Download Ethical Treatment of Research Participants: Traditional vs. Modern Approaches in Education and more Lecture notes Ethics in PDF only on Docsity!
Chapter 2
Ethics in Educational Research
KENNETH R. HOWE AND MICHELE S. MOSES University of Colorado at Boulder
The ethics of social and educational research has been significantly complicated over the last several decades as a consequence of the "interpretive turn" and the ever-increasing use of qualitative research methods that have accompanied it. In this chapter, we identify what came before and after the interpretive turn with the traditional and contemporary approaches to research ethics, respectively. The distinction is a heuristic one. We do not mean to suggest that the interpretive turn occurred at any precise point in time or that it has completely won out. In this vein, the traditional approach is no doubt still in currency. Embedded in the distinction between traditional and contemporary approaches is another between the protection of research participants ("research subjects" in the traditional vocabulary) and research misconduct. This, too, is a heuristic distinction, because it involves significant overlaps. In particular, research mis- conduct largely subsumes the protection of research participants. Nonetheless, it is a distinction that has the virtue of familiarity, since it parallels the way federal regulations and universities divide the issues in research ethics. We should observe here at the outset that medical research has been at the forefront of the ethics of research involving humans, both with respect to the development of vocabularies and frameworks and with respect to the formulation of federal policy. Social research in general and educational research in particular have generally followed this lead. We do not make this observation to suggest that social and educational researchers have remained on the sidelines, simply applying the precepts of medical ethics. On the contrary, as we shall see, for at least some theorists an adequate approach to the ethics of social and educational research requires significantly modifying the vocabularies and frameworks that have come down to them through the ethics of medical research. We make this observation instead to apprise readers of why we borrow so heavily from sources outside education and to alert them to an important part of the history of the ethics of educational research.
THE TRADITIONAL APPROACH The "traditional approach" draws a rather sharp line between the "prescrip- tive" (moral-political) component of social research and the "descriptive" (scien- tific-methodological) component (Beauchamp et al., 1982). It divides questions concerning the morals and politics of social scientific studies from questions
21
Howe, K. & Moses, M. (1999). Ethics in educational research. Review of Research in Education, vol. 24, pp. 21-60.This work may be downloaded only. It may not be copied or used for any purpose other than scholarship. If you wish to make copies or use it for a non-scholarly purpose, please contact AERA directly.
22 Review of Research in Education, 24
concerning their scientific merits and pursues them relatively independently. Indeed, not keeping these domains separated is often considered the mark of biased social research and advocacy. In our discussion of the traditional view, we follow suit and separate the ethics of social research from broader political and methodological issues. We save raising questions about this maneuver for our discussion of the "contemporary approach."
Protection of Research Participants Protecting individual autonomy has long been a central principle in Western moral-political thought. That it should also occupy a central place in Western thought about the ethics of research, particularly research involving human partici- pants, should thus come as no surprise. The traditional controversy about auton- omy has been cast in terms of Kantian versus utilitarian ethical frameworks. The Kantian (also nonconsequentialist or deontological) framework employs
categorical ethical principles, the most general of which is the following:
"Always treat persons as ends in themselves and never solely as means." In this framework, individual autonomy is fundamental, since respecting autonomy is tantamount to treating individuals as ends in themselves. By contrast, the
utilitarian (also consequentialist or teleological) framework employs hypothetical
ethical principles that are subsidiary to the uniform goal of maximizing the balance of benefits over harms. Thus, "always treat persons as ends" is subject to the condition " / f this maximizes benefits." In this framework, autonomy is instrumental; respecting it must serve the overall utilitarian goal. Utilitarian reasoning is widely criticized for sanctioning unacceptable moral conclusions in both theory and practice. In theory, one can imagine a society in which benefit is defined as what gives people pleasure and in which fights to the death among enslaved combatants serve to maximize the total balance of pleasure over pain--a society in which, by summing the pleasure experienced by the spectators and subtracting the pain experienced by the combatants, the total value would be higher than if the practice were prohibited. In practice, certain biomedical research (e.g., the Tuskegee study of the progression of untreated syphilis in African-American men [Jones, 1993]), as well as social research (e.g., Milgram's [ 1974] studies of obedience), receives its sanction from utilitarian reasoning when it is defended on the grounds that the harm done to research participants in the short term is outweighed by the long-term benefit of the knowledge produced. To be sure, these are research practices that many (perhaps all) utilitarian theorists would condemn. Arguably, however, it is only by adopting rule utilitari- anism and providing utilitarian reasons for following moral rules independent of the calculation of specific benefit-harm ratios--that is, only by providing utilitar- ian reasons for eschewing utilitarian reasoning in specific cases--that the unac- ceptable moral conclusions that follow from a thoroughgoing application of utilitarian reasoning can be blocked.
24 Review of Research in Education, 24
use to which such calculations can be put. This is true for moral wrongs as well because they involve the rights to self-determination and privacy, rights that, in Dworkin's (1978) suggestive phraseology, " t r u m p " utilitarian calculations. As mentioned earlier, there is a version of utilitarianism that putatively avoids the kinds of criticisms just advanced, namely rule utilitarianism. Kelman (1982), a self-described rule utilitarian, provides a good example of such a view applied specifically to the ethics of social research. The benefit that Kelman ultimately seeks to maximize is the "fulfillment of human potentialities" (1982, p. 41 ). He concedes, however, the extreme difficulty involved in determining whether this applies in specific circumstances and, for this reason, rejects act utilitarianism. He goes on to use "consistency with human dignity" as his criterion for moral evaluation (1982, p. 42), which he subsequently identifies (in language almost straight from Kant) with treating "individuals as ends in themselves, rather than as means to some extraneous ends" (1982, p. 43). In a related vein, under the rubric o f ' 'wider social values" (1982, p. 46), Kelman embraces the idea that social research should avoid engendering "diffuse harm," the "reduction of private space," and the "erosion of trust." The parallel between Kelman's and Maclntyre's views is striking. Correspond- ing to Maclntyre's admonition to avoid "moral wrongs," we have Kelman's admonition to treat persons as "ends in themselves"; corresponding to Macln- tyre's admonition to avoid "moral harms," we have Kelman's to avoid "diffuse harms." In both cases, confinement to utilitarian benefit-h~u-m calculations is viewed as morally inadequate. If moral justification is to be ultimately utilitarian, to ultimately fall under the rule of benefit-harm calculations, then it is not only individually defined benefits and harms that must be taken into account but also benefits or harms to the moral health of the human community overall. This should explain why we would say that thinking about the ethical treatment of participants in social research is de facto Kantian: There is rather widespread agreement that whatever the ultimate justification for moral conclusions regarding the treatment of research participants might be, certain ethical principles should constrain the manner in which researchers may treat research participants in meeting the traditional utilitarian goals of advancing knowledge and otherwise benefitting society. Informed consent is the most central of such ethical principles, and it is prominent in federal regulations governing social research. The basic idea is that it is up to research participants to weigh the risks and benefits associated with participating in a research project and up to them to then decide whether to take part. And they can do this only if they are informed about and understand what their participation in the research involves. In this way, their autonomy is protected in a way it was not in the Tuskegee, Tearoom Trade, and Milgram studies. Informed consent is de facto Kantian because refusal to participate on the part of research participants is binding, even if their refusal results in a failure to maximize presumed benefits.
Howe and Moses: Ethics in Educational Research 25
It should be observed that the doctrine of informed consent, as explicated in the Code of Federal Regulations for the Protection of Human Subjects (45 CFR 46, 1991, as amended), is not so permissive as to sanction any research in which humans agree to participate. Special protections are provided to "vulnerable populations" of various kinds (for example, children and prisoners) who are too immature or in too compromised a position to make a truly autonomous decision. Furthermore, researchers must seek a just distribution of the burdens of participat- ing in social research over different populations. Finally, proposed research projects may also be blocked if they are so methodologically flawed as to provide little or no promise of generating credible findings, rendering any potential risks or harms too great. Privacy is the second central principle in the traditional conception of the ethical treatment of social research participants, in addition to autonomy. The two vehicles for protecting it are anonymity (not gathering identity-specific data) and confidentiality (not revealing identity-specific data). The relationships among autonomy and informed consent, on the one hand, and privacy, confidentiality, and anonymity, on the other, are varied and complex. In one form of analysis, privacy is a kind of autonomy. For example, in the celebrated Roe v. Wade decision (1973) and the precedent on which it depended, Griswold v. Connecticut (1964), the right to privacy was invoked as equivalent to protecting the autonomy of individuals regarding abortion and birth control, respectively. In a related way, autonomy has also been advanced as the justifica- tion for protecting the privacy of medical records, school records, and social research data, on the grounds that the release of such information can restrict the options available to those about whom it is released. For example, depending on the circumstances, the release of medical records could reduce a person's employment opportunities, the release of school records could label a child and thereby restrict his or her life options, and the release of social science data could cause someone to be deported. In a second form of analysis, privacy has an intrinsic value tied to human dignity and security and distinct from its relationship to autonomy (e.g., Beauchamp et al., 1982). According to Arthur Caplan, "Privacy is a basic human need. Without privacy, it is not possible to develop or maintain a sense of self or personhood." Thus, the attempt to derive privacy from autonomy puts "the cart before the proverbial horse" (Caplan, 1982, p. 320). Save the limiting case of not choosing to be observed, the value of privacy may have little or no connection to self- determination. It is the value of having "private space" (e.g., Kelman, 1982), of being free from surveillance, from looking over one's shoulder, from humilia- tion and embarrassment, and the like, that privacy protects. Consider the reason one would not want to be observed by strangers going to the toilet or engaging in sex. As it turns out, the requirement of informed consent diminishes the importance of determining which analysis of privacy is the correct one, and in which contexts. For part of the informed consent process is describing to participants just what
Howe and Moses: Ethics in Educational Research 27
minds of today's researchers. In many instances, they are emotional reminders of how researchers can lose their integrity even in the search for knowledge and truth. These incidents of misconduct compromise the integrity not only of the researchers involved but of the entire research community (Chubin, 1985). Mis- conduct and dishonesty by some researchers reflect poorly on all, especially in the eyes of the government and the public. What, then, constitutes misconduct within scholarly research? A narrow inter- pretation defines research misconduct as intentional deceit and falsification of research, plagiarism, and misinterpretation or misrepresentation of results (Sten- eck, 1994). Broader definitions also include inappropriate collaborations, inappro- priate faculty-graduate student relations, denying knowledge of dishonest research practices by another, and conflicting interests with funding agencies. This much is uncontested: True misconduct must be distinguished from honest mistakes, shoddy work, and real disagreements about results or interpretations (Steneck, 1994). The outstanding problem here, of course, is determining the actual knowledge and intent of the researcher. Also uncontested (although perhaps less so) is that harms that are intrinsic to research should be distinguished from harms that are extrinsic (Warwick, 1982). For example, using deception as part of the research technique is intrinsic to the research process, and researchers are in control and should therefore be held morally responsible for the harms that are caused by the research. Extrinsic harms, on the other hand, are not part of the research process that is controlled by the researcher. If someone uses research findings for unethical aims that could not have been foreseen by the researcher, the researcher is not morally responsible. Determining the knowledge and intent of the researcher is once again problematic, because it is possible that the researcher could foresee negative and unethical uses for her or his research findings and yet continue the research in spite o f - - or, worse, because of--those reasons. In that type of case, researchers should not be able to evade moral responsibility for the uses to which their research is put. Certain research on race could fall into this category, such as using IQ data to establish certain races as inferior or superior. Donald Warwick (1982) proposes a taxonomy of harms attending research misconduct useful for broadening our discussion. He categorizes such harms into three main areas: (a)harms to research participants, (b)harms to society, and (c) harms to researchers and the research professions. We have already discussed ethical issues concerning research participants. Included under possible harms to society are the development of public cynicism and/or mistrust of academic research processes and the perpetuation of stereotypes or adverse images of certain societal groups. Increased public mistrust of scholarly work negatively affects the public perception of researchers and their activities, as well as the possibilities for funding and support of important research endeavors. One current example of the fallout from serious misconduct is that, in the United States after the Tuskegee deception, the federal government now needs to offer strong incentives for people of color to participate in federally sponsored health research
28 Review of Research in Education, 24
(Haworth, 1997). As for the harms to researchers and the research professions, Warwick cites the development of a deceptive, manipulative attitude toward others; increased restrictions on research activities; and lowered overall quality of research.
Plagiarism
Instances of plagiarism are perhaps the most common of all research miscon- duct, in any field. Plagiarism can take different forms: copying another research° er's work verbatim, which is the most blatant form; using intellectual property without the express permission of the owner of those ideas; or lifting substantial portions of another's work without any citation of that author. While it often may be obvious when someone actually copies the work of another, what makes plagiarism especially complicated to contend with is that it is often very difficult to locate the exact origins of ideas. Two prominent cases from the biomedical sciences, the Aisabti case and the Soman case, illustrate these issues. Elias Alsabti came to the United States in 1977 from Jordan to pursue postgrad- uate medical education. He was hired by a cancer research laboratory within Temple University's medical school, where he supposedly did cancer research. He ended up publishing more than 60 articles within 2 years of his arrival in the United States, some of which appeared in prestigious journals such as the Journal of Cancer Research and Clinical Oncology (Broad, 1980a). However, as he moved from one lab to another, his work became suspect, until he finally was accused publicly of severe plagiarism and of making up the names of various listed coauthors. For example, one article by Alsabti published in a European journal was found to have been copied almost word for word from a 2-year-old article in a Japanese journal. As the investigation of Alsabti continued, it was found that he never even had received a medical degree in Jordan (Broad, 1980a). This incident shook the world of medical research and publishing. People wondered how so many fraudulent articles could have slipped by the screening review systems. Apparently, even those who had noticed something fishy with Alsabti's work did not have him investigated. Rather, they just terminated him, which gave him the opportunity to move to other research laboratories and continue his plagiarism (Broad & Wade, 1982). Another prominent case occurred at around the same time. in 1979, a National Institutes of Health (NIH) medical researcher accused two Yale medical research- ers of plagiarizing a manuscript that she had submitted to the New England Journal of Medicine. She had been asked to review a paper submitted for publication by Philip Fefig, vice-chair of Yale's Department of Medicine, and his junior coauthor, Vijay Soman. The NIH researcher, Helena Rodbard, recognized the data and a portion of the writing as her own (Broad, 1980b). Concerned about priority of publication, she contacted the dean of the Yale University School of Medicine, who responded by asking the researchers whether they had conducted the study on which their paper was based. Felig and Soman said yes, and once he saw their data sheets, he considered the matter closed. His high respect for senior
30 Review of Research in Education, 24
politics. According to Rushton, Burt's findings oppose what he terms "genetic equalitarianism" (1994, p. 40). Rushton's own psychological research on cranial size differences by race and sex and their relation to IQ is also quite controversial. Burt's critics claim that he must have fabricated data because he could not have found so many cases of identical twins who were raised apart; he misrepresented the data so that they would fit with his predetermined theories (Joynson, 1994). So-called advocacy research, which places the researcher in the role of advocate for some view or another, is a frequent target of criticism. The ethical issues here are extremely tricky, and, most recently, Charles Murray and Richard Hermstein, authors of The Bell Curve (1994), have been accused of this type of unethical advocacy research (Strosnider, 1997). Some scholars believe that by using race as something more than a socially constructed label, researchers like Murray and Herrnstein play into a culture of racism (Anderson, 1992). Neil Gilbert (1994) has criticized advocacy research on any side of the political spectrum as eroding research standards and abusing statistics, especially in social research. What is perhaps most difficult to ascertain in many of these cases is the issue of intentionality. When there is clear evidence of data fabrication, as finally with Darsee, unethical conduct is clear. However, data fabrication is difficult to prove, especially when confidentiality issues arise. There are few checks on researchers within the research process. Data misrepresentation is even harder to prove. It is very tricky to distinguish for certain between willful misinterpretation of data and shoddy or incompetent research practice. In addition, the politics of certain types of research and research findings make ethical judgments very compli- cated indeed.
Pressures on Researchers What might compel some researchers to engage in unethical research behaviors, particularly when the potential harm to the research community and themselves is severe? Daryl Chubin (1985) identifies seven causes of research misconduct: "psychopathy, unbridled ambition, pressure for publication, competition for fed- eral support, the 'lab-chief' system, failings of the 'peer-review system,' and lack of replication and sheer sloppiness" (p. 177). Excluding psychopathy, these can be reclassified into four more general causes: researcher-researcher competi- tion, funding conflicts, publication pressure, and abuse of power. First, competition among researchers has stiffened immensely in recent years. This is due to, in part, the dearth of academic positions, especially tenure-track positions at top-notch research institutions. It is interesting to note that the most sensational cases of research misconduct have come from top-level institutions such as Yale University (Soman) and Harvard University (Darsee). Second, there is also a notable increase in competition for research funding, prompting conflicts of interest associated with increasing funding for higher education from private industry (Burgess, 1989). How much does a funding agency's agenda affect research practice and publication? A 1996 study of top biology and medicine journals found that in approximately one third of the articles
Howe and Moses: Ethics in Educational Research 31
sampled, the lead author(s) had some sort of conflict of interest. The conflict was usually a financial one, for example, holding investments in a company connected to the research in some way (Cho, 1997). Another recent investigation revealed that 98% of research studies that were funded by the pharmaceutical industry found new drug therapies to be more effective than the current drug, whereas 79% of studies that were not funded by the pharmaceutical industry found the new drug to be more effective. All of these studies were published in peer-reviewed biomedical journals (Cho, 1997). Third, the long-standing pressure to "publish or perish" within academia continues to put strain on young researchers. The top-level research institutions require numerous publications as a condition of tenure. Consider Alsabti, a junior researcher. When asked why he engaged in such misconduct and fraud, he blamed the pressure to move up within academia. He said that his "actions... were done in the midst of significant pressure to publish these data as fast as possible so as to obtain priority" (Broad, 1980b, p. 39). According to Alsabti, the cutthroat research atmosphere had compelled him into fraud. For Felig's part, although it was found that he had not been aware of Soman's unethical behavior, he made sure that, as a senior researcher, his name appeared on Soman's papers even when he was not involved in the research project. Certainly, an interest in adding to his list of publications played into Felig's ethical negligence. Finally, the various pressures just mentioned can also lead to abuse of power. One prominent example from the biomedical research community is the Baltimore case at the Massachusetts Institute of Technology (MIT). In 1988, David Balti- more, a Nobel-prize-winning biologist and director of a laboratory at MIT, was indirectly accused of data misrepresentation because the evidence presented in an article by Baltimore and five other colleagues did not support the conclusions drawn (Goodstein, 1991). The primary researcher and author of the article was Thereza Imanishi-Kari, the director of another laboratory at MIT. Baltimore's role was as senior scientist; it was Imanishi-Kari's lab team who did the primary work for this particular paper. Baltimore's lab team was collaborating with Imanishi-Kari's on a larger project from which the research in question came. After repeated denials from both Baltimore and Imanishi-Kari, the Office of Research Integrity's Commission on Research Integrity found that Imanishi-Kari had indeed falsified data to help support research findings that were published in the journal Cell. Much later, in 1996, Imanishi-Kari won her appeal to the Department of Health and Human Services' Integrity Adjudications Panel, when they decided that the Office of Research Integrity had never adequately proven its charges of intentional data falsification (Kevles, 1998). What makes this case fall under the abuse of power category is that it was two junior scholars, postdoctoral fellow Margot O'Toole and graduate student Charles Maplethorpe, who first questioned the Imanishi-Kari/Baltimore research. There were two other graduate students in the laboratory who also suspected data falsification, but when O'Toole brought the accusations, they refused to support her because they feared jeopardizing their degrees (LaFollette, 1994a).
Howe and Moses: Ethics in Educational Research 33
and operating principles. The former refers to the broad moral-political frame- works that undergird social and educational research; the latter involves the more specific principles used to govern and evaluate social and educational research vis-h-vis ethics.
Fundamental Perspectives The interpretive perspective jettisons the positivistic fact-value distinction and, along with it, the idea that social and educational researchers can confine them- selves to neutral descriptions and effective means toward "technical control" (Fay, 1975). Rather, value-laden descriptions and ends are always pertinent and always intertwined. Because each are part and parcel of social science research, the researcher has no way to avoid moral-political commitments by placing ethics and politics in one compartment and scientific merit in another. As stated by Maclntyre:
The social sciences are moral sciences. That is. not only dt~ social scientists explore a human universe centralize constituted by a variety or obediences to and breaches of, conformities to and rebellions against, a host of rules, taboos, ideals, and beliefs about goods, virtues, and vices... their own explorations ~/ that universe are no different in this respect from any other.[~Jrm ~!["httm,n ac'tivity. (1982, p. 175, italics added)
This general stance is one that a variety of contemporary perspectives converge on: In addition to communitarianism, care theory, and postmodernism, already mentioned, critical theory and contemporary liberalism may be added to the list. Nonetheless, these perspectives can diverge quite dramatically, and, although we will eventually draw them together again around several points of agreement, it is worth briefly describing where their differences lie. (We do not deny there may be other ways of distinguishing perspectives. For instance, feminism is perhaps conspicuous in its absence. But feminism cuts across the five perspectives. Furthermore, Noddings's care theory is one kind of feminist perspective we explicitly address.)
Communitarianism Communitarianism locates morality within a given community and its shared norms and "practices" (Maclntyre, 1981). Accordingly, what is conceived as the morally good life has to be known from the inside and varies from one community (or culture) to another. Because social and educational research cuts across communities that may differ from the social researcher's own, ensuring the ethical treatment of research participants who are members of such communities is doubly problematic. Not only are the normal problems involved in protecting autonomy potentially complicated by a lack of mutual understanding; a commit- ment to the fundamental values that undergird social research may not be shared. For example, certain communities do not place a high value on individual autonomy (the Amish perhaps being the most well-known case). As such, it is not up to individual community members to give their informed consent to have
34 Review of Research in Education, 24
social researchers peering into the social life of the community, for it is not always theirs to give. The community may reject the way of explaining and rendering community life transparent associated with social science and may not want its practices understood and portrayed in these terms. True, an individual community member who agreed to participate in developing such a portrayal might be viewed as a rogue who was wronging the community, but the social researcher could not avoid the charge that it was he or she who was the true instigator of such an "act of aggression" (MacIntyre, 1982, p. 179). The social researcher has no wholly neutral position from which to conduct research. "The danger" in believing otherwise, according to Maclntyre, "is that what is taken to be culturally neutral by the [social researcher] may be merely what his or her own culture takes to be culturally neutral" (1982, pp. 183-184). The ethical predicament for social and educational researchers raised here is close to the one historically raised under the anthropological concept of "cultural relativism." The difference is that it may now be recognized as a pervasive problem that applies to the broad range of "qualitative" social and educational research conducted across a broad range of cultural contexts and groups, not only exotic ones.
Care Theory
Care theory is a close cousin of communitarianism insofar as both emphasize concrete circumstances and specific demands on individuals ("the view from here") over ideal circumstances and the demands placed on individuals by abstract principles ("the view from nowhere") (Nagel, 1986). On the other hand, care theory embraces, if not a culturally neutral ideal, one that nonetheless is to be applied across cultural encounters; for Noddings (1984), caring is the ethical uni- versal. Noddings (1986) applies the ethics of care specifically to educational research. Her first thesis is that the relationship between researchers and participants ought to exemplify caring, particularly trust and mutual respect; her second thesis broadens the first so as to apply to the educational research enterprise as a whole. According to Noddings, the choice of research questions and the overall conduct of the research ought to be based on their potential to contribute to caring school communities. Educational research should not be conducted on the basis of mere intellectual curiosity; much less should it be conducted in a way that is likely to be harmful to individual students or groups of students or destructive of school communities. Educational research should be "for teaching," Noddings says, not simply "on teaching" (1986, p. 506). Ignoring these concerns renders the traditional emphasis on autonomy and privacy incomplete at best.
Postmodernism
Postmodernism shares the premise found in communitarianism and care theory that social and educational research cannot, first, isolate the descriptive component of social research from its moral component and, second, ensure the ethical
36 Review of Research in Education, 24
the form of testing/accountability regimens is left to the expert researchers. Finally, research-sanctioned testing regimens are then put in place with little or no input from those most affected: teachers, students, and parents. The means-ends bifurcation (a particular instance of the broader fact-value bifurcation) is open to at least three criticisms. First, means are relative to ends. Adopting the end of economic competitiveness ipso facto restricts the range of relevant means to those associated with achieving it. Accordingly, such means are laden with the end (read: value) of promoting economic competitiveness. Furthermore, whether something is a means or an end typically depends on its place in a longer chain of means and ends. For example, achievement in math and science is an end relative to instruction but a means relative to economic competitiveness. Second, means themselves are subject to value constraints, even relative to some end. If it could be shown that an effective means for improving economic competitiveness is putting all "at-risk" students in forced labor camps, we trust that no one would seriously entertain such a policy. Unfortunately, the general point this example illustrates often gets lost when dealing with less obvious examples of morally suspect means--talent tracking, for instance. Third, and most fundamentally, positivist technical control is irremediably undemocratic. Presupposing the ends of those with the power to formulate them, and then employing expert researchers to investigate the means to effect such ends, engenders technocracy rather than democracy. Genuine democracy requires that participation be respected as an end in itself. Social and educational research in service of democracy requires that no end or ends be settled on ahead of time, prior to and independent of the investigation of means. Rather, dialogue about both should be free, open, inclusive, and "undistorted," to use Habermasian language, by imbalances of power and by confining social science to the role of controlling social life. From the perspective of critical theory, an additional and fundamental role of social science is (should be) emancipation.
Contemporary Liberal Theory
Liberalism has been one of the primary targets of the four perspectives just described, particularly its utilitarian strand. Utilitarianism generalizes a certain conception of individual rationality (maximize benefits over harms) to the level of ethics and social policy (maximize benefits over harms in the aggregate) (e.g., Rawls, 1971). It is vulnerable at both levels. First, the view of rationality presupposed is not shared across groups and, worse, is not a particularly desirable one in any case. Second, extending the principle of utility maximization to the level of ethics and policy can result in wronging people in the way previously described by Maclntyre (1982). Finally, in the practice of social and educational research, the principle of utility maximization tends to work in tandem with the goal of technical control. Utilitarianism was the major strand of liberalism until the appearance of John Rawls's A Theory of Justice in 1971, which has since been the point of departure
Howe and Moses: Ethics in Educational Research 37
for liberal thinking. The kind of "liberal-egalitarian" view (e.g., Kymlicka, 1990) that Rawls formulated constrains the principle of maximizing utility in the name of justice. That is, not only aggregate utility is morally relevant. How utility (benefit) is distributed is paramount: Stated most generally, Rawls's principle of justice is that distributions (or redistributions) should tend toward equality. Although providing an advance over utilitarianism, Rawls's theory has nonethe- less been criticized for making several of the same general mistakes, including (a) presupposing a certain Western (and male) conception of rationality (i.e., maximize utility within constraints) and (b) conceiving of policy-making on the model of technical control (merely operating with different principles than utilitarianism). The most general difficulty is liberal-egalitarianism's commitment to the "dis- tributivist paradigm" (Young, 1990). The basic criticism is that liberalism defines and identifies the disadvantaged and then goes about the task of compensation. Compensation takes the form of various social welfare programs, including educational ones. Insofar as those targeted for compensation have been excluded from participation, what counts as rational and good is foisted upon them, and they are the pawns of technical control. And compensation, so called, can come at a cost. Consider a sexist curriculum in which girls fare poorly relative to boys. It is hardly a benefit to girls to compensate them so that they, too, can become sexist. Contemporary liberal-egalitarians have taken these difficulties seriously and have proffered remedies aimed at preserving the viability of liberalism. The general strategy is to tilt liberalism's emphasis on equality away from the distribu- tion of predetermined goods and toward participation in determining what those goods should be. As stated by Kymlicka:
It only makes sense to invite peopleto participatein politics (or for people to accept that invitation) if they are treated as equals.... And that is incompatiblewith defining people in terms of roles they did not shape or endorse. (1991, p. 89)
The "participatory paradigm" (Howe, 1995) exemplified in Kymlicka's admo- nition is much more attuned to the "interpretive turn" in social and educational research than the "distributivist paradigm." It fits with a model of research in which justice and equality are sought not only in the distribution of predetermined goods but also in the status and voice of research participants. The five perspectives we have portrayed differ in the ways we have indicated and, no doubt, in further ways we have not developed. We do not wish to deny that these differences can be deep, perhaps even irreconcilable. Still, there are several shared themes across these perspectives regarding the ethics of social and educational research. First, as we have indicated, there is a strong tendency in what we call the "traditional view" to distinguish the "descriptive" (scientific-methodological) component of social research from the "prescriptive" (moral-political) compo- nent. Each of the five alternative perspectives denies that social and educational
Howe and Moses: Ethics in Educational Research 39
on male participants, women were being held unjustly to a male standard of development. As the contemporary approach to research ethics has evolved in recent years, there have been cases similar to the Kohlberg-Gilligan studies that have involved bias not only against females but against people of color and the poor and working classes as well. In one such case within educational research, anthropologists John Ogbu and Maria Matute-Bianchi (1986) address the issue of school failure among students of color. By citing various sociocultural factors that contribute to less than optimal school environments for students of color, they confront theories that tend to place all responsibility (and blame) for such failure on the shoulders of the student of color. In doing so, they challenge a dominant view within educational research that, since school " w o r k s " for most White students, students of color must deserve the blame for their school disappointments. One more example is relevant here. Consider ethnographer Paul Willis's 1977 study of working-class male youth in England. Through his research, Willis attempts to shed light on what goes on in school for these young "lads." He discovers a culture of resistance against school knowledge and success that had developed among the young working-class lads. Through their rebellion, they were sabotaging their own chances at educational opportunities to get out of the working class, This finding challenged the more accepted theory that poor and working-class students were merely being manipulated by an education system intent on reproducing their social class roles. Through these examples from current research, we see that the influence of contemporary perspectives has caused researchers to examine things in new ways. This has led researchers to ask different questions and use different methods in finding the answers.
Operating Principles
The distinction between research ethics in the sense of operating principles and in the broader, fundamental sense is not hard and fast. What questions are worth asking and how researchers are to conduct themselves in the process of answering them cannot be divorced from the overarching alms that research seeks to achieve, one of the fundamental premises of the "contemporary approach." Nonetheless, there exists a "looseness of fit" between operating principles and competing perspectives, such that reasonable agreement on what constitutes ethical conduct is (or should be) possible in the face of broader theoretical disagreements. Bearing in mind, then, that broader ethical obligations associated with broader moral-political perspectives are always lurking in the background, there remain general ethical implications of the interpretive (qualitative) turn in educational research that may be best understood in terms of the methodological nitty-gritty of "techniques and procedures" (a description that owes to Smith & Heshusius, 1986).
40 Review of Research in Education, 24
Protection of Research Participants The techniques and procedures of interpretivist research possess two features that experimental and quasi-experimental research lack (at least to a relatively significant degree): intimacy and open-endedness (Howe & Dougherty, 1993; see Wax, 1982, for a similar analysis). The features of intimacy and open- endedness significantly complicate protecting participants' autonomy and privacy and complicate the researcher's moral life as well.
Interpretive (qualitative) research is intimate insofar as it reduces the distance
between researchers and participants in the conduct of social research. Indeed,
the growing preference for the term participants--who take an active role in
"constructing social meanings' ' - - o v e r subjects--who passively receive "treat-
ments' '--testifies to the changed conception of relationships among human beings engaged in social research that has attended the interpretive (qualitative) turn. The face-to-face interactions associated with the pervasive techniques of interviewing and participant observation are in stark contrast to the kind of interactions required to prepare "subjects" for a treatment.
Interpretive research is open-ended insofar as the questions and persons to
which interviewing and participant observation may lead can only be roughly determined at the outset. This, too, is in stark contrast to the relatively circum- scribed arena of questions and participants that characterizes experimental and quasi-experimental research. What intimacy and open-endedness mean for researchers employing qualitative techniques and procedures is that they are (whether they want or intend to or not) likely to discover secrets and lies as well as oppressive relationships. These discoveries may put research participants at risk in ways that they had not consented to and that the researcher had not anticipated. These discoveries may also put researchers in the position of having to decide whether they have an ethical responsibility to maintain the confidentiality of participants or to expose them, as well as having to decide whether to intervene in some way in oppressive relationships (see, e.g., Dennis, 1993; Roman, 1993). Researchers employing experimental and quasi-experimental techniques and procedures can face the same problems. For instance, information can simply fall into their laps in the process of explaining a protocol and recruiting participants, a treatment may prove so obviously effective (or harmful) that the trial should be stopped, and so forth. Still, the odds of facing unforeseen ethical problems are surely much higher for interpretive researchers. Generally speaking, then, interpretive research is more ethically uncharted, and thus more ethically hazard- ous, than experimental and quasi-experimental research. Once begun, it requires more vigilant ethical reflection and monitoring for that reason. Some interpretive researchers have recoiled at this suggestion, on the grounds that the current ways of thinking about and monitoring the ethics of social research are rooted in the experimentalist tradition and are therefore inappropriate for interpretivist research (e.g., Lincoln, 1990; Murphy & Johannsen, 1990; Wax,
