8.2. The Belmont Principles: Respect for Persons, Beneficence, and Justice

In the United States, the foundational document guiding government regulations and professional norms regarding research ethics is the Belmont Report, which grew out of the federal government’s efforts to establish a regulatory framework for research in the 1970s. As part of the 1974 National Research Act, Congress required all institutions receiving federal support to establish IRBs to oversee research projects and protect the rights of human subjects. The legislation also launched the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, which in 1979 produced the Belmont Report—named after the Maryland manor where the commission met. Inspired by the Nuremberg Code, the Belmont Report put forward three overarching ethical principles for research on human subjects: respect for persons, beneficence, and justice (National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research 1979). Let’s break down the implications of these three principles (you can also read the full text of the report online):

1. Respect for persons: Individuals should be treated as autonomous agents able to make their own decisions. This principle implies informed consent—honoring your research participants’ autonomy means that you should tell them about what you are studying, how you are studying it, and what risks and benefits their participation entails. Showing respect for participants also means you should safeguard their privacy, a topic we discuss further in regards to protection of identities. Finally, the principle of respect for persons applies to groups with diminished autonomy—such as children and those with disabilities—for whom scientists must show special care when involving them in research.

2. Beneficence: Although the word “beneficence” is typically associated with unconditional acts of kindness and charity, in a research context, it is a moral imperative, the Belmont Report argues. Beneficence in this sense involves two principles: (a) “do not harm” (in line with the no harm to participants principle we have previously discussed) and (b) “maximize possible benefits and minimize possible harms.” When researchers act with beneficence, they respect the decisions of their human subjects and secure the well-being of these individuals.

3. Justice: This principle focuses on who should receive the benefits of research and bear its burdens. An injustice occurs when some benefit to which a person is entitled is denied without good reason, or when some burden is imposed unduly or arbitrarily. The tenet of justice reinforces the need for informed consent: participants must receive clear information not only about the purposes of the study but also its risks and rewards.

We’ve already discussed the principle of no harm to participants that is explicit in the Belmont’s beneficence tenet and was egregiously lacking in many of the biomedical and social scientific studies we’ve described thus far. In this section, we elaborate on the key concepts of informed consent and protection of identities that follow from the overarching principles in the Belmont Report.

Informed Consent

What does it mean to treat our human subjects as autonomous agents—that is, individuals who are capable of making decisions for themselves? It means that we seek their voluntary participation based on informed consent. In any legitimate scientific study—certainly any study an IRB is willing to approve—we must have formal procedures in place to ensure that all human subjects involved in the research freely choose to participate. Their consent should also be “informed,” meaning that they should be given understandable and accurate information about the study, not some vague or misleading statements about it.

These formal disclosures are usually included in a consent form that researchers provide to subjects before the start of the study. In some cases, subjects are asked to sign the consent form indicating that they have read it and fully understand its contents. In other cases, subjects are simply provided a copy of the consent form, and researchers are responsible for making sure that subjects have read and comprehend its terms before starting any data collection. In either case, researchers should go over the form with their subjects at length—highlighting important sections of it, giving them ample time to read and review it, patiently addressing any questions or concerns they have about it, and providing them with a copy of the form to take home.

What sort of information should be laid out in a consent form and discussed during the consent process? First, researchers should describe what the general aims and goals of the study are, who is conducting and sponsoring the research, and how the research will be conducted—including what participants will do, and for how long. They should clearly communicate the potential risks associated with it, which obviously would include any psychological or safety risks but also any dangers to privacy or reputation. The informed consent process should also cover the possible benefits of participation—anything from monetary compensation to the opportunity for participants to share their personal views on an important issue. Finally, researchers must describe how they will protect the identities of their participants (covered in the next section) and whom they should contact for additional information about the study or their rights as study participants. This informed consent process should happen before data collection begins so that the knowledge can help participants decide whether or not they wish to be involved.

There are important practical implications of this general approach of informed consent:

• We cannot force anyone to participate in our research.
• We cannot include a person in our study without their knowledge or consent.
• We need to provide clear and accessible information to potential participants about the study they are being asked to be involved in.
• We cannot ask potential participants as a condition of their participation to not pursue legal recourse if the study somehow ends up harming them.

These last two points bear further discussion. In providing information about the study to potential participants, we need to describe the study in layperson’s terms—not with dense scientific jargon—and we need to make their rights and any potential risks clear and concrete to them, rather than concealing them in ambiguous language or legalese. We also need to make sure that every potential participant fully understands what is being asked of them. This is an especially important consideration for more vulnerable populations like minors or immigrants, who may have particular trouble comprehending the parameters of the research study they are agreeing to.

On the final point, it is essential not to confuse informed consent with any strategy to avoid liability. Human subjects should not be forced or expected to release a researcher or other parties from legal consequences should something go wrong while the research is being conducted. And in giving their informed consent, subjects should not be asked to waive any of their legal rights.

Informed consent is not always necessary when conducting research. Remember how the U.S. federal government’s definition of research on “human subjects” that we quoted at the beginning of this chapter stated that it involved “intervention or interaction with the individual” or the use or analysis of “private information.” When we study public speeches given by other human beings, or published books written by them, or publicly released videos or audio recordings created by them, we are not technically conducting research on “human subjects”; instead, we are studying the publicly shared information that those individuals produced. In these cases, we do not need to ask those individuals for permission to study their public communications, which required no intervention or interaction with the individual for us to obtain. Likewise, when we observe public spaces—street corners, for instance—the normal assumption is that no consent is needed to write about or otherwise record what one observes (although there are caveats to this general view, as the uproar over Laud Humphreys’s study of “public” restrooms suggests).

At the same time, it is not always clear what is “private” or “public” information. An everyday example of why researchers should insist on informed consent is the widespread and arguably intrusive analysis of user data for research conducted by companies and advocacy groups. Whenever you sign up for a new app or online service, you typically have to wade through thousands of words of legal disclosures—the dreaded end-user license agreement (EULA)—and click on the “Agree” button before you can upload your cat pictures. Embedded in these disclosures is often an acknowledgment that the provider of the service is using your data for research purposes. They want to know the demographics and habits of their users so that they can improve their service or hawk advertising to marketers who want to reach specific audiences. They often sell their data—including names, mailing addresses, email addresses, and other identifying information—to third parties who could use it for anything from sending you unsolicited emails to developing strategies to encourage (or discourage) you from voting.

In 2018, Facebook faced a major scandal when news reports revealed that a British consultancy called Cambridge Analytica had harvested the data of 87 million Facebook users and—without their knowledge or consent—had used their data to develop “psychographic” profiles that helped political organizations target more effective advertisements at supporters and swing voters across media platforms. The company’s clients included the 2016 presidential campaign of Donald Trump, a Super PAC supporting the Trump campaign, and the 2016 British campaign in favor of a referendum vote to leave the European Union. Individuals close to Trump helped run and fund the company. The data harvesting first occurred through a popular app, “This Is Your Digital Life,” that succeeded in getting hundreds of thousands of users to complete a paid survey said to be part of an academic study. These users agreed to language asking for their consent before taking the survey, yet Facebook allowed Cambridge Analytica to use their personal data for purposes other than what they actually consented to. The company was also able to pull data about the users’ friends—thereby expanding the company’s sample to encompass millions of people on Facebook. As a result of the scandal, Cambridge Analytica folded, and Facebook wound up paying billions of dollars in fines levied by U.S. and UK regulators.

While the deception uncovered in the Facebook/Cambridge Analytica data scandal was shamelessly brazen, it is clear that the “consent” process that companies typically use—an online agreement spelling out the data collection and its uses in fine print—is still a pale imitation of what an informed consent procedure looks like (or at least should look like) in academic research. An ethical researcher needs to be utterly clear about the risks and benefits of the research, convey that information in an accessible and straightforward way, and fully address any questions or concerns without coercion of any kind—which is often the opposite of what the EULA’s impenetrable legalese and “click to proceed” pressure tactics entail.

That said, there are exceptional cases when ethical researchers can justify not fully informing their subjects of the purpose or specific procedures of a study at its outset. In these situations, researchers may be allowed to speak vaguely about their research or even mislead their participants outright. IRBs may approve such strategies when there are serious concerns that participants who learn too much about the study beforehand will subsequently alter their responses or behavior, hampering efforts to get an accurate picture of the phenomenon being examined. For instance, suppose the purpose of a study is to examine to what extent subjects abandon their own views to conform with “groupthink.” To measure this phenomenon in the lab, researchers convene a group of participants who are encouraged to listen to each other’s opinions on a topic before voicing their own. In this research context, disclosing the study’s purpose beforehand might sensitize subjects to the idea of groupthink and—consciously or not—make them act differently than they would have without that information.

Not being fully upfront about the nature of a study is often unavoidable if the researcher wants to conduct a valid analysis of a highly stigmatized phenomenon. For that reason, ethnographers sometimes conduct covert observations—much like undercover officers do when conducting drug busts—so that the people they observe act naturally (and perhaps tolerate the researcher’s presence to begin with, as we discuss further in the next chapter). The social stigma associated with certain behaviors or beliefs makes informed consent particularly tricky for in-depth interviews. In this situation, the researcher usually knows the identities of the participants and must rely on their self-reporting about how they think or feel. Given that respondents can choose how they express themselves in these situations, they may tailor their answers to show themselves in a more favorable light. Specifically, they may hide less desirable behaviors or attitudes while exaggerating more desirable ones. In these cases, so-called social desirability bias will distort the researcher’s understanding of what’s actually going on. For instance, if a sociologist is conducting in-depth interviews with subjects to suss out their underlying racial biases, telling them the actual point of the study will prompt a good number of respondents to provide answers that paint themselves more positively than they actually are. Very few people will comfortably acknowledge that they are unrepentant bigots; at the very least, they will use euphemisms or code words and otherwise dial back the attitudes they might state bluntly to friends and family.

Weighing the pros and cons of not providing informed consent is a tricky matter, and sociologists and IRBs frequently have divergent opinions about what sorts of details can be justifiably concealed. Even if researchers have a strong case for keeping certain information from participants, they can and should conduct a debriefing session immediately following the data collection process. This ensures that the human subjects are eventually informed, even if it cannot occur before they participate in the study. In a debriefing, the researchers explain the true purpose of the study and why any deception was necessary. They also provide a full accounting of the potential risks or harm that the participant might have experienced during the experiment or might still suffer in the future. In all cases, the researchers need to provide ways to contact them or others who can offer assistance while the participants process their experiences afterward.

When obtaining informed consent from our potential participants, we should also keep in mind that some people in our pool of recruits may not be as able as others to give their genuine consent to participate in research. Here, we should revisit the Belmont Report’s principle of “respect for persons.” Recall that this principle not only aims to treat individuals as autonomous agents but also tries to provide protection for people with diminished autonomy. The latter type of human subjects are seen as members of vulnerable populations—people who may be at risk of experiencing undue influence or coercion.

Protection of Identities: Anonymity and Confidentiality

Another important principle of ethical research is protecting the identities of those participating in the study. At the most basic level, participants have a right to know what information about them is or is not shared with others, and these particulars should be spelled out to them during the informed consent process described earlier. It may also be the case that not sufficiently concealing the identities of research subjects endangers their privacy and their public reputations. For instance, if reading the study allows someone to piece together who a specific participant is—which can be very likely when sociologists study a group of people who already know each other—that participant might feel discomfort and embarrassment and could even be in danger of being manipulated through the use of that information. Furthermore, participants who talk about stigmatized behaviors, such as mental health problems, substance abuse, or criminal activity, could be singled out by employers (who routinely do web searches to vet job applicants as well as current employees) or even law enforcement (who sometimes decide to prosecute individuals who admit to crimes in writing or other media). Even beyond the ethical principle of treating participants with respect and protecting them from harm, researchers have practical reasons for promising to safeguard identities. Without such assurances, interview respondents may not feel comfortable speaking honestly about their views or experiences, in the fear that others will criticize, shun, or retaliate against them.

To address these potential problems, researchers often promise to maintain either the anonymity or the confidentiality of their research subjects. Anonymity means that not even the researcher conducting the study should be able to connect specific pieces of information provided by participants with their actual identities. An example of anonymity in scientific research is an internet survey in which no identification numbers are used to track who is responding to the survey and who is not.

Although anonymity is the most effective way to avoid any breaches of privacy in research, it can be impractical to guarantee. For instance, sociologists who use ethnographic observation and face-to-face interviewing get to know about participants in person as part of their interactions. It can be difficult or even impossible in these situations to keep participants’ identities hidden from the researchers.

As a result, sociologists often promise confidentiality rather than anonymity. Under this arrangement, participants allow the researcher to have access to identifying details in the data being collected, but only the researcher should be able to link specific participants to their stored data. In other words, the researcher can identify the specific person who said a particular quote or did a particular thing described in the research report, but they promise not to divulge that person’s identity in any published work or public forum.

A confidentiality agreement implies that a participant’s personal information is secure, but it is important to understand that these guarantees are not ironclad—as much as a well-intentioned researcher might wish them to be so. For one thing, depending on applicable laws, researchers who learn that someone (including the participant) is at “high or imminent risk of harm” may need to contact a health professional or the authorities; the same is true if they learn about the possible abuse of a child (Kodama Muscente 2022). These exceptions need to be mentioned in consent forms so that participants are not given a false sense of security.

Generally speaking, authorities can use court orders to compel researchers to release or discuss the data they collect, which means they may be unable to protect the identities of any participants suspected of involvement in activities under legal scrutiny. Conversations between social science researchers and their human subjects do not automatically enjoy the “privileged communication” status that those between health professionals and patients or between lawyers and clients do. Nor are researchers protected by so-called shield laws, legislation enacted in many states that allows journalists to keep their sources and unpublished notes confidential. In 2003, the U.S. Department of Health and Human Services began allowing researchers to apply for certificates of confidentiality, which prevent them from having to divulge the identities of their human subjects in any legal proceedings, but very few research projects qualify for this level of protection (Office for Human Research Protections 2010). Nevertheless, the code of ethics for the American Sociological Association (ASA)—discussed further in the sidebar Professional Codes of Ethics—is clear that sociologists who promise confidentiality should not cave into that pressure even in the face of legal action: “Confidential information provided by research participants should be treated as such by sociologists even if there is no legal protection or privilege to do so” (American Sociological Association 2018:10).

Researchers have been imprisoned for failing to disclose the identities of their participants. In 1993, sociologist Rik Scarce refused to testify before a federal grand jury investigating an act of vandalism. Then a graduate student at Washington State University, Scarce had been studying radical animal rights activism. Authorities were trying to identify individuals who had broken into research facilities at his university and freed animals used in experiments. They believed Scarce had met the perpetrators and demanded that he discuss his conversations with specific members of the Animal Liberation Front (Scarce n.d.). Citing his moral obligations as a sociologist, Scarce told the court that he would not testify about his conversations with activists, and only agreed—under duress—to disclose nonconfidential details not related to his research. He spent 159 days in jail under a contempt order before a judge finally released him (American Sociological Association n.d.).

Compared with anonymity, confidentiality is a weaker form of protection of identities, and even these agreements are not always possible or desirable for particular kinds of research. For instance, focus groups that involve interviews with multiple people at the same time (discussed in Chapter 10: In-Depth Interviews) cannot ensure confidentiality because the individuals involved will inevitably know at least some identifying details about their fellow respondents. Furthermore, some sociologists have argued that not using the real names of their participants in their published work can have negative repercussions that may outweigh the benefits of protecting privacy (Duneier 1999; Chen and Goldstein 2024). As journalists have long maintained, having individuals provide their quotes on the record with their names attached creates accountability, making it less likely that sources will be able to lie or manipulate without consequences. Some participants may also prefer to see their names in the finished work, perhaps because they want the public recognition or feel their experiences are silenced or trivialized by the use of a pseudonym. Acknowledging these concerns, some IRBs will allow researchers to follow this more journalistic practice of identifying respondents if the risks of the research are minimal and the persons involved explicitly request the use of their real names—sometimes, by signing a “release of information” form distinct from the consent form.

The consent form itself should provide all the key details about how the participant’s identity will be protected, including whether the researchers promise anonymity or confidentiality and how they will go about ensuring those conditions. For instance, when participant data is confidential, researchers will typically assign IDs or pseudonyms to their interview respondents and then keep a password-protected and encrypted document that has the real names associated with those IDs or pseudonyms. The consent form should describe what security measures are in place and who has access to the secure files. If the participants are being taped or video-recorded, the consent form should note that fact and clearly state what will happen afterward to those recordings, which by their very nature include a wealth of identifying details. For instance, recordings may be transcribed afterward and then destroyed; the transcripts may be stripped of any identifiers, such as mentions of the names of the respondents or their family and friends, adding a further layer of protection.

Whether the identities of a study’s participants are anonymous or confidential, it may still be possible for an individual who personally knows that person to spot them in the published paper—perhaps based on the configuration of details provided about their age, gender, race, or education. Qualitative researchers may try not to provide too much in the way of detail about a particular individual in order to minimize this risk. When quantitative researchers release their datasets for public consumption, they will sometimes provide the data in an aggregate form that conceals any microdata—the responses of specific individuals—so that people’s privacy is protected. For similar reasons, government agencies and polling organizations routinely restrict access to data collected from small geographical units, such as zip codes or small cities, given the risk that—within such a narrow pool of respondents—someone could conceivably connect a person’s survey answers to an actual identity.

Sometimes, researchers go so far as to change details—usually minor ones—in order to make individuals unrecognizable in published work even to people who know them. Again, the field of journalism has a different moral take on this practice. (Victor, a former journalist, writes about these divergent sets of professional ethics in the sidebar The Journalist and the Ethnographer.) Most journalists would probably be aghast at even slight tweaks to facts within a write-up. In their view, it would cross the ethical line that separates omission (justifiable in some contexts, such as when reporters omit the names of whistleblowers who fear retaliation) from falsification (inherently wrong).

It’s also worth noting that the ethical standards for protecting privacy in the digital age are evolving and the subject of much confusion among researchers. After all, a vast amount of data is now available about almost every individual across the online spaces they frequent. Sociologists wrestle with how much of the data—including freely shared social media posts by people who are not public figures—can be used for research purposes without violating reasonable expectations of privacy. For their part, IRBs are not always in agreement about what procedures should be in place to protect the creators of publicly shared digital content, much less data retrieved from less freely accessible spaces online.