Ethics in Research

Here's the frame to carry through this whole topic: every rule is a scar. The forms you sign, the boards that review studies, the debriefing you get at the end of an experiment — none of it was handed down from on high. Each protection exists because specific, real people were deceived, injured, or exploited, and the field decided never again. Read the rulebook the way you'd read a body's scars: each one is a story about a wound.

Tuskegee: the wound the whole system remembers

If you only remember one case from this unit, remember this one, because U.S. research ethics as we know it was largely built in its shadow. Starting in 1932, the U.S. Public Health Service, working with the Tuskegee Institute in Alabama, enrolled roughly 600 poor Black men — about 400 who had syphilis and about 200 who did not — into a study of untreated syphilis. The men were told they were being treated for "bad blood." They were not told they had syphilis, and they were not treated. They were given placebos, painful diagnostic spinal taps dressed up as care, and false promises.

The betrayal deepened. By the mid-1940s, penicillin had become the recognized cure — and the researchers deliberately withheld it, choosing instead to watch the disease run its course. The study ran for forty years, until whistleblower Peter Buxtun leaked it to the press in 1972. By then men had died, wives had been infected, and children had been born with congenital syphilis. In 1997 President Clinton offered a formal apology to the survivors. The legislative aftershock came fast: public outrage produced the National Research Act of 1974, which created the commission that wrote the Belmont Report and required the review boards you'll meet below. Tuskegee is not a distant footnote; it is the reason the very next section of this reading exists.

The APA's five principles: Belmont, translated for psychologists

Belmont governs all human research; the American Psychological Association's ethics code (APA, 2002) restates the same commitments as five general principles for psychologists specifically, and they map onto Belmont almost one-to-one. Beneficence and nonmaleficence — benefit the people you work with and, above all, do no harm (Belmont's Beneficence). Fidelity and responsibility — keep the trust participants place in you and take responsibility for what your work does in the world; this is the principle debriefing serves. Integrity — be accurate and honest in research, teaching, and reporting; deception, when used, must be justified, minimal, and corrected. Justice — fairness in who bears the burdens of research and who gets its benefits (Belmont's Justice). Respect for people's rights and dignity — privacy, confidentiality, self-determination, and special care for anyone whose ability to decide freely is compromised (Belmont's Respect for Persons). Underneath all five sits the plain obligation the guidelines exist to enforce: treat participants with kindness, respect, and fairness — as people doing you a favor, not as material.

The conundrum underneath everything: risk vs. benefit

Here is the central tension that never fully resolves. Research can hurt people, but research also produces knowledge that helps people. So the fundamental question an ethics review keeps asking is a balance: do the potential benefits of the knowledge gained justify the potential risks to the participants? That's it. That's the conundrum. It has no formula — reasonable people weigh it differently — which is exactly why we don't leave the judgment to the researcher alone (who is, after all, invested in their own study going forward). We hand it to a review board.

A key calibration point in that balance is the idea of minimal risk: risk no greater than what a person would ordinarily encounter in daily life or in routine physical or psychological exams. A study at or below minimal risk gets lighter scrutiny; anything above it triggers heavier protections, documented consent, and closer review. When you hear a board argue about whether a study is "greater than minimal risk," they're arguing about which side of that line it falls on.

Who watches the balance: IRBs and IACUCs

The Institutional Review Board (IRB) reviews proposed research with human participants before it can begin. Crucially, an IRB isn't just a room full of the researcher's colleagues — it must include at least one community member from outside the institution, precisely to check the conflicts of interest baked into academia (your career advances if your study runs; the IRB's job is to make sure that incentive doesn't quietly override participant safety). IRBs sort proposals into levels of review — exempt, expedited, and full — scaled to how much risk is on the table. Its parallel for animal research is the IACUC (Institutional Animal Care and Use Committee), which protects the welfare of animal subjects. An IACUC includes a veterinarian alongside scientists, administrators, and a community member; it must approve any animal protocol before it starts, requires that pain and distress be minimized, and inspects the animal facilities twice a year. Most psychological research with animals — the large majority of it on rodents and birds — exists precisely because the equivalent study on humans would be unethical, which is why the review bar is high rather than absent.

Informed consent: a conversation, not a signature

The single most common student misconception is that informed consent is a form. It isn't. It's an ongoing process — the communication of information about a study, in language the person can actually understand, from which they make a genuinely voluntary decision to take part (or not). The signed form documents that process; it doesn't replace it. Documented consent is required whenever a study poses greater than minimal risk. Walk through what a real consent process has to put on the table:

Deception: the controlled exception, and its costs

Social psychology has an awkward dependency: sometimes people won't behave naturally if they know the true purpose of a study, so researchers construct a cover story to keep behavior authentic. This is deception, and it's tolerated only under strict conditions. It comes in flavors worth naming. Deception by commission is actively giving false information — telling participants the study is about "learning and memory" when it's really about obedience. Deception by omission is withholding a key detail — not mentioning that the "other participant" is a confederate. Two finer distinctions round it out: technical deception misrepresents the equipment or procedure (a fake shock generator), while role deception misrepresents the people in the study (confederates posing as fellow subjects). Some textbooks add implicit deception: nobody states a falsehood, but the setup is arranged so participants draw a false conclusion on their own — a "waiting room" that is really the study. It counts as deception all the same, because the participant's understanding is what matters, not the researcher's technical honesty.

Deception is never free. Martin Orne (1962) warned that it can backfire into demand characteristics — cues that lead participants to guess the hypothesis and then, consciously or not, play along to help (or sabotage) the researcher, present themselves favorably, or otherwise stop behaving naturally. And there's a hard ethical floor beneath all of it: APA standards forbid deceiving participants about anything that would affect their willingness to take part in the first place — physical risks, significant discomfort, unpleasant emotional experiences. You may hide the hypothesis; you may not hide the harm.

Debriefing and coercion: the bookends of a fair study

When deception is used, debriefing is mandatory. After the study, the researcher explains the true purpose, describes the deception and why it was necessary, and works to undo any discomfort — restoring an honest relationship with the participant and, along the way, teaching them something real about how psychological science works. Debriefing is how the field pays back the trust it borrowed.

On the front end sits coercion. No undue pressure may be placed on anyone to join or stay in a study. Researchers can't order, threaten, or guilt people into participating — and, less obviously, they can't offer excessive payments that tempt people into risks they'd otherwise refuse (an "offer you can't refuse" isn't a free choice). This is also why a researcher shouldn't hold power over their participants, and why students who face a research requirement in a course must always have a genuine alternative way to fulfill it. Consent given under pressure isn't consent.

Vulnerable populations: each protection is its own scar

Certain groups get extra layers of protection, and — true to this whole topic — none of them landed on the list by accident. Each became "protected" because research went badly wrong with that group, and courts, agencies, or legislation responded. Walk the list and you're really walking a history of specific harms:

• Children / minors — can't legally consent and may not grasp risk the way adults do; past abuses led to rules requiring parental permission and child assent, with tight limits when there's no direct benefit.
• Pregnant women, fetuses, neonates — historically either exploited or excluded without justification; disasters like thalidomide forced added review and risk/benefit scrutiny (45 CFR 46 Subpart B).
• Prisoners — live in an inherently coercive environment, so "voluntary" consent is murky; a long record of coercion and undue inducement produced special federal rules (Subpart C).
• Impaired decision-making capacity — people with cognitive impairment or severe mental illness were experimented on without real understanding or consent; rules now require capacity checks, surrogate consent, and assent where possible.
• Economically / educationally disadvantaged — the poor, illiterate, or non-English-speaking are easiest to exploit and pressure; boards must now ensure understandable forms, interpreters, and guard against undue influence.

A more recent turning point widened the lens beyond individuals to whole communities. In the Havasupai case (Havasupai Tribe v. Arizona Board of Regents, 2004–2010), blood samples collected from tribe members for diabetes research were later used for unrelated studies — on schizophrenia, migration, and inbreeding — without consent. The case reshaped conversations about consent, cultural vulnerability, and the rights of Indigenous groups, and it exposed a limit in the classic model: consent isn't just a one-time checkbox for one study; it can be violated by re-use the participant never agreed to.

A catalog of cautionary cases

The field's ethics rules were written against a backlog of studies that could not run today. Here they are in order — each with the principle it trampled. Read them not as villains-of-the-week but as the raw material the Belmont principles were forged to prevent. Click any case to open a short account of what actually happened (drawn from the Tuffin reading).

• 1938 Henle & Hubbell hid under dormitory beds to record students' private conversations — extreme invasion of privacy, no consent (Respect for Persons).

  Henle and Hubbell wanted to know whether adult conversation is as self-centered as Piaget said children's is, so they went where people talk unguarded: they hid under beds in dormitory rooms and eavesdropped in lounges and washrooms, jotting down snippets of talk and coding how much of it was about the speaker. The students never knew, never agreed, and were never told. The finding (adults talk about themselves a lot) is almost beside the point; the method is the lesson. Privacy was treated as an obstacle to data rather than a right.

• 1956 Festinger, Riecken & Schachter (When Prophecy Fails) infiltrated a doomsday cult to watch dissonance when the world didn't end — deception, no consent, vulnerable believers (Respect).

  A small Midwestern group believed a flood would end the world on a set December date and that flying saucers would rescue the faithful. Festinger and colleagues, testing their brand-new dissonance theory, joined the group posing as sincere believers and took notes for weeks. When the date passed uneventfully, the most committed members did not abandon the belief; they decided their faith had spared the world and began recruiting harder, exactly as dissonance theory predicted. Homan (1991) called it "a remarkable case of deceit and invasion": the members never knew they were data, and the observers' presence may itself have propped up the very conviction under study.

• 1962 Orne had participants solve and destroy thousands of pointless math sheets for hours to show experiments generate their own compliance pressure — exploited trust, undue burden (Respect; Beneficence).

  Orne was hunting for a task so pointless that people would refuse it, so he could use refusal as a baseline in a hypnosis study. He couldn't find one. Participants were handed 2,000 sheets, each with 224 addition problems, and a card telling them to finish a sheet, tear it up, and take the next card, which said the same thing. Nobody quit; people kept adding and shredding for hours. Orne's conclusion is the birth of the idea of demand characteristics: simply being in an experiment makes people compliant, so any result has to be read against that pressure. The ethical cost was smaller than in most cases here, but the participants' trust was the raw material.

• 1962 Berkun et al. told military trainees mid-flight that their plane was crashing, complete with fake pilot chatter and "death preparation" forms — severe psychological harm, coercion, no real consent (Beneficence; Respect).

  Army recruits were flown to about 5,000 feet, where a propeller was cut and the men overheard the pilot reporting an emergency. Passing over the airfield, they could see ambulances and fire engines lining up. The pilot then announced the landing gear had failed and to prepare for a ditching at sea. Two forms were handed out: instructions for disposing of personal belongings in case of death, and a multiple-choice quiz on emergency procedures, supposedly so the insurers would know precautions had been followed. The papers went into a waterproof container to be thrown clear before impact. The plane landed safely; the whole thing was staged to measure performance under terror. The recruits had never been told they were research subjects.

• 1963 Milgram had participants deliver what they believed were dangerous shocks on an authority's orders; 65% went to 450 volts — deception, severe distress, thin debriefing (Beneficence).

  Participants were told the study was about punishment and learning, then assigned (by a rigged draw) to be the teacher who shocked a learner for every wrong answer. The shock generator was a convincing prop with switches climbing to 450 volts; the learner was a confederate; no shocks were delivered. That is technical deception (fake apparatus) stacked on role deception (a fake fellow participant). Twenty-six of forty went all the way, while sweating, trembling, and pleading to stop. The debriefing was real but, critics argued, no match for what had been done to people's image of themselves. You'll meet this study again in Obedience.

• 1972 Piliavin & Piliavin staged a subway collapse (sometimes with fake blood) to study bystander helping — no consent, risk of panic and injury to real passengers (Beneficence).

  On a moving Philadelphia subway car, a confederate collapsed to the floor. In one condition a trickle of stage blood ran from his mouth; in the other there was none. The hypothesis, from a cost-of-helping model, was that blood raises the cost of stepping in, so help would drop. It did. Field realism came with field problems: transit police discovered the team, and some riders reacted in ways nobody had scripted, pulling the emergency cord and, in a few cases, panicking at the sight of blood. In one trial two passengers rose to help, one gasped "Oh, he's bleeding," and both sat back down. No passenger consented to any of it.

• 1974 Harris had confederates cut in line to provoke reactions — no consent, real risk of escalation to aggression (Beneficence).

  Harris sent confederates to push into queues at supermarkets and movie theaters, ahead of unsuspecting shoppers, to see how people respond to a small, pointed frustration. The expected reaction, and the usual one, was a muttered complaint. But the design had no way to guarantee that. A shopper at the end of a bad day could have answered with a shove or worse, and the confederate would have been the one in harm's way, too. Tuffin's point about this study is the general one: when people are drawn into research without knowing it, the researcher has surrendered control over what happens next.

• 1975 West, Gunn & Chernicky recruited students to help stage a burglary for a fake private investigator; 45% agreed if promised immunity — extreme deception, legal jeopardy, no debriefing (Respect; Beneficence).

  In the Watergate era, undergraduates were approached by a man claiming to be a private investigator working for the IRS. He wanted help breaking into a local advertising firm to photograph records that would prove the company was defrauding the government. Depending on the condition, participants were promised immunity from prosecution, a cash reward, or nothing. With immunity on the table, 45% said they'd take part. Only after they had answered did they learn it was a study. The finding is about how ordinary people talk themselves into wrongdoing; the method asked people to agree to a felony before telling them it was hypothetical.

• 1976 Middlemist, Knowles & Matter watched men urinate via a hidden periscope to test personal-space invasion — gross invasion of privacy in an intimate act, no consent (Respect for Persons).

  In a university men's restroom, a confederate stood either at the urinal directly beside a user, one urinal away, or not at all. In a nearby stall an observer with a periscope timed how long it took each man to begin urinating and how long he continued. The closer the confederate, the longer the delay and the shorter the stream, which the authors read as physiological arousal from having personal space invaded. Nobody was asked. The men were observed at one of the most private moments of the day and never knew it. The result is a legitimate question about personal space answered in a way that treats dignity as a nuisance.

• 1976 Shotland & Straw staged a public argument between confederates, framed as strangers or spouses, to see who'd intervene — deception, distress to unwitting bystanders (Beneficence).

  Paid actors staged a loud, physical fight in which a man attacked a woman in front of bystanders who had no idea it was theater. The only thing that varied was one line she shouted: either "I don't know you" or "I don't know why I ever married you." Onlookers were far more likely to step in for a stranger than for a wife, which says something uncomfortable about how people read domestic violence. The bystanders were frightened, some were placed in a position where intervening could have gotten them hurt, and none had agreed to be part of an experiment. Tuffin notes the obvious: the results would have looked very different had they known the couple were acting.

• 1970/75 Humphreys' Tearoom Trade — covert observation of anonymous sex, then tracing men by license plate and visiting their homes under a false identity — voyeurism, deception, breach of privacy (all three principles).

  Humphreys, a sociology doctoral student, wanted to know who actually sought anonymous sex in public restrooms, since the stereotype said openly gay men. He posed as the "watch queen," the lookout who coughs when a stranger or the police approach, and in that role observed the encounters. Men who didn't confide in him were traced through their license plates; a year later he appeared at their homes, disguised, as an interviewer for a health survey, and collected details of their marriages, jobs, and families. Only 14% identified as gay; most were married men living conventional lives. Some sociologists demanded his degree be revoked. Gay advocates praised the work for demolishing a myth. Both reactions are still worth taking seriously, and the page returns to that below.

• 1981 Zimbardo, Andersen & Kabat used post-hypnotic suggestion to induce temporary hearing loss, leaving participants confused and paranoid — psychological harm, deception, inadequate debriefing (Beneficence).

  College students were told the study concerned hypnosis and problem solving, and the early sessions were pleasant enough: better concentration, less fatigue. Then came a post-hypnotic suggestion that they would have trouble hearing, with no memory of where the trouble came from. Asked to interpret a set of slides while others talked around them, they grew confused, angry, and suspicious that people were whispering or lying. That was the actual hypothesis: paranoia in older adults may begin as a reasonable attempt to explain hearing loss the person hasn't recognized. Deny it, and every reassurance sounds like a cover-up. The participants got to live that logic for an afternoon without having agreed to. Zimbardo's other famous work comes up later too.

The hard one: what do we do with knowledge gained unethically?

Now the question this unit refuses to answer for you — because it's a genuinely open one. Consider Laud Humphreys' Tearoom Trade (1970). Ethically, it is close to a worst case: covert observation of men engaged in anonymous sexual encounters, followed by tracking those men down through their license plates and interviewing them at home while posing as a health worker. Voyeurism, deception, breach of privacy, exploitation of a criminalized and stigmatized group — it fails every principle we've built.

And yet. Humphreys' findings challenged a damaging stereotype of the era. His data showed that only about 14% of the men he observed identified as gay; many were married, middle-class men living conventionally heterosexual lives. At a time when the culture cast this behavior as the exclusive province of openly gay men, his results complicated that story and were welcomed by advocates as a blow against a harmful myth. So the knowledge was real, and arguably reduced harm — even though the method inflicted it. That's the wrinkle. Sit with three honest positions, none of which the field has fully settled:

• Discard it. Using knowledge obtained by violating human dignity ratifies the violation and creates an incentive for the next researcher to cut the same corner. The data is fruit of a poisoned tree.
• Use it. The harm is already done and can't be undone by ignoring what was learned; refusing to use knowledge that could reduce suffering wastes the one thing the participants involuntarily gave. Throwing it away helps no one now.
• A middle path. Acknowledge and even use the knowledge, but never launder its origins — cite it alongside the harm, teach the ethics with the finding, and let the case stand as a permanent warning rather than a clean result.

The point isn't to hand you the right answer; it's to make sure you feel the weight on both sides. The same dilemma recurs — with Nazi hypothermia data, with any finding whose value is real and whose origins are indefensible. A person who can hold that tension without collapsing it into a slogan is thinking like an ethicist.

Check yourself

The one thing to carry out of this topic

Research ethics can feel like bureaucracy — forms, boards, boxes to check — right up until you learn where each rule came from. Then it reads differently: as a memorial. Informed consent exists because men in Alabama were lied to for forty years. Debriefing exists because people were terrified into believing their plane was crashing. Community members sit on IRBs because researchers, left alone, will find reasons their own studies are fine. So when you design a study, run a survey for a class, or just evaluate a claim someone made "in the name of science," ask the two questions this whole topic reduces to: Do the benefits justify the risks to these particular people? and Would I be comfortable if this were done to me, or to someone I love? Every rule is a scar. Your job is to not add new ones.

References

American Psychological Association. (2002). Ethical principles of psychologists and code of conduct. American Psychologist, 57(12), 1060–1073. https://doi.org/10.1037/0003-066X.57.12.1060

Berkun, M. M., Bialek, H. M., Kern, R. P., & Yagi, K. (1962). Experimental studies of psychological stress in man. Psychological Monographs: General and Applied, 76(15), 1–39. https://doi.org/10.1037/h0093885

Festinger, L., Riecken, H. W., & Schachter, S. (1956). When prophecy fails. University of Minnesota Press.

Harris, M. B. (1974). Mediators between frustration and aggression in a field experiment. Journal of Experimental Social Psychology, 10(6), 561–571. https://doi.org/10.1016/0022-1031(74)90079-9

Henle, M., & Hubbell, M. B. (1938). "Egocentricity" in adult conversation. The Journal of Social Psychology, 9(2), 227–234. https://doi.org/10.1080/00224545.1938.9921699

Humphreys, L. (1970). Tearoom trade: Impersonal sex in public places. Aldine.

Middlemist, R. D., Knowles, E. S., & Matter, C. F. (1976). Personal space invasions in the lavatory: Suggestive evidence for arousal. Journal of Personality and Social Psychology, 33(5), 541–546. https://doi.org/10.1037/0022-3514.33.5.541

Milgram, S. (1963). Behavioral study of obedience. The Journal of Abnormal and Social Psychology, 67(4), 371–378. https://doi.org/10.1037/h0040525

National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research. (1979). The Belmont Report: Ethical principles and guidelines for the protection of human subjects of research. U.S. Department of Health, Education, and Welfare.

Orne, M. T. (1962). On the social psychology of the psychological experiment: With particular reference to demand characteristics and their implications. American Psychologist, 17(11), 776–783. https://doi.org/10.1037/h0043424

Piliavin, J. A., & Piliavin, I. M. (1972). Effect of blood on reactions to a victim. Journal of Personality and Social Psychology, 23(3), 353–361. https://doi.org/10.1037/h0033166

Shotland, R. L., & Straw, M. K. (1976). Bystander response to an assault: When a man attacks a woman. Journal of Personality and Social Psychology, 34(5), 990–999. https://doi.org/10.1037/0022-3514.34.5.990

West, S. G., Gunn, S. P., & Chernicky, P. (1975). Ubiquitous Watergate: An attributional analysis. Journal of Personality and Social Psychology, 32(1), 55–65. https://doi.org/10.1037/h0076466

Zimbardo, P. G., Andersen, S. M., & Kabat, L. G. (1981). Induced hearing deficit generates experimental paranoia. Science, 212(4502), 1529–1531. https://doi.org/10.1126/science.7233242