Chapter Summary
This chapter surveys ethical matters and dilemmas that may be faced by researchers during fieldwork. Questions are posed to consider issues such as participant consent, benefits, risk, confidentiality, and data ownership.
Contents
Introduction
Agreements With Study Participants Ethical Issues
Worthiness of the Project Competence
Informed Consent
Benefits, Costs, and Reciprocity Harm and Risk
Honesty and Trust
Privacy, Confidentiality, and Anonymity Intervention and Advocacy
Research Integrity and Quality Ownership of Data and Conclusions Use and Misuse of Results
Conflicts, Dilemmas, and Trade-Offs Closure and Transition
Introduction
We cannot focus only on the quality of the knowledge we are producing, as if its truth were all that counts. We must also consider the potential “wrongness” of our actions as qualitative researchers in relation to the people whose lives we are studying, to our colleagues, and to those who sponsor our work. All researchers must be guided by the classic principle of humane conduct: First, do no harm.
Any responsible qualitative researcher ponders moral and ethical questions such as the following:
Is my project really worth doing? Am I exploiting people with my research? Do the respondents have a right to see my report? What good is anonymity if people and their colleagues can easily recognize themselves in my report? When they do, might it hurt or damage them in some way? What do I do if I observe harmful actions during my fieldwork? Who owns the data, and who owns the report?
The qualitative literature is full of rueful testimony on such questions, peppered with sentences beginning with “I never expected . . .” and “If I had only known that . . .” and “I only later realized that . . .” We need to attend to the ethics of what we are doing before, during, and after a study.
We cannot begin to deal with all ethical issues, but we do hope to cut across some of the problems involved and raise as much ethical consciousness as we can. In this chapter, we describe agreements with participants and specific issues that often arise in qualitative research, exploring their implications for analysis. We then examine some of the conflicts, dilemmas, and trade-offs involved and conclude with some general advice.
Agreements With Study Participants
Usually, study participants and researchers need to reach some explicit agreements about shared expectations. Our intent here is to examine the analysis-related aspects of agreements made with those whose daily lives are being examined. The main issue is what explicit expectations we want to build with study participants that will maintain and improve the quality of our conclusions.
We might think first of a “meta-agreement”: Are we contemplating an equal-status, polyvocal model, in which researchers and participants are jointly telling their stories? Or are we heading for a collaborative or participatory action research model, in which researchers join forces with people facing a problem in their community to help them study and resolve it? Or will it be a more traditional model that differentiates researchers from participants, on whom the research is conducted? The first two models imply more shared control over the design and conduct of the project than does the last.
Whatever the basic relationship implied in such a meta-agreement, some matters need to be clarified with participants at the outset. (It’s important to remember that these matters may not be fully understood on both sides until the study unfolds.) Display 3.1 presents issues that might be considered as the elements of an initial set of expectations. Many different agreements can be reached. We comment on the implications of some for doing analysis.
Data collection agreements (see Question 2 in Figure 3.1) that include active involvement of participants, such as journal writing, are threatened when such participation is coerced, even gently (see Question 3). Such agreements also tend to move the study in the direction of shared study design and steering (Question 4).
Vagueness about confidentiality (Question 5), as when a researcher voluntarily or involuntarily passes on a participant’s comments to another, often has “drying up” or distorting effects on subsequent data collection; relationships may get strained, and subsequent analyses may be biased.
The same goes for anonymity (Question 6). An individual, group, or organization not assured in advance of nonidentifiability in study reports may provide biased data (self-censored, defensive, and rosy) if it is believed that an accurate, identifiable account would jeopardize some interest. In any case, anonymity of individuals is difficult or impossible to assure when a case study of a group or organization is read by its members. (An alternative agreement is to use real names from the outset;
under these circumstances, individuals only provide information they regard as public or nondamaging. This, too, has its conclusion-narrowing aspects.)
Display 3.1
Questions for Agreement With Study Participants 1. How much time and effort will be involved?
2. What kind of data collection is involved (e.g., observation, interviewing, journal writing, life histories)?
3. Is participation voluntary?
4. Who will design and steer the study?
5. Will material from participants be treated confidentially?
6. Will participants’ anonymity be maintained?
7. Who will produce descriptive and explanatory products?
8. Will participants review and critique interim and final products?
9. What benefits will accrue to both participants and researchers?
Source: Miles, M. B., & Huberman, A. M. (1994). Qualitative data analysis: An expanded sourcebook (2nd ed.). Thousand Oaks, CA: Sage Publications.
A typical agreement is that researchers will produce the products of the study (Question 7). This agreement rests on the traditional assumption that well-prepared researchers will get good data and will draw well-founded conclusions. In some models, however, expertise resides in participants as much as in researchers. The collaborative or participatory action research model implies that participant expertise is developed through the researcher’s facilitation during the process—for example, in community-based projects. In either case, issues of “goodness of analysis” are just as pressing as in the traditional model (see Chapter 11).
Study products are sometimes fed back to participants (Question 8) as a way of providing member checks on the accuracy of descriptions, explanations, and interpretations. Agreements can vary: Does an individual see material about his or her case before others do? Can an individual or group censor or veto material, or even block publication? Or is the agreement only that errors of fact will be corrected and alternative interpretations will be included in footnotes? Such agreements can improve the quality of both the data and the final conclusions, but they also can result in truncated or distorted conclusions if someone has been given, and exercises, the right of censorship.
Researchers usually benefit (Question 9) from their studies through insights, recognition, promotion, new grants, and consulting. That is why they keep on researching. Participants’ benefits are often posed quite vaguely at the start: “the chance to reflect,” “clarifying ideas,” “learning what others are doing.” Advance agreements that mention assistance (as in collaborative research), consultation or training, joint authorship, or shared royalties as expected benefits may improve the quality of the data and the conclusions. If the benefits do not materialize, data and conclusion quality may suffer.
We offer these final pieces of advice on agreements with study participants:
1. Be clear in your mind what you want your agreement with participants to be like. Commit it to paper as a vehicle for discussion with them while negotiating entry and access. Once the agreement is clear, a copy of a summary of the study and its ground rules is helpful for a study that has multiple participants.
2. Incorporate in data collection plans an explicit procedure for logging participants’
understanding (or misunderstanding) of agreements, including any threats to data or conclusion quality that you see.
3. Researchers held to their institutional review board (IRB) regulations may have specific requirements imposed on them that affect researcher–participant relationships, agreements, and data collection protocols.
Ethical Issues
Most professions have well-defined codes of ethics, which include guidelines ranging from research participants’ rights to inappropriate forms of researcher–participant relationships.
Organizations subject to an IRB or comparable overseeing agency also mandate guidelines for securing permissions, maintaining confidentiality, working with minors, and other legal matters. But fieldwork and its accompanying dilemmas are often quite unpredictable and site specific. Plus, research questions, instrumentation, and sampling all may evolve over the course of a qualitative study, making the IRB’s requirements for these details to be precisely specified before a study begins nearly impossible to comply with. Nevertheless, some planning and forethought are necessary for application and review.
We now outline some specific ethical matters and issues for reflection.
Worthiness of the Project
Ask yourself: Is my contemplated study worth doing? Will it contribute in some significant way to a domain broader than my funding, my publication opportunities, and my career? And is it congruent with values important to me?
Implications for Analysis
In general, a study that is only opportunistic, without a larger significance or real meaning to you, is likely to be pursued in a shallow way, with less care devoted to design, data collection, and analysis. The report will be written to “look good” rather than to be right. Initial conclusions may not be questioned; follow-up analyses with rival hypotheses may be rare. There’s nothing wrong with taking on a commissioned research project for personal income. But if you slant your findings to accommodate and please your funders, you become complicit and dishonest.
Competence
Ask yourself: Do I (and my colleagues) have the expertise to carry out a study of good quality? Or (because researchers, both novice and experienced ones, are always exploring things they do not quite know how to do) am I prepared to study and be supervised, trained, and consulted with? Is such help available?
Implications for Analysis
Unacknowledged (or not understood) incompetence is, we think, responsible for a certain pattern of analytic weakness in qualitative studies: accumulation of large amounts of poorly collected, unanalyzed data and drawing superficial and hasty conclusions as deadlines loom. This picture often occurs when lone researchers fail to seek help from friends, colleagues, or mentors. Graduate students often understand their own inexperience but sometimes cannot get support and help from their teachers or dissertation supervisors. That is research malpractice.
Informed Consent
Ask yourself: Do the people I am studying have full information about what the study will involve?
Is their consent to participate freely given—fully voluntary and uncoerced? Does a hierarchy of consent (e.g., children, parents, teachers, administrators) affect such decisions?
IRB regulations are strict but thorough when it comes to recruiting voluntary participants, informing them about the study’s goals, and assuring them of their rights throughout the project.
Research with minors or other vulnerable populations (e.g., prisoners, pregnant women) includes additional guidelines for permissions processes, such as securing parental consent and children’s assent.
Be an open book with your project; this will help develop a sense of trust between you and the setting’s individuals. Also, acknowledge that participants are doing the researcher a great favor by volunteering to be part of the study. Respect their gifts of time, insight, and privacy, for the root meaning of “data” is not something that is collected but something that is given. If children and adolescents are part of the mix, they too have a say in the matter. There are only extremely rare occasions when covert research and deceit are necessary for the integrity of a research project’s goals (e.g., investigating the social processes of humans engaged in illegal activity). If you have something to hide from participants, it better be for a good reason.
Implications for Analysis
Weak consent usually leads to poor data. Respondents will try to protect themselves in a mistrusted relationship or one formed with the researcher by superiors only. Ambiguity about the later stages of analysis also can be damaging to study quality and to the interests of people in the case. If you plan to use “member checks” to verify or deepen conclusions, that expectation and specific procedures need to be clear as the study proceeds. Securing a participant’s permission is not a single hurdle to be jumped; dialogue and ongoing renegotiation are needed throughout the study.
Benefits, Costs, and Reciprocity
Ask yourself: What do participants have to invest in time, energy, or financial resources? What will each party to the study gain from having taken part? Is the balance equitable?
As we suggested earlier, researchers are often “paid” in one way or another. They usually enjoy their work and learn from it, they may get a dissertation out of it, and their papers, articles, and books not only contribute to their fields but also can bring them recognition, royalties, new funding, and career advancement.
Study participants have a somewhat different set of benefits: They get to be listened to, they may gain insight or learning, they may improve their personal practice, a program or policy they are involved with may be strengthened, and they may get help in taking effective action on some recurring problem. But study participants rarely share in publication, and they usually don’t become famous. Most don’t get paid for their research contributions.
The question of costs and who bears them is important. The researcher’s time is repaid—usually not fully—in cash or a class grade or dissertation approval. Research participants normally must take time from or beyond whatever else they are doing and are usually not recompensed. The local organization may well have added costs (e.g., for teacher substitutes).
Implications for Analysis
Study participants’ concern about the inequity of benefits and costs serves to jeopardize access and thin out data. But we are always moved and amazed when people do keep talking to us thoughtfully, inviting us into their lives day after day, when the benefits to them seem so slim, so intangible, and often so delayed.
Researchers vary in their perspectives about monetarily compensating their participants (e.g., offering a gift card or a token $20.00 for each hour of interview time). Some feel that the offer of money may unduly influence participants’ responses to become more favorable and positive than they really think. Other researchers feel that working-class professionals, such as teachers and social workers, deserve some type of financial compensation for their valuable time and opinions. This is an issue to be resolved by each individual researcher and his or her particular project.
Researchers traffic in understanding. Most study participants are preoccupied with action—how to work and live better. It can be argued that if you approach your analytic work with a deeper sense of its action implications, your understanding will be deeper—and the benefits to participants more equitable.
Harm and Risk
Ask yourself: What might this study do to hurt the people involved? How likely is it that such harm will occur?
Harm to participants can come in many varieties: from blows to self-esteem or “looking bad” to others, to threats to one’s interests, position, or advancement in the organization, to loss of funding for a program, and so on, up to being sued or arrested. The information from a qualitative study is never value-free, and it may have negative consequences.
Harm cuts both ways. We like the story told by a New York Times reporter who asked a drug dealer if he really felt comfortable about talking frankly. The dealer said cheerfully, “Sure. If I don’t like what you write, I’ll kill you.” As researchers, we have occasionally been threatened with litigation and with warnings to intervene with our funding agency when a draft report was seen as threatening a key interest.
Sieber (1992) points out that it’s important to think of varying vulnerability to harm. More vulnerable persons (and institutions) include those who are publically visible, lack resources or autonomy, are stigmatized, are weakened or institutionalized, cannot speak for themselves, are
involved in illegal acts, or are too closely associated with those studied.
Setting risk levels for potential harm is very difficult—perhaps impossible—in qualitative studies.
It’s wise to assume that the chances of some type of harm are better than others and to consider in advance ways of reducing that likelihood.
Implications for Analysis
As with inequitable benefits and costs, if harm is expected, access and data quality may suffer.
The prospect of immediately impending harm—which well may occur when reports are made to local participants, sponsors, or funding agencies—can lead to pressure on you to revise or delete conclusions or to self-censor them in advance.
Honesty and Trust
Ask yourself: What’s my relationship with the people I’m studying? Am I telling the truth? Do we trust each other?
Most qualitative researchers are unlikely to lie, cheat, or steal in the course of their work. But broken promises are not unknown. And some researchers have reported deceiving respondents about the true nature of the inquiry (as in some participant observation studies such as Humphreys’ [1970]
study of homosexuals, where he posed as a “watch queen” lookout inside public toilets).
Typically, dishonesty is more subtle. The field-worker may project a “fake persona” (the friendly listener or the would-be insider) to gain knowledge or access. At some levels, as van Maanen (1979) notes, whenever the field-worker works to “penetrate fronts,” symbolic violence is being done: “People are, to a degree, coaxed, persuaded, pushed, pressured, and sometimes almost blackmailed into providing information to the researcher that they might otherwise prefer to shield”
(p. 545). There are always individually drawn moral limits to this violence: A researcher may decide not to push on a delicate matter, or to leave an embarrassing scene. Nevertheless, the question of just how coercive and unauthentic relationships with respondents are cannot be ignored, or be defined away by the pious stipulation that “my relationship is fully honest.”
Implications for Analysis
If people feel betrayed by you when they read a report, it becomes almost impossible for them to accept it as a reasonable interpretation of what happened, because of their natural defensiveness when “the truth hurts,” as it well may, and their feelings of anger at having been misled.
Deceptiveness and broken promises—especially if benefits and costs have been inequitable or harm has occurred—will make any continuation of the inquiry problematic. We will have wronged not only our respondents but also our colleagues.
We will not pretend to have any easy or foolproof answers for this slippery category of ethical dilemmas. All we can pass along is some classic advice: When in doubt, tell the truth.
Privacy, Confidentiality, and Anonymity
Ask yourself: In what ways will the study intrude, come closer to people than they want? How will information be guarded? How identifiable are the individuals and organizations studied?
Sieber (1992) makes these helpful distinctions among three terms, which often are confused or used interchangeably in research practice:
• Privacy: control over others’ access to oneself and associated information or preservation of boundaries against giving protected information or receiving unwanted information
• Confidentiality: agreements with a person or organization about what will be done (and may not be done) with their data—may include legal constraints
• Anonymity: lack of identifiers, information that would indicate which individuals or