Research Ethics and Confidentiality in Criminal Justice Research

Every field of social research owes its participants basic protections, but criminal justice research concentrates several risks that rarely combine elsewhere. First, many of its participants are literally captive: prisoners, people on remand, and juveniles in detention cannot leave the setting where a researcher approaches them, which strains the idea that consent is freely given.

A prisoner who believes that refusing an interview might affect how staff or a parole board view them is not choosing as freely as a member of the public approached on the street.

Second, the subject matter is frequently self-incriminating or stigmatising. A study of active drug dealing, gang membership or intimate partner violence asks people to describe conduct that could expose them to prosecution, retaliation or social exclusion if it became known. Unlike a study of consumer habits, the researcher's field notes and interview recordings can function as evidence, which is why confidentiality protections in this field carry legal as well as ethical weight.

Third, criminal justice research studies victims as well as offenders, and victims bring their own vulnerability: a recent assault, an ongoing custody dispute, or trauma that a poorly designed interview can reopen. A researcher studying intimate partner violence must weigh the value of first-hand testimony against the real possibility of re-traumatising the person who provides it, and against the danger that a perpetrator might discover the participant spoke to a researcher at all.

These pressures explain why criminal justice research sits under some of the strictest ethical review in the social sciences, and why the principles and procedures covered in this chapter are treated as a precondition for publishable, fundable research rather than as bureaucratic overhead.

Modern human-subjects protection in the United States traces to the Belmont Report (1979), written by the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research after revelations about abusive studies, including the Tuskegee syphilis study. The report set out three principles that still frame ethical review internationally: respect for persons, beneficence and justice.

Respect for persons requires treating participants as autonomous agents capable of deciding for themselves whether to take part, and it is the principle that generates the informed consent requirement. Consent must be voluntary, informed by an honest description of the study's purpose and risks, and revocable, meaning a participant can withdraw at any point without losing any benefit they were promised.

Where a participant cannot give full autonomous consent, a juvenile or a person with a cognitive impairment, respect for persons requires extra protections rather than exclusion from research altogether: a parent or guardian consents, and the researcher still seeks the participant's own assent in language they can understand.

Beneficence requires that a study maximise possible benefits and minimise possible harms, and that this balance be assessed honestly rather than assumed. A criminal justice study that asks juveniles to relive an assault for a marginal gain in academic knowledge fails this test even if consent was properly obtained, because consent does not by itself make a harmful design acceptable.

Justice requires that the burdens and benefits of research be distributed fairly. Historically, prisoners and other captive populations were overused as research subjects precisely because they were easy to recruit and could not refuse convincingly, while the knowledge gained rarely benefited them. Contemporary ethical review treats disproportionate recruitment of a vulnerable group as a justice violation, not a matter of convenience.

An institutional review board, called a research ethics committee in the United Kingdom and much of the Commonwealth, is a standing panel that reviews proposed research involving human participants before data collection begins.

In the United States this review is required under the Common Rule (45 CFR 46) for federally funded research; equivalent university-level review operates in the United Kingdom under frameworks published by bodies such as the Economic and Social Research Council, and in India through institutional ethics committees that apply guidance issued by the Indian Council of Medical Research.

The board checks that risks to participants are minimised and are justified by the study's likely value, that the consent process is adequate, that vulnerable groups receive extra protection, and that data-handling plans meet the confidentiality standard the study promises.

Studies are typically sorted into review tracks by risk: an anonymous survey of university students about their view of policing may qualify for an expedited or exempt review, while an interview study inside a prison, or one that asks about active offending, requires full board review and periodic re-approval for as long as the study continues to collect data.

Review is not a one-time gate. Boards can require a study to report adverse events, amend its consent form, or stop recruiting if a risk emerges that was not anticipated at approval. A researcher who changes their method mid-study, adding a new interview question about self-harm, for instance, is normally required to seek an amendment rather than treat the original approval as covering it.

Ethics review has critics: some researchers argue that boards apply biomedical risk models poorly suited to interview-based social research, or that review timelines discourage timely fieldwork on fast-moving issues like a riot or a policing scandal. The consensus response in the methodology literature is not to abandon review but to train boards in social-science risk assessment and to build proportionate, fast-track pathways for low-risk qualitative work.

Confidentiality and anonymity are often used interchangeably but describe different protections. Anonymity means no one, including the researcher, can trace a response back to the person who gave it, because no identifying information was ever recorded; an anonymous online survey with no login or IP capture meets this standard.

Confidentiality means the researcher does know who provided which data but commits to controlling access to it, through measures such as coded identifiers, encrypted storage, and destroying the key that links code to name once it is no longer needed.

Most criminal justice interview research can only offer confidentiality, not anonymity, because a face-to-face interview or a case file review makes the participant identifiable to the researcher by definition. The ethical weight then falls on how rigorously that confidentiality is protected: who can access the raw data, how transcripts are stored, and what a researcher will do if a court, an employer or a family member demands to see them.

In the United States, a Certificate of Confidentiality, issued for research collecting sensitive identifiable information, lets a researcher decline to disclose participant data even under subpoena in most civil, criminal, administrative or legislative proceedings; the protection was made available automatically for qualifying federally funded research after amendments tied to the 21st Century Cures Act.

No exact equivalent exists in most other jurisdictions, so a UK or Indian researcher promising confidentiality to a participant describing an ongoing crime must be candid that the promise has a narrower legal shield than a US Certificate provides.

Data-protection law adds a separate layer on top of research ethics. The European Union's General Data Protection Regulation, in force since May 2018, and the United Kingdom's Data Protection Act 2018 both classify data about criminal convictions and offences as a special category requiring extra safeguards, and both give participants rights to access or request erasure of their data that a research design has to accommodate from the outset.

Prisoners are the paradigm case of a captive population in research ethics, because incarceration itself can be experienced as pressure to cooperate with anyone perceived to hold influence over conditions or release.

US federal regulations (45 CFR 46, Subpart C) impose additional safeguards on prisoner research: a prisoner representative typically sits on the reviewing board, the study cannot offer incentives large enough to be coercive given the restricted prison economy, and any benefit of the research to participants must be weighed against the risk that participation, or refusal, becomes visible to staff who control daily life.

Juveniles raise a parallel but distinct problem. A child or adolescent in conflict with the law cannot give fully autonomous legal consent, so researchers seek parental or guardian permission alongside the juvenile's own assent, and many ethics frameworks require a juvenile's assent to be honoured even where a guardian has already agreed to their participation.

Where a study touches abuse or neglect, most jurisdictions also impose mandatory reporting obligations on the researcher that override any promise of confidentiality made to the child.

Crime victims bring the risk of secondary victimisation: describing an assault, a burglary or an act of trafficking in a research interview can reopen psychological injury, particularly if the interview is conducted without trauma-informed technique or without a clear route to support services, a concern set out further in foundations of victimology.

Ethical designs for victim research typically build in a debrief protocol, referral information for counselling and, where relevant, a pathway into victim rights, support and the impact of crime, and boards increasingly ask for this as a condition of approval rather than a researcher's optional good practice.

A further, less discussed group is justice-system staff, police officers, prosecutors, prison officers, whose institutional position creates its own coercion risk if a superior effectively directs staff to take part in a study of their own agency. Ethical review treats consent obtained through a chain of command with the same scepticism as consent obtained inside a cell.

Covert research, observing or interacting with participants without disclosing that a study is underway, denies people the chance to consent at all, and most ethics frameworks treat it as justified only where three conditions hold together: the research question cannot be answered any other way, the topic has clear social value, and the risk of harm to participants is low.

Classic studies of closed subcultures, from gangs to extremist groups, have used covert or partially covert methods on this reasoning, but a covert design still requires board approval in advance, and a researcher cannot simply decide alone that deception is warranted.

Where covert observation is approved, researchers typically owe participants a debrief after the fact wherever it can be delivered safely, and the ethics board weighs whether debriefing itself would create risk, for instance by exposing an informant inside a group under study. Overt alternatives, disclosing the researcher's role to gatekeepers while remaining unobtrusive to wider participants, are generally preferred wherever they can still answer the research question.

Data security is the practical backbone of any confidentiality promise. Good practice separates identifying information from research data through coded identifiers, stores the key to that code separately from the data itself, encrypts files and devices, and sets a destruction date for identifying material once it is no longer needed for the study.

A confidentiality promise that is not backed by these measures is not a real protection, and boards increasingly require a written data-management plan before approving a study that touches sensitive criminal justice material.

Confidentiality is never absolute. Most jurisdictions recognise limits where a participant discloses an ongoing risk to an identifiable third party, a pattern that follows the reasoning in Tarasoff v. Regents of the University of California (1976), where the California Supreme Court held that a mental health professional's duty to protect an identifiable, threatened person can override confidentiality.

Many jurisdictions also impose mandatory reporting duties for disclosed child abuse regardless of a researcher's promise. Ethical practice requires stating these limits to a participant before they disclose anything, in the consent process itself, rather than discovering the boundary only when a disclosure forces the researcher's hand.