From: Graduate Research Ethics: Cases and Commentaries - Volume 7, 2006,
Edited by Brian Schrag
For the past few years, Sophia (a fourth-year graduate student) has been involved in bilingual research using a computational model that examines language acquisition. During the course of her research, however, she has become increasingly involved in issues concerning homeland security. Recently, this interest has led her to work with the NSA, FBI, and CIA to develop the model in a novel way for application as a text-screening tool. When fully developed, the model will go through text samples provided by these agencies and look for patterns of word usage that indicate characteristics such as deception or mental instability. For the current purpose of her dissertation, however, she will combine her present and past interests and use the model to detect key patterns of language usage that are indicative of language background (i.e., what languages the person speaks or has been exposed to). This research entails selecting specific language groups and comparing/contrasting patterns of their language usage using the output created by the model. Successful completion of this work is important because it will allow the agencies that use this tool to have more information about the ethnicity or nationality of the authors of the texts they receive (e.g., threats).
The nature of her research involves collecting large quantities of text samples from participants. As such, the first experiment for her dissertation will take approximately three hours to complete and requires subjects to write an essay/letter during the final two-hour session. The essay is to be written from one of three perspectives: a terrorist writing a bomb threat to the authorities, a kidnapper writing a ransom note, or a charity organizer requesting money from a large corporation (which will act as a neutral control). These topics were chosen to be representative of the kinds of texts received by the government agencies that will benefit most from this research. Furthermore, only Chinese (Mandarin or Cantonese)-English bilinguals, Spanish-English bilinguals, and English monolinguals are to be used as subjects (they are screened at sign-up), since they represent the predominant language groups on campus and subjects must be grouped according to a meaningful linguistic criterion that allows for as many subjects as possible to be placed in a group. Students who participate in this experiment receive three hours of credit—one credit for the first session, and two credits for the second—essentially fulfilling their introductory psychology course requirement.
Sophia is using a data analytic approach that has never been used before and so has little guidance from the literature concerning how the task might affect participants, especially after 9/11. Much research has shown that writing about certain events in one’s life is beneficial to the health of the writer in many ways. Two studies, however, suggest that writing about negative events may produce the opposite effect. No study that Sophia can find uses her particular negative role-playing strategy, so a direct comparison cannot be made. In addition, it will be obvious to subjects that only certain language groups are being used in this context (and that these groups represent different ethnicities, as well).
Understandably, the IRB showed concern for the usage of such controversial essay topics and required adequate justification for their use. After a year of review and much back-and-forth, Sophia convinced the board that her topics were necessary and that the experiment fulfilled an important need for theoretical and practical data in this field. She is finally allowed to begin her dissertation work with an important caveat. She must include in the informed consent a statement warning the subjects ahead of time that they may have to write a bomb threat or a ransom note during the second session of the experiment and that if this makes them uncomfortable, they can leave at any time. The experiment progresses until the end of the term without any problems. Then, a subject attends session 1 with a Middle Eastern language background, contrary to the explicit screening instructions given to undergraduate students at the time of signing up for the experiment. The experimenter tells him that she will give him partial credit for showing up, but that he cannot participate in or receive credit for the second session. The subject begs to be included and appears to be desperate for all three hours of experimental credit. When the request is refused, the subject begins to argue with Sophia and complains that the experiment targets particular races and is unethical. The argument becomes heated and the subject continues to pressure her to give him credit. In addition, his attitude becomes very sexist and his bullying makes her feel uncomfortable. He continues to claim that he should get credit because of the “unethical and racist undertones” of the experiment. Although concerned that the controversial nature of her experiment could cause alarm in her subjects, she does not think that this is the case in the present situation and that the student is simply manipulating her for credit.
At the end of the year, Sophia must reapply for IRB approval and report any indication that the experiment had a negative effect on a participant. Sophia must decide whether to mention the above incident and risk more restrictions/cancellation of her experiment or to not report the incident and assure her dissertation is completed on time and with relevant data.
Posted 3 years ago
This case highlights ethical issues in research design and effectively brings out a number of additional ethical issues. Temptations faced by social science researchers in dealing with these issues are forcefully presented. The case and other commentary make the social science researcher’s resistance to the IRB’s interference understandable, and at the same time they underline the importance of the role of the IRB.
The initial focus of this commentary is on the research itself, an investigation in the field of psycholinguistics. Sophia is engaged in bilingual research on language acquisition using a computational model. In every case, presumably, English is one of the languages. For her dissertation, Sophia will use the model to discern particular patterns of language usage that indicate what languages a person speaks or has been exposed to. This is the core scientific research. It is applied research, for the research product is designed for use in government agencies, such as the NSA, FBI, and CIA (perhaps also Homeland Security) to help them gain information about the ethnicity or nationality of the authors from texts they receive. Of particular interest are texts containing bomb threats, ransom notes, and the like.
The research involves selecting many participants from specific language groups and comparing patterns of their language usage on the basis of output from the model. It is essential to collect a large number of text samples from research participants. In addition, to obtain results that will be useful to the agencies for ascertaining ethnicity and nationality of authors, Sophia has determined that she needs text samples of bomb threats and ransom notes from participants.
Presumably, Sophia’s dissertation advisor(s) agreed on the need for such texts. This is an important point because the case raises an ethical question about the justification for an experiment that requires students to furnish such texts, as well as a question about research design. Accordingly, before turning to the risks for students from participating in the experiment, it is important to ascertain the scientific rationale for requiring students to write bomb threats and ransom notes and for allotting such a large portion of the three-hour experiment to obtaining these texts.
Another question of research rationale focuses on the justification for apparently accepting the agencies’ interest in determining the ethnicity and nationality of authors. Why are these the target variables? What about other variables of interest to psychologists, such as mental instability? If such psychological characteristics are the ultimate target, why are nationality and ethnicity thought to be appropriate foci of investigation? One is reminded of a point often made concerning the apprehension of terrorists. Using ethnicity or nationality profiles rather than certain behavioral profiles is a time-consuming distraction from tracking people exhibiting odd behavior that might indicate questionable clandestine activity. Accordingly, academic psychologists could serve the applied research users by questioning and discussing with them their interest in the ethnicity and nationality of the authors of texts.
These questions of justification need to be discussed with participants, for the rationale points to the value of the research, i.e. the basis for justifying risks for which participants’ consent is sought. In addition, students are to be awarded academic credit for participating in the research; it is important that they have an intellectual understanding of the research and the reasons for pursuing it. The credit issue will be explored more fully below after comments on Sophia’s stance regarding risks.
The claim that there is little guidance in the literature regarding the risks to participants from the tasks of writing bomb threats and ransom notes seems open to question. Sophia may not find guidance related just to these tasks, but by using her imagination to identify relevantly similar tasks and discussing this issue with her advisor(s), she may uncover some helpful literature. Such research is an appropriate component of the dissertation research, even necessary to defend experiments such as the one Sophia intends to conduct.
If, contra expectations, Sophia finds no guidance, she must proceed with great caution, given that there is evidence that writing about certain positive events may have positive effects. In Sophia’s experiment, students are to produce texts that in certain contexts would expose them to legal consequences. The absence of guidance is not license simply to proceed.
Presumably, students are to take the tasks seriously, not play games with them. Is it possible to tell the difference? This leads to further questions about whether such tasks are necessary or even useful, particularly in view of the goal to obtain from the texts information about ethnicity and nationality of authors. Moreover, as the author of the case notes, the restriction of research participants to certain language groups might well raise questions in the minds of participants (and others) about profiling. In this connection, there may be relevant studies that indicate risks, and negative effects may not be so difficult to predict. One would not expect a decline in health among negative effects. More likely effects are anxiety, irritation, resentment, and even fear. The researcher has an obligation to learn about likely effects.
The research subjects are entitled to know that this is applied research and which government agencies are the intended users. Respect for students as persons requires that they have an opportunity to chose whether to participate in this particular applied research and to withdraw from participation at any time because of objections to the application or for any other reason. This respect is also owed to students as participants in what is presented as an educational activity. Students are presumably allowed credit because this is intended to be an educational experience.
Students are more vulnerable than some other populations because of their status. They are subject to the power and influence of their instructors, and they need credit, grades, letters of recommendation and other sorts of consideration that affect their careers in college and in life afterward.
Three hours of course credit for participation in a three-hour experiment of writing texts seems excessively generous, especially if one accepts the ground rule that students should not perceive participating as subjects in research as an easier way to obtain extra credit than other available options. (There should, of course, be other options.) Such generous terms can rank as undue inducement, even coercion in some circumstances. This is a concern that the IRB should not fail to address. It is incumbent on IRBs (and psychology departments as well) to establish and enforce policies, such as the ground rule suggested above, out of respect for students and the need to protect them. The course credit should not count in the balance against the risks of the research.
It is altogether appropriate that IRB approval required a year’s time, considerable discussion, and the caveat that students be on notice through the consent form that they might be required to write a bomb threat or ransom note. The IRB justifiably pressed Sophia for justification. It is surprising that there is no detail on the considerations that finally persuaded the IRB. For example, how did the IRB come to agree with the distribution of time, two thirds to be spent on the bomb threat or ransom note?
The incident that occurred at the end of the term is related to two ethical issues already identified: the generous course credit awarded for participation and the apparent profiling of certain language groups. The Middle Eastern student’s perception is that he was discriminated against in not being allowed into an experiment that would have answered his credit needs. This is the sort of outcome Sophia might have thought about in advance. She might, as a result, have decided upon less seductive credit terms. It is also conceivable that if she had questioned the agencies’ presumed interest in the ethnicity and nationality of text authors, she might not have gone ahead with a research protocol that visibly targeted certain language groups.
Sophia is not entitled to conclude that the student is simply manipulating her for credit. Our motives most often are mixed. Sophia’s concession to give him partial credit may not have been well considered, for he did not have the qualifications of a proper participant. He might have concluded that whatever consideration had inclined her to give him some credit might work to extract more credit. The narrative about the student suggests that he might have been genuinely aggrieved (not “alarmed”, as Sophia opines) at being excluded on what he plausibly perceives to be “racist” grounds. After all, Sophia had foreseen that there might be a problem with the appearance of profiling. She thought of the effect on qualified participants rather than on those excluded.
Sophia must report this incident to the IRB when she reapplies for approval. This is knowledge the IRB must have in deciding whether and with what modifications or caveats to grant approval. The incident was intense and related to ethical questions raised by the research. The researcher is at risk in such situations. Sophia cannot be confident that such incidents are not harmful to students and will not occur again.
There are no general, bright line boundaries defining what in the first year of experience must be reported to the IRB. It is helpful to think about what information the IRB needs to make sound decisions for the present and the future to protect research participants. It seems obvious that since Sophia has become involved in a quite heated incident with a student, she is not in a good position to gauge the motives and level of stress of the student. Regarding all three questions following Part 1, the experimenter should discuss her responses with advisors, department members, and perhaps others. In general, researchers should test out their views on ethical questions concerning their research with others whose reactions may be useful to confront.
It is no longer a question whether an IRB should have authority to intrude into research. Since the appearance of the Belmont Report, we have well established knowledge demonstrating that researchers cannot be left to determine for themselves whether they have adequately dealt with risks posed by their research. The consent form is very useful for pressing researchers to consider and address ethical questions raised by their research, and it is essential for providing appropriate protections to research subjects. The IRB should be the final authority to be satisfied on these points.
The other commentary raises the question of whether fully informing research participants might affect the results of the experiment and the benefit of the research in respect to homeland security. It might affect the results, but that prospect is not enough to overbalance the importance of fully informing participants. Indeed, the effect might not be for the worse. The benefit to homeland security is in the realm of speculation.
The argument that fully informing participants jeopardizes the research has long since lost its hold. Openness and transparency are as important in research as in other areas of activity, and restrictions require strong justification. The demand for openness should trigger ingenuity in designing the research. Government funded research that bars fully informing research participants should not be conducted with students. Whether and under what conditions such research can be justified with other participants is beyond the scope of this discussion.
Although this case involves a specific experiment in psycholinguistic research, several general ethical questions are addressed that can be applied to work outside of the area, including risk assessment, formation of the informed consent, subject selection, credit/participation, and reporting to the Institutional Review Board (IRB). To aid in this treatment, the American Psychological Association (APA)’s Ethical Principles of Psychologists and Code of Conduct will be consulted as the ethical standard of this field. APA guidelines consist of five overarching principles that are meant to be general and aspirational coupled with ethical standards that are meant to address specific incidences that may arise in the course of psychological research. This commentary will address Part 1 and Part 2 of the case study in turn.
Part 1 introduces the experimental situation and raises background issues that may arise with research of this sort. The following are several themes that can be elaborated upon in discussion. Underlying these themes is a more general moral tension that runs throughout this case concerning the obligations of a researcher to science and their obvious need to protect the rights of their subjects.
The APA’s ethical standard 3.04 (Avoiding Harm) states that researchers must “take reasonable steps to avoid harming their clients/patients, students, supervisees, research participants, organizational clients, and others with whom they work, and to minimize harm where it is foreseeable and unavoidable.” This scenario, however, is meant to provoke thought on this standard when the appropriate level of safeguarding in an experimental situation is less than obvious. In Sophia’s case, the literature provides no guidelines for use of a negative role-playing task. Research suggests that writing about negative events may be harmful, but it is not clear that that is what subjects are doing. How should researchers assess such situations? What are the “reasonable steps” that could be taken to minimize harm if the experiment should be allowed to proceed?
Also underlying this dilemma is the role of the experimenter in making these judgment calls and in deciding whether the benefit of the research outweighs its potential risk. Question 1 challenges this role. Further discussion can center on the position of the IRB versus the professional responsibilities that are placed upon members of academia. For example, APA explains that their guidelines were purposely written in such a way as to allow professional judgment on the part of psychologists (stated in introduction). How does this judgment come into play when ethical dilemmas arise? When should potential biases be protected against? In other words, how much responsibility should be given solely to the investigator rather than to a governing board such as the IRB?
An important concern in this study lies in the formation of the informed consent. This is raised early on in the fact that the data are intended for development of government technology. It is plausible that some subjects would not want to participate in such an endeavor. An obvious course of action would be to include this information in the informed consent. However, this may change the results substantially and affect the benefit such research has on homeland security. Other issues may be brought up in discussion that stem from this problem. For example, what if the scenario is slightly changed such that the experiment is being funded by these agencies and they put this information under security clearance? Does this change the moral obligations of the researcher from that of subject to country? Should the experiment not be run if subjects cannot know the use of their data? Would it be enough to let subjects know of this restriction?
An additional issue concerning informed consent formation raised in this case lies again in the potential risks students face from participation in this task and how much information concerning this should be divulged. This is a classic ethical dilemma when conducting research (applicable also to the previous issue). On one hand subjects have the right to know what they are agreeing to do. APA ethical standards dictate that researchers must inform participants of any “reasonably foreseeable factors that may be expected to influence their willingness to participate such as potential risks, discomfort, or adverse effects” (ethical standard 8.02a). However, if the task is divulged the experiment may be jeopardized. The argument from the literature for a potential risk is not very strong, but does this matter? Where is the line and who decides this? When does it become necessary to include hypothetical problems in an informed consent?
A third general issue addressed in this case deals with subject selection and recruitment. Question 3 raises issues concerning screening and use of students as subjects. Use of language groups, though seemingly innocent, sometimes involves separation of ethnic groups (in this case: Hispanic, Asian, and Caucasian). Combined with the essay topics (terrorism and crime), this may cause discomfort in participants just by its implications. What are the ethical responsibilities of a researcher in this situation? Additionally, the vulnerability of students as subjects can also be addressed in discussion at this point. Should they be treated with more care than other sampling populations?
This scenario also touches on the use of incentives. Having the experiment fulfill all of the student’s course requirements induces students to want to participate (leading to problems like those seen in Part 2). APA’s recommendation for use of inducements seems inappropriate for this situation in stating that psychologists must “make reasonable efforts to avoid offering excessive or inappropriate financial or other inducements for research participation when such inducements are likely to coerce participation” (ethical standard 8.06a). The current incentive is not excessive. But, is it coercive? An interesting point for discussion centers around the potential distinction between personal ethical choices and principles laid out by an institution. Is simply following standard guidelines enough? What if these guidelines do not specifically address the moral issue in question?
Part 2 is concerned not so much with experiment preparation as in Part 1, but with issues that may arise during the experimental situation. More specifically, the problems Sophia faces concern subject credit/participation and reporting to the IRB.
Sophia, in managing the concerns and behavior of the participant, chooses to refuse him full credit and participation in her study. Was this the correct solution? APA guidelines state that when the concerns of the researcher are in conflict, they must “attempt to resolve these conflicts in a responsible fashion that avoids or minimizes harm” (Principle A). Is this what occurred? Did Sophia let her personal annoyance get in the way of resolving the situation peacefully? Challenge students in discussion to come up with alternative courses of action along with the pros and cons of each.
Several concerns arise when considering the role of the IRB in this case study. The crux of the dilemma lies in whether or not Sophia should report the incident with the offending participant. Doing so would jeopardize her research and its use, yet provide a safeguard against potential future harm of participants as well as provide a second opinion on a judgment that is potentially biased. This raises two topics for discussion. First, how much information needs to be given to the IRB? What qualifies as a harmful situation? Second, should researchers rely on their own subjective judgment? What about experimenter bias? Are there ever situations where experimenters can rely on their own judgment calls? In the discussion, it might be interesting to highlight the conflict between thorough reporting and wasting the IRB’s (usually taxed) resources.
The IRB is not only in place to protect the participants, but the experimenters as well. A second thread of discussion—not often addressed—concerns the potential harm that researchers face in some experimental situations. Sophia was bullied by the male participant and sexually harassed. Is this something she should report to the IRB as well? Is the task designed in such a way that these situations may reasonably arise in the future? Should the experiment be re-evaluated for her safety as well? Should she make this decision or allow the IRB to decide?
American Psychological Association. 2002. Ethical Principles of Psychologists and Code of Conduct.