What Should Researchers Do When a Participant Appears to Pose a Serious Risk to Someone Else?

What if a Participant Says Something That Suggests Someone Else May Be in Danger?

During an interview, a participant describes wanting to seriously hurt a particular person. A mental-health questionnaire is followed by a statement about retaliating against someone. A participant becomes agitated and makes a threat that research staff cannot confidently dismiss.

Researchers now face a difficult conflict. Confidentiality protects participants and supports candid research participation. Yet ignoring a credible threat could expose another person to serious harm.

The temptation is to reduce the problem to a familiar phrase: “duty to warn.” That is too simple. Duties to warn or protect vary by jurisdiction, often depend on professional role and the characteristics of the threat, and were largely developed in clinical relationships rather than as one universal rule governing all researchers.

Serious Threats Require Prompt Assessment, but There Is No Universal Researcher Duty to Warn

“Duty to Warn” Is Not One Universal Rule

The legal idea comes largely from clinical practice

Much discussion of threats toward others begins with Tarasoff v. Regents of the University of California, the influential U.S. case associated with psychotherapists' responsibilities when a patient poses a serious danger to another person.

But “Tarasoff duty” is often used too casually. APA materials emphasize that jurisdictions differ significantly. Some impose duties to warn, some duties to protect, some permit protective disclosure without requiring it, and others use different or no comparable statutory requirements.

Researchers therefore should not copy a clinical rule from another jurisdiction into a research protocol and assume the legal problem has been solved.

Professional role matters

A licensed psychologist conducting clinical research may have professional obligations that do not apply identically to a non-clinician sociologist, research assistant, computer scientist, or data analyst encountering the same words in research data.

Some duty-to-protect laws are specifically framed around mental-health professionals or particular therapeutic relationships. APA's legal discussions repeatedly emphasize that the applicable duty depends on the jurisdiction and professional role.

The study therefore needs to identify who on the research team may have independent professional duties in addition to responsibilities arising from the research protocol.

A violent statement is not automatically a credible imminent threat

Participants can describe anger, fantasies, past behavior, hypothetical violence, or rhetorical threats without necessarily posing an immediate serious danger. Conversely, a specific and credible threat can require urgent attention.

Researchers who lack training in violence-risk assessment should not be expected to determine dangerousness independently. The protocol should specify what information triggers escalation and who is qualified to evaluate it.

The first can be recognized by appropriately trained research staff without requiring them to perform the second.

Specificity and identifiability can matter legally

Some duty-to-protect frameworks attach particular importance to serious threats directed toward an identifiable or reasonably identifiable person. Other jurisdictions formulate the threshold differently.

APA's discussion of Pennsylvania's Emerich decision, for example, describes a duty arising in the context of a mental-health professional determining that a patient presented a serious danger of violence toward another and an identifiable threatened person. California's Tarasoff-related rules have developed differently.

These examples illustrate legal variation rather than a checklist researchers should apply universally.

Protecting someone does not always mean warning them directly

The phrase “duty to warn” can obscure the broader concept of a duty to protect. Depending on the governing law and circumstances, reasonable protective action might involve warning an identifiable person, contacting law enforcement or emergency services, arranging urgent clinical evaluation, taking steps through a healthcare system, or another intervention permitted or required by law.

APA materials note that many jurisdictions frame the obligation as protection rather than requiring one specific warning mechanism.

Researchers should therefore follow the applicable professional and institutional process rather than assuming that direct contact with a potential victim is always the legally correct response.

Confidentiality remains important even when safety is at stake

The existence of a possible threat does not make all participant information freely disclosable. Any protective disclosure should have an appropriate ethical and legal basis and should ordinarily be limited to what is necessary for the protective purpose.

APA ethics materials distinguish between jurisdictions that require disclosure and those that permit it, emphasizing that legal assessment and clinical judgment interact in determining what action is appropriate.

This is precisely why the broader question of when confidentiality gives way to a duty to act or report should be addressed prospectively.

Certificates of Confidentiality require careful interpretation

For U.S. research covered by NIH Certificates of Confidentiality, identifiable sensitive research information receives substantial statutory protection. Current NIH policy specifies limited circumstances in which disclosure is permitted, including disclosures required by federal, state, or local law.

Older OHRP guidance also explains that Certificates do not themselves prevent researchers from voluntarily disclosing information such as a participant's threatened violence to self or others, while emphasizing that intended voluntary disclosures should be explained in consent.

Because current statutory protections, local law, institutional policy, and the study's consent language all matter, research teams should obtain current institutional or legal guidance rather than treating “we have a Certificate” as a complete answer.

The consent form should explain foreseeable safety exceptions

If a study is likely to elicit threats of violence or other safety-sensitive disclosures, participants should understand relevant limits to confidentiality before deciding to participate.

A consent statement might explain that confidentiality will ordinarily be protected but that particular information indicating a serious risk of harm may need to be shared as permitted or required under the applicable framework. The precise wording should be developed for the actual study and jurisdiction rather than copied generically.

Promising absolute confidentiality and then discovering that a professional or legal duty requires disclosure places both participant trust and the research team in an avoidable conflict.

Threat assessment should not become interrogation

Research staff may need enough information to activate the correct safety pathway, but they should not conduct an improvised forensic investigation.

If the participant's statement meets the protocol's escalation threshold, the appropriate next step may be to involve a qualified clinician, safety officer, principal investigator, institutional safeguarding contact, or another designated professional.

Immediate danger changes the pace

A vague statement about past anger and a specific threat suggesting immediate serious violence should not move through the system at the same speed.

If the information suggests an immediate danger, the study's urgent safety procedure should be activated promptly. Researchers should know in advance how to reach appropriate clinical, institutional, security, or emergency resources.

The broader principle is the same as for research information affecting health or safety: greater credible urgency calls for faster escalation, without pretending the researcher knows more than the evidence supports.

Remote research creates additional problems

A participant may make a concerning threat during a video interview while physically located in another state or country. An online platform may identify a concerning statement hours after it was posted. A research team may not know the participant's current location.

These circumstances complicate both emergency response and legal analysis. APA guidance for remote mental-health practice stresses the importance of understanding the law where the person is located because duties to warn or protect vary substantially across jurisdictions.

Remote studies likely to encounter acute safety concerns should therefore determine prospectively what location information is available, which jurisdiction's procedures may apply, and what the study can realistically do.

Researchers should avoid overpredicting violence

Threat assessment is difficult. APA's legal materials have long cautioned against assuming that mental-health professionals can predict violence with certainty, much less expecting untrained researchers to do so.

A safety protocol should therefore create a defensible process for assessment and protection rather than pretending to eliminate uncertainty.

Threats to others and suicide risk should not be treated as identical

Both situations may require urgent safety responses, but they involve different potential beneficiaries of protection and different legal doctrines. A suicide-risk protocol should not simply be renamed a violence-risk protocol.

When the potential harm is directed toward the participant, the process described for self-harm or suicide risk during research is more directly relevant. When another person may be at risk, duties to protect third parties and confidentiality law can become central.

A Participant Makes a Specific Threat During an Interview

The interviewer did not need to become an expert violence predictor. They needed to recognize the safety trigger and know exactly where the responsibility moved next.

Common Mistakes When a Participant May Pose a Risk to Others

Build a Threat-Escalation Process, Not a Researcher Guessing Game

If your research could plausibly elicit threats of serious violence, determine the relevant law and institutional requirements before enrollment. The protocol should specify what triggers escalation, who performs further assessment, who can obtain urgent legal or ethics consultation, and what protective resources are available.

The goal is neither reflexive disclosure nor reflexive secrecy. It is a defensible process that protects participants and others while respecting confidentiality as far as the circumstances and applicable rules allow.

Before Conducting Research That May Elicit Threats

Questions About Participants Who May Pose a Risk to Others

Take Serious Threats Seriously Without Inventing a Universal Duty to Warn

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.