature of that perfect working" (p.3211).

Wigmore takes the lawyer-client relation to be an example of a relationship that would not exist in the absence of privilege. Why or how could anyone talk freely to a lawyer for the purpose of creating a legal defence if one's lawyer could then be subpoenaed to testify for the prosecution? He contrasts this with the physician-patient relationship, which he believes does not meet this criterion, since he doubts whether patients would stop going to physicians, or stop telling physicians about their symptoms or conditions, in the absence of a recognized privilege. Even where these might be considered "sensitive" – Wigmore gives the example of matters in relation to abortion and sexually transmitted diseases – he asserts,

But where does the researcher-subject relationship sit with respect to the second criterion? Is the provision of confidentiality an essential feature of the researcher-subject relationship? In the case of many kinds of research we believe it is, and that it is embedded in the relationship itself. Perhaps the clearest example of the essential nature of privilege to the researcher-subject relationship is research on law breaking, especially when it concerns undetected or unreported law violations. What offender would talk openly and freely about undetected offences for the purpose of contributing to research if they knew that a researcher could be forced by a court to testify for the prosecution? Academic privilege (cf. Judge Steinberg in Ogden v. SFU), or what we would prefer to call "research subject privilege," is indispensable to the relationship.

Several features of the researcher-subject relationship suggest that confidentiality is essential to it, including the motives of the subject (they are voluntary participants with little to gain) and the access confidentiality gives to unique information

We note that in most of the relationships that have or aspire to some degree of privilege – for example, police officer-informant, therapist-client, priest-penitent – the person providing the information has something to gain by the transmission of that information. The informant seeks immunity from prosecution; the therapists' client hopes for healing; the penitent seeks forgiveness.

In the social sciences, however, researchers have little to offer beyond a chance for someone's voice to be heard, or the opportunity to contribute to the generation of knowledge and the understanding that comes with it. Indeed, when we do have something more concrete to offer – money, or a lesser sentence, or a potential cure – ethical alarms are set off, and we are obliged to consider whether the incentive is "undue," or even tacitly coercive. Every code of ethics we know includes the requirement that participation must be completely free of direct or indirect coercion.

Also, unlike the patient, the penitent and the lawyer’s client, all of whom initiate their interaction with the person to whom their secrets are entrusted, the research subject usually does not initiate the interaction with a researcher. The researcher seeks them out, and asks for their assistance. Given that: (a) we usually approach them for information; (b) their participation is entirely voluntary, and (c) we (and society) are the primary beneficiaries of their participation, our ethical burden to protect research subjects is enhanced, and the protections to which they are entitled are that much greater than if they initiated the relationship for their direct gain.

But does the provision of confidentiality really make a difference to research with human subjects? Authors of textbooks in research methodology and research ethics would seem to think so. In Ethics in Social Research, for example, Bower and de Gasparis (1978) state,

Indeed, the assumption is so pervasive as to be thought self-evident. Also, an examination of US cases involving subpoenaed research information shows that, time and again the courts have accepted that confidentiality is essential to the research enterprise. In fact, with few exceptions (which we describe below), the courts have shown enormous respect for research confidentiality, and severed identifiers from the records. The same is true of the privacy certificates described in section 4.2 above: if confidentiality is not essential to ethical and valid social research there would be no reason for privacy certificates to exist!

Generally speaking, the methodological belief is that the more clearly anonymous or confidential the data, the greater their probable validity, particularly when the topic under discussion is a sensitive one and/or where there can be negative repercussions for the subject (e.g., consequences at work, in his/her social group, possible incarceration).61

Our research experience with persons involved in repeated law violation, most of which remains undetected, as well as with persons involved in catching and dealing with them, leaves us with no doubt that confidentiality is essential to the researcher-subject relationship. Because of the premium put on confidentiality by our research contacts, we believe that if research of this type were to be conducted without an assurance of unlimited confidentiality, it would not yield valid and reliable information.

The Policy Statement offers unequivocal support for the idea that confidentiality is essential to the researcher-subject relationship. Most directly, in a section entitled "Privacy and Confidentiality" (Section 3), it asserts:

The Tri-Council Policy Statement, which will soon govern the conduct of all university researchers in Canada, could thus be referred to in any defence of research confidentiality made in court.

The third criterion asks whether the community believes the researcher-subject relationship is one that should be sedulously fostered. Having determined that the ability to offer confidentiality is integral to the "perfect working" of the researcher-subject relationship, we next consider whether there is reason to believe that the community values the conduct of social research. The way we address this criterion depends in part on which "community" is considered to be the referent. We consider two: the research community itself, and the broader community.

With respect to the former, we suggest it is a point nolo contendre that the research community believes its relation with subjects is worth both fostering and safeguarding. The existence of codes of ethics are themselves evidence for the degree to which the academic community believes the researcher-subject relation is valuable and should be fostered. We know of no code of ethics that does not recognize the provision of confidentiality as integral to that relationship. Added to this are dozens of research texts that assert that human subjects are the natural resource on which the human research enterprise thrives, and that subjects deserve our utmost care and protection.

Of course, the nature of the relationship between researcher and subject also varies to some extent by research tradition. The psychologist and experimental research subject, for example, typically have a relationship that is transient, limited in scope, and where the circumstances of the interaction are highly programmed and constrained. The participation may regularly last only 50 minutes, for example, and one’s role as subject may involve being exposed to certain stimuli and/or simply responding to some structured questionnaire items about the experience.

Other traditions involve quite different relations. Reinharz, for example, discusses a range of feminist perspectives on this issue, and suggests that being a feminist, any kind of feminist, means doing one’s all to create an egalitarian interaction between researcher and subject. Feminists deliberately set out to avoid the hierarchical and exploitative relations that they see as characterizing much "malestream" research.62 Accordingly, the ideal researcher-subject relation is one in which the researcher does not impose her power to unilaterally define the research question, constrain responses, and simply wave goodbye at the end of the session. Instead, an egalitarian exchange is sought in which the subject is as much collaborator as informant, and where the researcher does her best to ensure the subject receives at least as much for her participation as the researcher gains from having done the research. Various kinds of participatory-action research follow the same mandate.

Similarly, in anthropology and sociology, with their rich traditions of ethnography, participant observation, and other forms of field research, much emphasis is placed on the extended duration of interaction that is required to appropriately build rapport and mutual trust. In the literature there are frequent admonitions to the effect that research conducted without this rapport is less valid, and hence less valuable, than that which involves it.63 Also, field researchers tend to emphasize a more inductive approach, where research questions and concepts are expected to emerge from ongoing interaction, and hence to better reflect the "realities" of the situation, than would be the case if the researcher merely imposed his or her understandings from the start.

These latter relations – characteristic of phenomenology, ethnography, standpoint feminism, research with "key informants," participant observation, participatory action research, and other qualitative research traditions – place the researcher in a different relation vis-à-vis the subject than is the case with more experimental and/or quantitative paradigms. Confidentiality is valued because it allows one to understand people in a manner that is not threatening to them, and is the basic expression of trust that allows access into people's lives. John Lofland, when he was Chair of the American Sociological Association's Committee on Professional Ethics, stated it this way:

With respect to the appropriate stance to ensure the protection of confidential information derived in fieldwork, criminologist Marvin Wolfgang explained:

With respect to the broader community, there is abundant evidence again of the value that is placed on social science research, and the researcher-subject relationship on which that research is based. Russel Ogden illustrated this well in so far as the media and various governmental Committees and Commissions avidly sought his views, which were developed in the context of the researcher-subject relationship. Indeed, the decision by the Coroner that when Ogden refused to violate confidentiality, he was not in contempt of court involved a similar recognition of the importance of the research enterprise, and the role of confidentiality within it.

The hundreds of millions of dollars that granting and contracting agencies spend annually on research is another indicator of the importance attached to the research enterprise, and the researcher-subject relationship on which an important component of it is based. Further, if we believe that formulating policies, procedures, treatments, and law is better to the extent it is based on "facts" rather than stereotypes and empirical knowledge rather than uninformed opinion, any damage to the researcher-subject relationship short-changes our long-term ability to deal with these issues in an informed way. Stated simply, our role involves critically questioning and analyzing all aspects of society, including asking questions about why social arrangements – e.g., law – are the way they are; and looking at the borders between pathological and normal, criminal and non-criminal, and stigmatized versus mainstream. It is only by preserving and fostering the researcher-subject relationship that we can hope to provide the understanding and knowledge that society demands for its own long-term benefit. How and why would anyone ever talk to us about their private lives and view of the world if they thought we were going to simply turn around and hand the information over to the state? As criminologists, our studies often focus on undetected criminal activity. Why would anyone divulge information about law violations if they thought we would give evidence to a court for their prosecution?

The value attached to research confidentiality is also evident in the one place we know of in Canada where researcher-subject privilege has been codified in law, and that is in the Statistics Act. This Act gives Statistics Canada employees such a privilege to ensure that the information unearthed by research by one portion of government (Statistics Canada) cannot be used by any other branch of government, or in any other litigation, in a manner that would violate the confidence of any individual respondent.

Section 17(b) of the Statistics Act, entitled, "Prohibition against divulging information," asserts that:

Section 18 is even more assertive:

There appear to be three main reasons for this provision of privilege, i.e. the Federal Government's recognition that:

1. information to be gathered through social research is essential to informing government so that it is best able to fulfil its mandate;
2. confidentiality is essential to the gathering of accurate information; and
3. because Statistics Canada is itself a part of government (i.e., the state), research subjects must be assured that any information they provide will not later be used against them by other parts of government.

A fourth reason should also be noted, which is related to the fact that, when the first Statistics Act was proclaimed in 1917, the primary mandate of Statistics Canada (then the Dominion Bureau of Statistics) was to undertake the census. Because of the mandatory participation requirement – Canadians are liable for fine and/or imprisonment if they do not participate in the census – Parliament traded mandatory reporting for guaranteed confidentiality. Although this root of the privilege embedded in the Statistics Act is important, we note that Statistics Canada now engages in many other types of surveys that do not involve mandatory participation requirements, but are still governed by the same oath and privilege. Thus, at this time, the mandatory participation requirement itself cannot explain the existence of the privilege in these other endeavours.

Finally, we emphasize the fact that, as far as we know, no researcher in Canada has ever been ordered by a court to divulge confidential research information. As we outlined above, this may be taken to indicate a traditional acceptance of researcher-subject privilege in Canada. If a general presumption and recognition of privilege did not exist, one might expect many cases to be on the books, as is the case of claims that have been made for journalistic privilege, client-therapist privilege, physician-patient privilege, and so on.

There is a preponderance of evidence suggesting that confidentiality is essential to the "perfect working" of the researcher-subject relationship. Confidentiality may not be essential to every type of research with human subjects – e.g., social surveys regarding innocuous topics. However, when it comes to sensitive research involving the collection of information that could cause subjects anything from embarrassment to serious harm, methodologists and field researchers by the score maintain that it could not be done without an unqualified guarantee of confidentiality.

Ironically, the major threat to the ability of researchers to protect research confidentiality using the Wigmore criteria thus comes from the SFU limited confidentiality consent statement, since it can be taken by the courts as a statement that all research can proceed without an unqualified guarantee of confidentiality. The implication of this is that the most counterproductive thing the Task Force can do when creating a new Ethics Policy for SFU would be to repeat the mistake made by the SFU Ethics Committee since 1995. Making anything less than an unambiguously clear statement of the importance of research confidentiality and the need of researchers and the university to defend it vigorously when challenged in court will undermine, if not destroy, our ability to protect confidential research information to the extent permitted by current Canadian law.

There are several ways in which the Policy Statement is consistent with the view that the researcher-subject relationship is socially valuable and should be sedulously fostered. Indeed, the need for researchers to sedulously safeguard the researcher-subject relationship is evident in the Policy Statement's admonition to consider the rights and interests of research participants at every turn; it is the one principle that pervades the entire document. Indeed, in the event that conflicts of interest are created or exist, we suggest that the Task Force has an even greater obligation to research participants than it does to the Policy Statement itself.

If there is one sentence that can capture the general flavour of the Policy Statement, it would be that the Tri-Council is adamant that research ethics must be considered with the rights and perspectives of research subjects in mind. The point would seem so central to what makes "ethics" ethical as to be a truism.

And yet we have, on several occasions, argued that a central problem with a "limited confidentiality" approach to ethics is that it seemed more concerned with university and researcher liability than with protection of the research subject. The aspect that most concerns us about the URERC's caveat emptor approach is its implicit view of the research subject as a supplier of information to whom we have no more than a contractual obligation: an objectification of the research subject as the supplier of information to be harvested. The practice of "limited confidentiality" has institutionalized the exploitative view that, as long as researchers warn subjects that confidentiality is limited, they are free to provide the information without contesting a court order. This is liability talking, not ethics. In the words of a correspondent for the SFU Ethics Committee in one case we discovered all is fine as long as researchers and the university are "covered."67

We, and now apparently the Tri-Council, view such behaviour as ethically unacceptable. When researchers, who should and do know far more about the possible legal repercussions of research than subjects, take whatever data they are offered, and offer limited protection in return, they are engaged in exploitation.68

The Policy Statement addresses that approach to confidentiality in several respects. First, it states clearly that the research subject is to be treated as more than simply a means to an end:

The Tri-Council went further to express not only the debt we owe to the subjects who volunteer for our research, but also the special responsibility researchers incur to protect them. Consistent with recognition of the importance of the researcher-subject relationship to the research enterprise is the Policy Statement’s ethical requirement that researchers and Research Ethics Boards consider ethics from a subject-oriented perspective. For example:

It is clear throughout the Policy Statement that a researcher's first obligations are to research subjects. Section 4 on conflict of interest, for example, cautions researchers about conflicts of interest. Such conflicts include the economic interest that occurs when a consultant receives money from an organization or agency, but to fulfil the contract, gathers information from persons who are lower in the organizational hierarchy, or who are "clients" of the agency paying the bills. Researchers are required to either make such conflicts explicit to subjects, or avoid them altogether (see pp.4.1-4.2). These issues are also discussed in a section on informed consent (Section 2), where the unique dynamic of situations involving "captive" or otherwise obliged populations (e.g., prisoners, employees of the hiring agency) are considered:

We highlight these conflict of interest provisions because the imposition of limited confidentiality involves the realization of such a conflict. At SFU, university authorities interested more in preventing liability than protecting the rights of research subjects chose a strategy – the limited confidentiality consent statement – that would "cover" the University.69 By limiting confidentiality this way, the URERC effectively requires subjects to waive their right to assert researcher-subject privilege in court. We contend that it is the ethical responsibility of the researcher and the university to be the guardians of the privilege, not to cause subjects to waive their rights as a precondition of participation. Hopefully the Policy Statement requirement that, "The consent of participants shall not be conditional upon, or include any statement to the effect that, by consenting, subjects waive any legal rights" (p.2.6) will put an end to use of the SFU limited confidentiality consent statement.

The first three criteria could be considered as "eligibility" criteria for a particular relation, and every indication is that the researcher-subject relation passes those three tests with flying colours. We point here to the precedent established by Ogden’s defence of researcher-subject privilege in Coroner’s Court (see Daisley, 1994)70 and to an obiter dictum by the Honourable Judge D. Steinberg of the British Columbia Provincial Court in a small claims action brought by Russel Ogden against Simon Fraser University. Judge Steinberg noted that "principles of academic freedom and privilege … are fundamental to the operation of any accredited University," and that "There is no question that the research of [Ogden’s] was of great social value and very topical in today's society" (see Steinberg, 1998, p. 20).71

Notwithstanding Judge Steinberg’s apparent recognition of what, for lack of better terminology, might be considered a general presumption of privilege, the essence of the Wigmore criteria lies in criterion four, which calls for a balancing of two competing social considerations:

1. the importance attached to the relation for which privilege is sought, and the adverse impact that would fall upon that relationship if confidentiality were to be thwarted; versus
2. the importance attached to the state’s obligation to ensure fair trials, and of all citizens to provide evidence when called, and the impact that the non-availability of evidence due to recognition of privilege would have on the particular trial in which the privilege is claimed.

In order to evaluate the social worth of research relative to specific litigation, we examine decisions made in cases where this balance has been assessed by the courts.

In Canada, there is only one case to consider, and that is the case in which Russel Ogden was subpoenaed by the Vancouver Coroner to testify in the case of the "unknown female." For the Coroner in Ogden’s case, the trade-off involved a comparison of the value attached to:

1. The ability of Ogden and other researchers to do research and contribute to society’s knowledge in controversial areas such as assisted suicide, and the knowledge that any denial of privilege would have an adverse effect on the ability of such research ever to be done again; versus
2. His need, as Coroner, to identify those who have died from "unnatural causes," and to unambiguously determine cause of death, as well as for the criminal justice system to bring to justice persons who might conceivably be charged with murder under the criminal code.

In the end, the Coroner decided for Ogden, and granted the affirmation of privilege he sought. He exonerated Ogden of any "stain of contempt." Ogden's case, described in The Lawyer's Weekly as "a precedent setting ruling on privilege and confidentiality", and "an excellent example of how the Wigmore test can be used to protect confidentiality" (Daisley, 1994, p.28), was a victory for all researchers and research subjects.

There are several more examples in the United States. We include summaries from twelve such cases, i.e. all the relevant cases referred to in the 1996 special issue of Law and Contemporary Problems devoted to the topic of court-ordered disclosure in Appendix A (see also section 3.1.iv, above). Although Canadian courts do not necessarily follow the course charted by their US counterparts,72 we are heartened by the US outcomes, where various trial and appeal courts have consistently respected the needs of researchers, and the importance of confidentiality to the research enterprise. The few losses are accounted for by the extraordinary procedural and design problems characterizing these cases (for description, see section 4.5, below).73

In the US cases, the value attached to research confidentiality is perhaps nowhere better expressed than in the most recent such case (not included in Appendix A) which pitted the richest corporation in the world – Microsoft – against two researchers from two of the most prestigious universities in the world (Harvard and MIT).

The US Department of Justice is currently prosecuting Microsoft Corporation for allegedly violating anti-trust laws by engaging in unfair competitive practices. One area of Microsoft's practices under scrutiny is their effort to become dominant in the Internet browser market. Netscape got the head start with Netscape Navigator, and, at one time, held 80% of the market, compared to Microsoft's Internet Explorer, at 20%. More recent figures show a 50-50 split. The prosecution is arguing that Microsoft rose from 20% to 50% because of its predatory business practices; Microsoft argues it improved market share because of better business practice than Netscape, which made some significant blunders allowing Microsoft to increase its share of the market.

In the midst of this case Microsoft lawyers heard of a forthcoming book by Drs. Michael Cusumano (of MIT) and David Yoffie (of Harvard), entitled Competing on Internet Time: Lessons from Netscape and the Battle with Microsoft. As part of their research, Cusumano and Yoffie had interviewed 40 current and former employees of Netscape. Microsoft secured a prepublication draft of the book, and were encouraged by what they saw. They wanted more, and subpoenaed Cusumano and Yoffie's tapes, notes, and transcripts, to get it.

We note first that MIT and Harvard did not abandon their researchers – they appeared in court with their researchers to challenge the subpoena in the interests of preserving research confidentiality and academic freedom.

Most recently, the case was heard by the US Court of Appeals (First District), which, in a December, 1998 decision,74 noted that there is a huge amount at stake in this case – literally, hundreds of millions, if not billions, of dollars. And, as the court explained, there is much on the Microsoft side of the balance: Microsoft's need for Cusumano and Yoffie's research information was recognized as "substantial" in terms of its utility to their defense in the anti-trust case. The court notes also that Microsoft was specific in their request and not simply engaged in a "fishing expedition." And, as the court writes:

It is also instructive to note that the legal position Microsoft's eight lawyers argued in trying to secure the information was to say that the communications between Cusumano and Yoffie and their research participants were not "really" confidential, and hence could not be considered privileged, i.e., they did not arise in a confidence that they would not be disclosed. No doubt they would have had an easy victory had Cusumano and Yoffie used the SFU Ethics Committee’s limited confidentiality consent statement. The point seems so obvious as to make us wonder yet again how the URERC can continue using it.

Notwithstanding all that is at stake for Microsoft, and the utility that access to the researchers' transcripts, notes, and tapes would have for them, the US Court of Appeals (First District) sided with Drs. Cusumano and Yoffie and their research subjects. As the court writes:

And later in the decision:

We only wish that Simon Fraser University's VP-Research and the University Research Ethics Review Committee valued research as much.

We have described above how the court's consideration of the fourth Wigmore criterion involves a balancing of the value to the research enterprise and the researcher-participant relation of the preservation of research confidentiality, versus what is necessarily an unknown at the time we embark upon our research – the specific case facts against which the value of the research enterprise will be weighed, and for which a subpoena is issued. The relevance of the Tri-Council to that consideration is given by comments in the Policy Statement that reflect on the value of "our" side of the balance, i.e., the value that the federal government and the granting councils attach to the research enterprise. They have done so well, and in a manner that will be helpful to any researcher who is faced with defending research confidentiality in court.

First, it is noteworthy that The Tri-Council Policy Statement offers an unequivocal affirmation of the value of free and unfettered research to the long-term benefit of society. For example:

The Policy Statement continues:

As this last passage suggests, the Tri-Council definition of "research" is a broad one. So, too, is its appreciation for the role that academic freedom plays in the research quest:

In sum, research is a formidable social good, warrants weighty consideration in comparison to other social goods, and deserves protection from those who would undermine the research enterprise and the academic freedom required for it to blossom and bear fruit.

Notwithstanding all the positive evidence above regarding the ability of the academic research community to vigorously defend research confidentiality through the case-by-case privilege afforded by the Wigmore criteria, it is also instructive to examine those few cases where researchers have made such a claim and lost.75 In order to find such losses in court regarding confidentiality, again one has to look to the experience in the United States. In this regard, our reading of the Cecil and Wetherington (1996) special volume of Law and Contemporary Problems (see Appendix A) and other cases with which we are familiar suggests it is useful to distinguish between two types of court ordered disclosure of confidential research information:

1. the compelled disclosure of research information not in the public domain, such as databases, field notes, interview transcripts, audio tapes, and the like; and
2. compelled disclosure of participant identities.

With respect to the compelled disclosure of non-identifying research information, Cecil and Wetherington’s (1996) review shows that the courts have generally operated on the belief that litigants are entitled to the information gathered by university researchers:

1. when that information is relevant to the issues being litigated;
2. when it is the best information available;
3. when the information is not being subpoenaed with the intent to harass researchers, and/or when disclosing the information would not lead to the future harassment of researchers; and
4. when the disclosure of the information will not create an undue burden for the researcher because of any of the following considerations:

The courts' resolution of those cases seems reasonable for a variety of reasons. First, they have, by and large, shown considerable respect for the ongoing viability of research programmes and the special role that the sort of independent evidence gathered by university researchers plays in informing not only the courts, but also policy makers and society generally. Second, we believe that university researchers should be open about the information they gather. Our raison d'être as a social institution is not only to do research on all aspects of society, but also to disseminate that information to our professional colleagues as well as the general public, so that they may benefit from our work. Accordingly, within reason, researchers should be prepared to testify about their research as long as they do nothing to harm research subjects, such as revealing their identities.

When it comes to the compelled disclosure of subject identities, the US courts have taken a very protective stance. They have shown clear recognition that the provision of confidentiality is vital to the research enterprise, and they have gone out of their way to maintain subject anonymity. At the same time, we know of four losses. Our reading of these examples of court ordered disclosure of subject identities shows that three problems were crucial to the outcome:

1. the researcher did not make an unequivocal guarantee of confidentiality;
2. the researcher had an ambiguous relationship with the subject (i.e. it was not clear their relationship was one of researcher and subject); and
3. the researcher gave up the fight and divulged the information when it probably was not necessary.

We examine each of these in turn.

Two cases are relevant here. The first is Atlantic Sugar, Ltd. v. The United States of America. In this case, described by Traynor (1996), Atlantic Sugar had participated in a survey undertaken by the International Trade Commission in which confidentiality was guaranteed "except as required by law." The addition of this phrase was the kiss of death for confidentiality, since the court stated that it "required" the information, and noted that as this requirement had been anticipated in the informed consent statement provided to the respondents, there was no reason to withhold it. This is an example of precisely the situation that has led us to refuse to use the SFU limited confidentiality consent statement – it is unethical because it exposes research participants to harm, and thus violates SFU’s Research Ethics Policy.

The second case is that of Mario Brujaha, a New York University graduate student who was doing his thesis research on the "sociology of the American restaurant" while working as a waiter (he was a participant-observer). Brujaha, it should be noted, did not enjoy the support of his university in his legal battle with the Grand Jury who attempted to subpoena his field notes. When he first appeared, he represented himself. Subsequently he received some pro bono representation from a New York law firm. Only later did he receive assistance from the legal aid lawyers at his university. The greatest difficulty Brujaha ran into was that he had guaranteed confidentiality to "many" of his research subjects, but not all of them, and had not kept a detailed record of those guarantees. This raised the problem of establishing for the court exactly to whom he had guaranteed confidentiality, and which parts of his field notes were thus governed by privilege and which were not. It is important to note that the court seemed quite prepared to accept and respect Brujaha's pledges of confidentiality, and in that sense can be considered a victory. The problem lay not in his provision of confidentiality, but in his failure to articulate, in detail, where and to whom those guarantees of confidentiality were made.

The lesson to be learned from his case is that researchers need to be unambiguously clear about their undertakings of confidentiality if they expect them to be treated seriously by the courts, and should be able to demonstrate that to be the case. In the least intrusive situation, researchers who give oral pledges of confidentiality should be required to make their intended statement a part of the record of their application for ethics approval. Another way to make the pledge of confidentiality clear to all concerned would be to tape record at least the part of the interview in which the pledge of confidentiality is made. When the entire interview is recorded, the pledge of confidentiality should be left as part of the written transcript with all identifiers removed. In the most intrusive situation – where a written and signed informed consent statement is used – the pledge of confidentiality will be most clearly offered and accepted. However, a signed consent form may create its own problem if the record itself could be used as evidence of the subject's participation, and thereby violate confidentiality in the process. This is less worrisome in situations where there is no concern about the subject's being identified as having participated in the research and the only worry is to ensure that particular individuals cannot be tied to particular responses.76

The negative results of these two cases show once again how the SFU limited confidentiality statement is nothing short of legal suicide. The Brujaha case also points to the wisdom of: (a) understanding the legal issues involved and addressing them during the research design process; (b) securing legal representation at the first sign that a legal challenge exists; and (c) of the university actively supporting the researcher.

The relevant case in this regard is that of Richard Scarce,77 a graduate student at Washington State University at Pullman. Scarce was doing long term field research with animal rights activists. He did not submit any sort of research proposal to the WSU research ethics committee, and thus did not obtain university ethics approval. One summer, while Scarce went on a holiday with his family, he engaged one of his research subjects to house sit. At about the same time, considerable damage was done to a university animal research facility by the Animal Liberation Front; the person who house sat for Scarce was a prime suspect.

When Scarce was subpoenaed, he asserted researcher-subject privilege, but both the trial and appeal courts were unsympathetic to his claim. The central problem for the courts was that it was not clear to them whether Scarce's claim was "real," or whether he was merely trying to hide behind researcher-subject privilege in order to protect a friend. In the end, Scarce lost the case, and went to jail for contempt of court (for approximately four months) in order to maintain his promise of confidentiality. Two problems conspired against Scarce in court: (1) the lack of a formal proposal and university ethics approval created the impression that his was not an authentic research project embodying appropriate rigour, peer review, and so forth; and (2) by engaging the research subject as a house sitter, there was an appearance of a conflict of interest. The court was left with the impression that there was more to the relationship than the pact between a "researcher" and a "subject."

Lesson 2, therefore, is that researchers would appear to gain credibility when they obtain ethics approval from the appropriate university authorities. Also researchers should be careful about the impressions that are created when research relationships move beyond the compartmentalized interaction between a "researcher" and a "subject." There is an aspect of this latter element that we find disquieting – it seems more respectful of the detached/aloof/objectified interactions that characterizes quantitative and positivistic research than it is of the intimate relationships that characterize feminist and other types of qualitative field research.

Richard Leo was a graduate student, at the University of California at Berkeley, at the time he was subpoenaed. He had been engaged in research with a major American police department on the topic of police interrogation procedures when a subpoena arrived. It pertained to a case in which one of the people who had been interrogated had brought suit against the police. Since the interrogation had been one that Leo had observed, it appears the belief was that his field notes would contain independent information that would be useful to the litigant in his action against the police. Although Berkeley provided legal assistance at the request of Leo's dissertation supervisor, Leo's impression was that they did so reluctantly, and that they went into it looking for a deal rather than providing the vigorous defence Leo thought warranted.

The irony of it all, and something that caused Leo and his advisor great consternation as they fought the subpoena and prospect of incarceration, was that Leo believed his notes actually vindicated the police officers, who he felt had bent over backwards to ensure that the suspect's rights were observed. Should Leo violate his independent stance and violate the confidentiality of his notes in an effort to help the police officers who were his primary research subjects? Eventually, Leo decided to do so, and produced evidence in the proceedings without protecting confidentiality "to the extent possible within the law." His retrospective remarks on this action are a lesson to us all78:

The Wigmore test provides a legal mechanism for the social science research community to consolidate recognition of researcher-subject privilege through common law, and prevent other researchers from having to confront Richard Leo’s dilemma when he was subpoenaed, and asked to yield confidential information to a court. We urge those who develop and administer university research ethics policies to design those policies and assess research protocols with the Wigmore criteria in mind.

Early on in the Policy Statement, in a section entitled "Ethics and Law," the Tri-Council acknowledges there are times when ethics and law might conflict:

"Ethical approaches cannot pre-empt the application of the law": researchers are not above the law, but the Policy Statement does not say researchers must obey court orders, and we believe it would infringe academic freedom if it did. Indeed, the Policy Statement acknowledges that "ethical approaches may well affect [Law's] future development or deal with situations beyond the scope of the law." It is important to recognize that the common law of privilege creates a situation where research decisions must be made ahead of time, while law is made after the fact. This puts researchers in an unenviable position. In order to satisfy the first criterion of the Wigmore test in the process of establishing privilege a researcher must guarantee unlimited confidentiality. This prerequisite for winning privilege means that, until we have a statutory privilege, or some mechanism akin to US privacy certificates, researchers and universities will have to have the courage of their convictions, and confidence in the courts to rule in favour of researcher-subject privilege.

As to having the courage of their convictions, we recall Judge Steinberg’s comments in Russel Ogden v. SFU about SFU’s failure to appear in Coroner's Court to defend Russel Ogden’s research subjects:

Blomley and Davis echo these same sentiments in their independent consideration of the university's decision-making in the Russel Ogden case:

And we see it again when Judge Steinberg directs a final injunction to all researchers:

We take President Blaney's acceptance of the Blomley and Davis report and its recommendations – including an apology to Russel Ogden, and an announcement that, in future, graduate students would have the same access to legal representation as faculty when academic freedom is challenged by a court – as a statement that the university wishes to put the Russel Ogden chapter behind us. The VP-Research and URERC's imposition of limited confidentiality is the final vestige of that ethics regime; it is inconsistent with current university policy.

When it comes to faith in the law, we remind readers of the unique position we are in as researchers. There are many persons who must place their bets ahead of time with respect to concepts that become defined after-the-fact. An Aboriginal person who exercises what s/he believes are his/her Aboriginal rights, might be arrested and have the event lead to a Supreme Court decision that adds to the case law regarding the definition of the term. An owner of an art emporium might acquire an erotic statue, be arrested, and have the event lead to a Supreme Court decision that adds to the case law defining "obscenity." We, too, place bets, i.e., that the research we (individually and collectively) engage in is of such social importance that it warrants the granting of researcher-participant privilege, since it is only because of research subjects' confidence in its existence, and our willingness to defend it (and them), that they agree to share what they know.

The element that makes our wager unique is that we have to stake someone else's well being in addition to our own. It is also what makes the SFU URERC's caveat emptor approach to ethics all the more unethical. We suggest that "informed consent" should not be turned into a mechanism for researchers to exploit subjects or for universities to manage liability. Subjects make a choice as to whether to participate in research, but they do so only because we ask them. We believe we have a unique responsibility to maintain what is really the subject's privilege, unless they unchain us from that responsibility. Also, we believe this feature of the relationship makes a researcher’s claim to privilege all the more compelling. As Wigram stated in an 1851 decision: "The rule [as to whether privilege exists] does not rest simply upon the confidence reposed; … it seems to rest, not upon the confidence itself, but upon the necessity of carrying it out."82

The implication for us is that once the decision is made to engage in research, then we have no choice but to have faith in courts to decide in our favour, and be prepared to live with the consequences if the Supreme Court were to happen to disagree with us. This would include the possibility of defying a court order, and facing the consequences should the Supreme Court disagree. Traynor, a US lawyer writing in the Law and Contemporary Problems special issue on court-ordered disclosure, addresses exactly this situation:

Our approach is guided by the view that a formal recognition of the confidentiality requirements of the researcher-subject relationship requires assertion of the Wigmore criteria whenever the principle of confidentiality is challenged. According to the Supreme Court of Canada in R. v. Gruenke, the law of the land requires persons who claim privilege to have made an unequivocal assertion that their communications originated in the understanding that they would not be disclosed. Having considered the law and followed it at every turn, we have every confidence that the courts will agree with our logic.

If we get to the Supreme Court of Canada and it does not agree, then we also believe that, ethically, we would have no choice but to accept the sanctions that derive from our misplaced faith. To do any less would have made our pledge of unlimited confidentiality a lie. If we believe in the long-run applicability of the Wigmore criteria, then it is only by taking this path that researchers would send an unequivocal statement that they are so committed to the importance of confidentiality to research and to the belief that the researcher-subject relationship is a social good that should be sedulously fostered, they are willing to accept the sanctions of the court to protect it. No code of ethics can force a researcher to do this. Nor should any ethical code of ethics ever preclude it. The Tri-Council Policy Statement does not. We trust the new SFU research ethics policy will not do so either.

Case Reference:

• Richards of Rockford, Inc., v. Pacific Gas and Electric Co., 71 F.R.D. 388 (N.D. Cal. 1976)

Outcome: Confidentiality of subject identities maintained.

Commentary:

• The District Court considered confidentiality of research sources. "A non-party university professor was subpoenaed to produce documents concerning confidential interviews of the defendant's employees that were undertaken as part of a research project. The court refrained from deciding whether a constitutional privilege existed that would protect researchers from disclosing the confidentiality of their sources, although it noted 'that society has a profound interest in the research of its scholars, work which has the unique potential to facilitate change through knowledge.' (Crabb, p.23)
• "The court held that the societal interest in protecting the confidential relationships between academic researchers and their sources outweighed the interests of this litigant and the public in obtaining the research data." (Crabb, p.24)
• "A Harvard public health professor, Marc Roberts, had conducted extensive interviews with employees of Pacific Gas and Electric Company (PG&E) in Northern California, during a study of how utilities make decisions about environmental issues. Later a construction company sued PG&E for breach of contract in federal court. During pre-trial discovery, the plaintiff suggested that information from Roberts's study might bear on the corporate judgement that triggered the alleged breach. PG&E then sought access to the notes of the study interviews; however, Roberts resisted granting PG&E the right to view his research." (O'Neill, p.38)
• "Roberts asserted that 'compelled disclosure of confidential information would without question severely stifle research into questions of public policy, the very subjects in which the public interest is greatest." (see O'Neil, p.38, footnote 17)
• Wiggins & McKenna give another description of this case, and cite the same kinds of conclusions as authors above, i.e., that the finding was "that the public interest in maintaining confidential relationships between academic researchers and their sources outweighed the plaintiff's interest in the subpoenaed information …" (pp.74-75)
• In reference to the case, Traynor notes, "Despite its failure to recognize a researcher's privilege, the court affirmed the importance of maintaining confidentiality. It recognized that '[c]ompelled disclosure of confidential information would without question severely stifle research into questions of public policy, the very subjects in which the public interest is the greatest.'" (p.128)

Case Reference:

• Atlantic Sugar, Ltd. v. U.S., 85 Cust. Ct. 128 (1980)

Outcome: Researcher ordered to disclose data, and complies.

Commentary:

• "Researchers frequently qualify their assurances by adding a proviso that confidential data will not be disclosed except as required by law. Such a proviso may alert the source to the possibility of compelled disclosure and may strengthen the researcher's defense against a claim of liability premised in contract, promissory estoppel, or tort in the event of such a disclosure. On the other hand, such a proviso could lead the party subpoenaing the data to contend that the possibility of compelled production was anticipated and that enforcement of a subpoena, therefore, is not inconsistent with the qualified assurance given." (p.122) Traynor cites the Atlantic Sugar case here (in footnote 12 on p.122), and describes it as "compelling disclosure of a non-party's answer to an International Trade Commission questionnaire, noting that persons who responded were informed that the information would not be disclosed "except as required by law."

Case Reference:

• Dow Chemical Co. v Allen, 672 F.2d 1262, 1274-77 (7^th Cir. 1982)

Outcome: Confidentiality of subject identities maintained.

Commentary:

• "In Dow Chemical Co. v. Allen, the Court of Appeals for the Seventh Circuit suggested that the First Amendment might offer protection against a 'judicially authorized intrusion' into the scholarly research activity of university researchers. The court was addressing a researcher's challenge to a subpoena issued by the Environmental Protection Agency at the request of the Dow Chemical Company. The subpoena sought information about ongoing studies at the University of Wisconsin involving dietary ingestion by rhesus monkeys of the chemical 2,3,7,8-terachlorodibenzo-p-dioxin Dow wanted the information for use in cancellation hearings before the agency on certain uses of two herbicides manufactured by Dow. The First Amendment issue was raised in the district court but not reached in granting the researchers' motion to quash. The court of appeals chose to discuss the issue [of First Amendment protection], characterizing Dow's efforts to subpoena ongoing research as a threat of 'substantial intrusion into the exercise of university research … capable of chilling the exercise of academic freedom,' which would 'inevitably tend[ ] to check the ardor and fearlessness of scholars …'. The court did not go so far as to hold that researchers were entitled to an evidentiary privilege, but concluded that where a researcher's interest in academic freedom is involved, 'the interests of the government must be strong, and the extent of the intrusion carefully limited. The court defined academic freedom as 'the right of the individual faculty member to teach, carry on research, and publish without interference from the government, the community, the university administration, or his fellow faculty members." (Crabb, pp.21-22)

Case Reference:

• Deitchman v. E.R. Squibb & Sons, Inc. 740 F.2d 556 (7^th Cir. 1984)

Outcome: Confidentiality of subject identities maintained.

Commentary:

• Squibb and other drug companies were defendants in civil actions brought by plaintiffs who alleged their mothers took DES when pregnant, and that it gave them adenocarcinoma of the vagina. Squibb served deposition on Dr. Arthur Herbst, Chair of the Dept. of Obstetrics and Gynecology at U of Chicago, seeking every document in the records of a registry he had established of vaginal and cervical adenocarcinoma contracted by women since 1940. Herbst had promised information would remain confidential; maintained that if confidence were breached, sources would dry up. Federal District court agreed with Herbst. The Court of Appeals reversed decision, and referred matter back to lower court, saying subpoena was OK, but confidentiality must be maintained. (Crabb, pp.10-12)
• "All society would be poorer … [because] a unique and vital resource for learning about the incidence, causes[,] and treatment of adenocarcinoma would be lost" -- Squibb acknowledges this, too. (Crabb, p.15)
• Wiggins & McKenna give a fairly detailed description starting at p.79. Some interesting new information: "In every case, Dr. Herbst promised the submitting physician that the information received would be maintained in confidence. Once records were received for an individual patient, Dr. Herbst would contact her and her mother's obstetrician to request additional records and information about the use of DES. In his subsequent contacts with patients and their mothers' physicians, confidentiality of the information was again assured. By this assurance, Dr. Herbst intended to promise that the information received would be used for research purposes only and would be released only in a form that did not allow identification of any patients." (pp.79-80)

Case Reference:

• In re Grand Jury Subpoena, 583 F. Supp. 991 (E.D.N.Y.), rev'd, 750 F.2d 223 (2d Cir. 1984)

Outcome: Researcher won at trial, but lost on appeal. The problem was not with the promise of confidentiality, but with the non-documented and sporadic way it was offered.

Commentary:

• Graduate student's study brought to a halt by involuntary involvement in a Grand Jury investigation of a restaurant fire. Name is Mario Brujaha.
• Court of Appeals not receptive to the idea of a First Amendment privilege, though they seemed ready to consider it on a case-by-case basis. Court reversed a lower court decision that quashed a Grand Jury subpoena. The student had testified, but refused to produce the journal. From the quote in Crabb (p.23), it looks like they justified the reversal by saying that he had not shown he was engaged in real "research" governed by real "standards," and that he hadn't demonstrated why the materials should be subject to privilege. (Crabb, pp.22-23)
• Has a good page long description of the case.
• Trial court ruled in the student's favour. Judge noted, "Affording social scientists protected freedom is essential if we are to understand how our own and other societies operate. Recognized by cultural anthropologists since at least the turn of the century as a basic tool, fieldwork is used widely in other disciplines, particularly in sociology and political science. In order to work effectively researchers must record observations, communications and personal reactions contemporaneously and accurately." (O'Neil, p.41)
• Appeal court felt that the case for privilege had not been made, and said that it would require an argument of the precise nature of the adverse effects that disclosure would have. The court stated, "The district judge's approach had raised 'an arguable question as to the validity of a qualified privilege where a serious academic inquiry is undertaken pursuant to a considered research plan in which the need for confidentiality is tangibly related to the accuracy or completeness of the study.'" (O'Neil, p.41)
• "Mario Brajuha received pro bono representation from a New York attorney and later from a legal clinic associated with New York University Law School, but only after university counsel refused to represent his interests and he had testified in court unrepresented." (Wiggins & McKenna, p.78)
• "Three briefs were submitted on behalf of Mario Brajuha; one by three professional associations, one by the American Association of University Professors, and one by the New York Civil Liberties Union." (Wiggins & McKenna, p.78)
• "Mr. Brajuha has described in detail the vacillating level of university and departmental support he received." (Wiggins & McKenna, p.79) -- [See Mario Brajuha and Lyle Hallowell, "Legal intrusion and the politics of field work: The impact of the Brajuha case." Urban Life, 1986, 14, 454-?.]
• Seems to have been some ambiguity regarding promises of confidentiality. At p.82, Wiggins and McKenna note that, "Mr. Brajuha stated that 'many' of the research sources identifiable from the notes had been promised confidentiality." More facts of the case are discussed on pp.82-83.

Case Reference:

• Farnsworth v. Proctor and Gamble Co., 758 F.2d 1545, 1546-47 (11^th Cir. 1985)

Outcome: Confidentiality of subject identities maintained.

Commentary:

• Notes that one of the factors considered by the courts is whether the researcher's sole concern is with respect to confidentiality (as opposed to the general burden to testify). Notes summary statement from case that "production of highly personal information given in a study on 'toxic-shock' syndrome might inhibit future studies. (Crabb, p.27)
• Suit involved the Centers for Disease Control (CDC). "The Eleventh Circuit accepted the CDC's plea to keep confidential the identity of subjects who had taken part in the toxic shock studies, even though the subjects had apparently not been given express pledges of anonymity. In reaching the result, the Eleventh Circuit stressed two factors: first, that the CDC's mission was to protect the health of the US public through 'scientific and social research supported by a population willing to submit to in-depth questioning'; and second, that subjects had a reasonable expectation of confidentiality even in the absence of express promises that their names would not be revealed." (O'Neill, p.40)
• "CDC wanted to remove from the documents before production all information that would allow identification of study participants who did not consent to release of information pertaining to them. P&G pressed for identifying information to enable it to contact the subjects of the CDC investigations personally, "to determine whether the studies were properly conducted." Declining to rule on the CDC's claim of a general confidential privilege for information provided by study participants, the courts in the reported decisions invoked Federal Rule of Civil Procedure 26(c) and quashed the subpoenas to the extent they sought production of identifying information." (Wiggins & McKenna, p.73)
• More description of this case by Wiggins & McKenna at p.81, including quotes from decision regarding importance of maintaining confidentiality.

Case References:

• Wright v. Jeep Corp., 547 F. Supp. 871, 876 (E.D. Mich. 1982)
• In re Snyder, 15 F.R.D. 211 (D. Ariz. 1987)

Outcome: Confidentiality of subject identities maintained.

Commentary:

• An early reference to Snyder (on 9.) makes reference to him being subpoenaed to appear in 80 cases, and that he is devoting his retirement to fending off subpoenas. Case reference at left described as: "Denying Professor Snyder's motion to quash subpoena requiring him to provide Jeep Corporation with all of the research data, correspondence, and other materials pertaining to study of on-road crash experience of utility vehicles in which he participated." (see Crabb, footnote 1, p.9)
• Then Buchanan v. American Motors Corp., No. 81-436 (E.D. Mich. Oct. 23, 1981), aff'd, 697 F.2d 151 (6^th Cir. 1983) is cited, as follows: "Granting Snyder's motion to quash subpoena requiring him to provide the same research data, on ground that it was unreasonably burdensome." (see Crabb, footnote 1, p.9)
• Seems lots of litigants -- American Motors as well as lots of people trying to sue American Motors -- were interested in Michigan State Professor Snyder's data regarding the Jeep CJ-5's propensity to roll over. "In an early round, the company prevailed, though only after disavowing any desire to discover 'confidential sources' and agreeing to the removal before discovery of the names of any individual accident victims who had been the subject of the study. Later, when AMC took a harder line, a federal judge in Arizona quashed the company's subpoena for Snyder's data, noting 'the potential for a chilling effect on research.' The court cautioned that, despite the absence of a blanket privilege 'discovery offers an avenue for indirect harassment of researchers whose published work points to defects in products or practices. There exists also a potential for harassment of members of the public who volunteer, under a promise of confidentiality, to provide information for use in such studies.'" (O'Neil, p.40)

Case Reference:

• In re American Tobacco Co., 880 F2.d 1520 (2d. Cir. 1989).
• In re American Tobacco Co., 866 F2.d 552 (2d. Cir. 1989).

Outcome: Confidentiality of subject identities maintained.

Commentary:

• "In one case, various tobacco company defendants sought detailed information about certain studies conducted by Dr. Irving Selikoff, in association with the American Cancer Society, at the Mount Sinai School of Medicine." (p.69) It was a very broad subpoena, and Selikoff won; the court cited the undue burden it would put on Dr. Selikoff to comply. In a subsequent two cases (referenced at left), American Tobacco Company was more specific in its requests. The tobacco company won access to his computer tape, while the court entered "a protective order that permitted the researchers to redact the information identifying research participants and that prohibited the tobacco companies from using the released data to identify the research participants." (Wiggins & McKenna, p.70)

Case Reference:

• Kennedy v. Connecticut Department of Public Safety, 115 F.R.D. 497 (D. Conn. 1987)

Outcome: Confidentiality of subject identities maintained.

• "A federal magistrate judge rejected a motion to quash a subpoena calling upon a researcher to disclose all material relating to her study of the integration of women into the Connecticut State Police. The court was unpersuaded by the researcher's assertion that production would be burdensome and that the plaintiffs had alternate sources of information. The court allowed her to redact the names of persons to whom she had promised confidentiality, but required that she do the redactions at her own expense." (Crabb, pp.23-24).

Case Reference:

• Defendant's Notice of Out of State Deposition on Oral Examination and Request for Production of Documents and Things, Mangini v. R.J. Reynolds Tobacco Co., No. 939359 (Super Ct., San Francisco Cty., Cal., Mar. 30, 1992)
• Defendant's Notice of Out of State Deposition on Oral Examination and Request for Production of Documents and Things, Mangini v. R.J. Reynolds Tobacco Co., No. 939359 (Super Ct., San Francisco Cty., Cal., Mar. 30, 1992)
• Motion to Quash, Fischer v. R.J. Reynolds Tobacco Co., No 93-RCCV-230 (Ga. Super. Ct., Richmond County, Apr. 16, 1992)
• Fischer v. R.J. Reynolds Tobacco Co., No 93-RCCV-230 (Ga. Super. Ct., Richmond County, Apr. 28, 1992) (order granting motion to quash)
• R.J. Reynolds Tobacco Co., v. Fischer, 207 Ga. App. 292 (1993)
• R.J. Reynolds Tobacco Co., v. Fischer, 427 S.E.2d 810, 811 (Ga. Ct. App. 1993)

Outcome: Confidentiality of subject identities maintained.

• Researchers required to review and delete identifying information from 97 file drawers of documents. The Joe Camel study. [By the way, Crabb cites the case incorrectly here; she confuses the Fischer case with another involving Selikoff; [the apparently correct case reference appears above.]
• Mangini was suing R.J. Tobacco in California because they were leaving health warnings off of merchandise, such as caps and t-shirts. Fischer's research was not cited in the litigation, but he got the subpoena anyway. (see Fischer article, p.160)
• The researcher, Fischer, won the battles in the courts regarding privilege. They ruled the data were irrelevant to the case that RJReynolds said they needed it for. But then RJR went after Fischer through a freedom of information request. "The state's attorneys gave Fischer little solace or support, insisting at one point that he would be suspended within forty-eight hours if he refused to accede to demands to release the data. The files were then turned over to the court pending resolution of the competing claims. In the summer of 1994, the trial judge ruled against Fischer on the freedom of information claim, and Reynolds immediately obtained the data it had sought." (O'Neil, p.43).
• "From the beginning, the tobacco industry attempted to discredit this research and harass the researchers. My experience in confronting the tobacco industry has taught me how easily the courts can become the unwitting accomplices of an industry whose goal is profit, not the identification of scientific truth." (Fischer, p.159)
• Fischer was overwhelmed by the magnitude and inclusiveness of the subpoena, and while he thought he might be subpoenaed, "I was not, however, prepared to receive a subpoena of this breadth that would require turning over the names of three- to six-year old children." (Fischer, p.161)
• "I had also anticipated that the Medical College of Georgia ('MCG'), on whose faculty I was a full professor and under whose auspices the research had been conducted, would provide full legal support for my position. … I refused to comply with the subpoena and MCG refused to provide me with legal assistance." (Fischer, p.161)
• After Fischer won in court, the university attorney stated in a newspaper article that they had not supported Fischer because the university thought it was an issue under the Open Records Act, which it believed would have required them to release the info. RJR hadn't known this, but read about it, and immediately sent a letter to the Chancellor of the Georgia university system, and the President of MCG, requesting that the research records be released. "I was given forty-eight hours to turn over all of the previously described records with the exception of the children's names. Clay Stedman, as MCG legal counsel, indicated that I would be suspended if I did not turn over the documents. Francis Tedesco, M.D., President of MCG, indicated that the Attorney General would have me arrested if I did not comply with the request." (Fischer, p.162)
• Not a pleasant university experience. The one good thing is that the kids' names were never released, and even RJR backed off on that aspect of the request. Fischer resigned and went into private practice at the end of that year.
• In a couple of his final thoughts, he notes:
• "Despite institutional affiliation and responsibilities to protect academic freedom, universities may provide poor legal counselling to scientists facing compelled disclosure. This problem may become greater due to the increased reliance of universities on corporate support. We might expect to see university presidents siding with corporate contributors rather than their academic faculty." (Fischer, p.167)
• "Human subject confidentiality, promised as part of the research process, must be protected at all costs." (Fischer, p.167)

Case References:

• Motion for Protective Order, In re the Exxon Valdez Re: All cases, Misc. 92-0072 RV-C (S.D. Ala. Oct. 13, 1992)
• Protective order, id. (S.D. Ala. Oct. 14, 1992)
• Defendants' Motion to Compel, In re the Exxon Valdez Re: All cases, Misc. 92-0072 RV-C (S.D. Ala. March 18, 1993)
• Order at 7-8 n.3, In re the Exxon Valdez Re: All cases, Misc. 92-0072 RV-C (S.D. Ala. July 1, 1993) (order)
• Motion to Amend Court Order, In re the Exxon Valdez Re: All cases, Misc. 92-0072 RV-C (S.D. Ala. July 2, 1993)
• Order at 3, Id. (S.D. Ala. July 9, 1993)
• Exxon's Request for Further Hearing With Respect to the Court's Order Of July 1, 1993, In re the Exxon Valdez Re: All cases, Misc. 92-0072 RV-C (S.D. Ala. August 9, 1993)
• Outcome: Confidentiality maintained.

• "Courts may also craft procedural safeguards to protect the confidentiality of research subjects. For example, in 1992 a University of Alabama sociologist, J. Steven Picou, was subpoenaed by the Exxon Shipping Co., to make available his work on community stress in Alaskan coastal villages following the Exxon Valdez oil tanker disaster. The Federal District Court for the Southern District of Alabama granted Exxon's request but stipulated that the information sought could be reviewed only by the expert sociologist whom Exxon had retained to assess Picou's files. In turn, the expert was warned that he could be cited for contempt if he failed to protect confidentiality." (Jasanoff, p.108; footnote 57; see also p.111)
• "As an independent, third-party researcher, I directed the collection of survey data related to the disaster in several small Alaskan communities from 1989 to 1992." (Picou, p.150) Funded through the National Science Foundation on a 4-year grant.
• "The longitudinal research design necessitated a detailed record of respondent identities for re-interviewing, required constant field work and data management, and included specific methodological procedures to protect respondent privacy.
• Respondents were selected by random procedures and guaranteed confidentiality." (Picou, p.151)
• Started off by receiving subpoena. Main priorities were protecting survey data and respondent confidentiality. University (Alabama) sought and received a protective order (see Traynor regarding protective orders). Next month he was deposed by an Exxon attorney, and gave up all sorts of documents and files, but retained data. "The next four months were frantic and extremely stressful for me. Throughout this time, Exxon sought access to the data by negotiating directly with the university attorney and through motions and affidavits offered to the court by their attorneys and experts. Various news accounts of my litigation were published in Alaska newspapers, and rumors about my case reached me from the communities I studied. Several respondents expressed fear and concern regarding the release of the data, respondent identifiers, and information on their involvement in the Exxon litigation. The ethical priority of protecting respondent confidentiality became my obsession when a 'concerned' respondent committed suicide." (Picou, p.152)
• Some of the cases they used included Deichtman v. Squibb, Farnsworth v. Proctor and Gamble, and Dow Chemical v. Allen. In the end, it sounds like Exxon was perturbed because a key variable of interest to them -- which respondents were involved (or not) in the Exxon Valdez litigation -- was not part of the data. They pressed for it anew, and were told that they were changing the initial application and should cease and desist, (p.154) In the end, "no disclosure of respondent identifiers occurred and the court specifically denied any form of 'deductive disclosure' by prohibiting the identification of individuals through a protective order." (Picou, pp.154-155)
• "In my case, I was fortunate to have strong and continuing support from my university, department, counsel, colleagues, and family." (Picou, p.155)
• "The researcher also takes the risk of having his or her professional integrity, research methodology, and personal ethics challenged through the 'deconstruction' of his or her work. In my case, this deconstruction began with a comparison of the procedures used in the research design I employed with a gold standard design used for conducting ideal experiments. Obviously, this tactic was exclusionary; that is, Exxon's experts attempted to build a case 'to deny scientific status' to any results that expanded the range of damage claims by plaintiffs. In short, 'mindless deconstruction' of one's research is a very real risk to the researcher." (Picou, p.155)

Case Reference:

• Daubert v. Merrell Dow Pharmaceuticals, Inc. 509 U.S. 579 (1993)

Outcome: N/A

• US Supreme Court decision says that trial judge has a gatekeeping responsibility, to ensure the data are reliable and valid, what its status is in scientific community. (Crabb, p.14)
• She refers to this case as "arguably the most influential decision concerning the admissibility of scientific evidence in the last half century" because it "encouraged judges to ask what makes science 'scientific' and to apply those criteria in discriminating good scientific evidence from bad." (Jasanoff, p.112). One of these criteria was peer review. Jasanoff discusses aspects of this criterion on p.112.

Case Reference:

• Scarce v. United States, 5 F.3d 397 (9^th Cir. 1993), cert denied, 510 U.S. 1041 (1994).

Outcome: Disclosure ordered by trial and appeal courts, but maintained by researcher by going to jail

• Washington State University doctoral student "was called before a federal grand jury that was probing a destructive raid on the university's laboratory animal facilities conducted by the Animal Liberation Front ('ALF'). A prime suspect in the case was a member of the ALF who had house-sat the graduate student's residence during his summer absence." (O'Neil, p.42).
• District Court ruled against Scarce, and so did Court of Appeals. The things that compromised his ability to protect the anonymity of his research participants were (a) he had not received ethical approval from the university; and (b) the house-sitting suggested that his relationship with the suspect was more than just a "research" relationship. In combination, the two factors seemed to have raised doubts as to whether the claim for privilege was legitimate, or simply an expedient way of protecting a friend.

Crabb quoting Wigmore at 2192:

O'Neil (p.36) talks about four devastating effects that forced disclosure can have on the research process: (1) the researcher loses control of the reporting and disclosure process; (2) dealing with the subpoena may severely hamper the research process (e.g., huge time required to reproduce documents); (3) if subpoenaed before publication, thereby involves the use of unverified information, i.e., that hasn't gone through the peer review process; (4) special concerns regarding confidentiality.

Wiggins & McKenna discuss the aftermath to researchers of being drawn into the court-ordered disclosure process. The article is based on published reports and interviews done with the principal researchers involved in such cases. By way of summary, the authors note,

Wiggins and McKenna note that none of the researchers they contacted were required to pay the costs of defending against the subpoena:

Traynor (p.137) notes a case involving Sweezy v. New Hampshire, (354 U.S. 234 (1957)), in which Paul Sweezy had refused to cooperate with a state attorney general's investigation; he had given a lecture at a public university, and was allegedly a member of a political party declared subversive. The US Supreme Court believed that academic freedom and freedom of association were more important than the state's need to maintain order, and that "academic freedom and political expression [are] areas in which government should be extremely reticent to tread." (p.137). He quotes the court:

All cases reported in this Appendix were reported by one or more authors in a special issue of Law And Contemporary Problems devoted to the topic of court-ordered disclosure. The full reference is:

Of the ten articles in the volume, the following contributed information to the summary of cases above:

• Barbara B. Crabb, "Judicially compelled disclosure of researchers' data: A judge's view," pp.9-34.
• Robert M. O'Neil, " A researchers' privilege: Does any hope remain?" pp.35-50.
• Elizabeth C. Wiggins and Judith A. McKenna, "Researcher's reactions to compelled disclosure of scientific information." pp.67-94.
• Sheila Jasanoff, "Research subpoenas and the sociology of knowledge." pp.95-118.
• Michael Traynor, "Countering the excessive subpoena for scholarly research." pp. 119-148.
• J. Steven Picou, "Compelled disclosure of scholarly research: Some comments on "high stakes litigation."" pp. 149-158.
• Paul M. Fischer, "Science and subpoenas: When do the courts become instruments of manipulation? pp. 159-168.