What Should You Consider Before Planning Research With Vulnerable or Hard-to-Reach Populations?

What Should You Think About Before Designing Research Around a Population That May Be Vulnerable or Difficult to Reach?

Some research questions cannot be answered well by recruiting whoever is easiest to find. You may need to study children, prisoners, displaced people, people with impaired decision-making capacity, people experiencing poverty, stigmatized communities, undocumented migrants, survivors of violence, people with rare conditions, geographically isolated populations, or others whose circumstances make research participation particularly complicated.

Two concerns often become entangled. Some prospective participants may be in situations that increase their susceptibility to harm, coercion, undue influence, exploitation, or exclusion. Others may simply be difficult for researchers to identify, contact, recruit, or retain. A population can be both, but the concepts are not interchangeable.

Before finalizing a study around either situation, you need to ask more than whether you can recruit enough participants. You need to understand why this population is scientifically necessary, what creates the vulnerability or access difficulty, whether participation can be genuinely voluntary and sufficiently safe, and whether your research design respects rather than exploits the circumstances that made the population important to study in the first place.

Start With Scientific Necessity, Then Examine the Population's Actual Circumstances

Vulnerability, Access, and Ethical Inclusion Need to Be Considered Together

Do not treat "vulnerable population" as a permanent label

Older approaches to research ethics often described entire categories of people as vulnerable. Contemporary guidance is more cautious about that approach.

CIOMS guidance for health-related research emphasizes examining the particular characteristics and circumstances that may make individuals or groups more likely to be wronged or experience additional harm. Vulnerability may arise because people have limited capacity to protect their own interests, but it can also arise from social circumstances such as marginalization, stigma, exclusion, limited resources, or environments in which other people may not adequately protect their interests.

The 2024 Declaration of Helsinki similarly recognizes, within medical research involving human participants, that individuals, groups, and communities may be in situations of vulnerability because of factors that can be fixed or contextual and dynamic.

This matters because two people who appear to belong to the same population may not face the same research risks. Conversely, someone who would not normally be described as vulnerable may become vulnerable in a particular research context because of dependency, crisis, power imbalance, stigma, or limited alternatives.

A hard-to-reach population is therefore not automatically vulnerable. A geographically dispersed professional group, for example, may be difficult to recruit without having any particular impairment in autonomy or increased susceptibility to research harm. Meanwhile, a readily accessible population can still be vulnerable to coercion or exploitation.

Ask why this population needs to be included

The first question should not be "How can I recruit them?" It should be "Why does my research question require them?"

Sometimes the justification is straightforward. A study of educational barriers experienced by students with a particular disability cannot simply substitute students without that disability because they are easier to recruit. Research concerning the experiences of incarcerated people cannot necessarily be answered by surveying the general population. A study intended to improve services for a marginalized community may require evidence from the people actually affected.

In other cases, researchers may target a population primarily because it is available, controllable, inexpensive, or dependent on the institution conducting the research. That requires much greater scrutiny.

The Belmont principle of justice is particularly relevant here. Participant selection should not place research burdens on particular groups simply because they are convenient or relatively easy to influence. Under the U.S. Common Rule, equitable selection of participants is also among the criteria for approval of covered research.

Scientific necessity therefore works in both directions. It can justify inclusion of a population requiring additional protection, but it can also expose situations in which that population is being recruited without a sufficiently good reason.

Exclusion can create ethical problems too

Protecting people from research should not quietly become protecting research from people who require more careful inclusion.

Historically, groups perceived as vulnerable have sometimes been excluded from research. That may reduce certain immediate research risks, but systematic exclusion can create evidence gaps. Interventions, policies, technologies, or services may then be developed using evidence from populations who differ from the people expected to use or experience them.

The 2024 Declaration of Helsinki explicitly recognizes this tension in medical research. It notes that default exclusion of people in situations of vulnerability has contributed to gaps in knowledge and potentially to disparities, and it calls for the harms of exclusion to be considered alongside the harms of inclusion.

That principle is useful more broadly even where the Declaration itself does not govern the research. Ethical protection should not be reduced to automatic exclusion.

Vulnerability may come from power rather than personal characteristics

Researchers should pay particular attention to relationships in which one party controls something the potential participant values.

A professor recruiting their own students, a supervisor recruiting employees, a physician recruiting patients, a prison authority facilitating access to prisoners, or a service provider recruiting people who depend on that service can create pressures that would not exist between strangers.

The participant may reasonably wonder whether refusal could affect grades, employment, treatment, privileges, services, recommendations, or relationships. Even when the researcher has no intention of imposing consequences, the participant's perception matters.

The Belmont Report distinguishes coercion from undue influence and recognizes that authority relationships can create unjustifiable pressure. Under the U.S. Common Rule, consent must also be sought under circumstances that minimize the possibility of coercion or undue influence.

Researchers should therefore examine who approaches participants, who explains the study, who knows whether they participate, and whether the person recruiting them holds authority over them.

Incentives need to be interpreted in context

Compensation for participants' time, inconvenience, expenses, or other contributions is not inherently unethical. The ethical question concerns how an offer functions in the circumstances of the prospective participant.

An amount that seems modest to a researcher may carry very different significance for someone experiencing severe economic hardship. Conversely, automatically reducing compensation for economically disadvantaged participants can itself be unfair if it undervalues their time.

There is no universal monetary threshold at which an incentive becomes undue influence. OHRP guidance notes that institutional review boards must assess the circumstances and consider whether incentives could interfere with voluntary informed consent.

The relevant questions include what the payment is for, whether it is proportionate to participation, how and when it is provided, whether participants lose earned compensation if they withdraw, and whether the offer could distort decision-making in the particular context.

Hard-to-reach populations can create methodological problems before ethical ones

Difficulty reaching a population affects much more than recruitment speed.

If researchers can access only the most visible, connected, service-engaged, digitally active, or institutionally recognized members of a population, the resulting sample may systematically differ from those who remain unreachable.

Suppose a study seeks to understand barriers faced by homeless young adults but recruits exclusively through one support organization. People who use that service may have experiences quite different from people who avoid or cannot access formal services. The recruitment strategy therefore shapes the population about which the study can credibly make claims.

Researchers should ask who their access route systematically misses. A large sample does not repair a sampling frame that excludes important segments of the target population.

Gatekeepers can provide access and simultaneously shape the sample

Hard-to-reach populations are often accessed through organizations, community leaders, clinics, schools, charities, online communities, local authorities, advocacy groups, or other gatekeepers.

Gatekeepers can be invaluable. They may help researchers understand the setting, communicate appropriately, identify safety concerns, and establish legitimate access. But gatekeepers also have interests and relationships of their own.

A gatekeeper may refer only cooperative members, exclude critics, discourage participation by people perceived as troublesome, or create an impression that participation is expected. Community permission may also be important in some contexts without replacing the individual's own consent where individual consent is required.

Researchers should therefore ask what authority the gatekeeper legitimately holds, what information the gatekeeper will receive, whether potential participants can decline privately, and how gatekeeper involvement may affect who enters the sample.

Community engagement can improve research, but it is not a substitute for participant protection

Research involving marginalized or historically exploited communities may benefit from meaningful engagement before recruitment begins. Community representatives may identify risks, cultural assumptions, inappropriate terminology, inaccessible procedures, or consequences that outsiders would miss.

CIOMS guidance gives substantial attention to community engagement, particularly for health-related research, and emphasizes that engagement should be meaningful rather than merely symbolic.

But "the community agreed" does not automatically establish that each individual's participation is voluntary or that the study is ethically acceptable. Communities are rarely homogeneous, and a community leader does not necessarily speak for every member.

Engagement can strengthen the research relationship. It does not transfer the researcher's ethical responsibilities to the community.

Consent procedures may need adaptation without lowering the standard of consent

Some populations may encounter language barriers, limited literacy, unfamiliarity with research, communication disabilities, impaired decision-making capacity, or cultural differences in how formal documents are understood.

The solution is not simply to provide the standard consent form and obtain a signature.

Researchers may need accessible language, translation, interpretation, alternative formats, additional explanation, comprehension checks, communication supports, or legally authorized representatives where the applicable framework permits or requires them.

The precise requirements depend on the research context and jurisdiction. The underlying objective is that participants, or an appropriate representative where applicable, receive information in a form that supports a meaningful decision.

Limited literacy, for example, should not be confused with limited capacity to decide. Nor should a signature be treated as proof that understanding occurred.

Decision-making capacity can be specific and dynamic

Some studies involve people whose capacity to consent may be impaired, fluctuate, or change during the study. Capacity should not be assumed absent merely because someone has a diagnosis, advanced age, disability, or other characteristic.

Where decision-making capacity is genuinely limited, the ethical and legal requirements become more complex. Researchers may need to determine whether the study can be conducted with people capable of providing consent instead, whether participation is responsive to the needs of the population concerned, what role a legally authorized representative may have, and how the participant's own preferences or objections will be respected.

CIOMS specifically cautions against treating entire classes of people as vulnerable and instead encourages attention to the characteristics that create vulnerability in the particular context.

Researchers should follow the applicable legal and institutional rules because standards for capacity, permission, assent, and legally authorized representation vary across jurisdictions and populations.

Recruitment methods can reveal the very characteristic you are trying to protect

For stigmatized, criminalized, or otherwise sensitive populations, simply identifying someone as eligible for a study may reveal private information.

An email with an explicit subject line, a telephone call answered by another household member, a recruitment poster placed in a revealing location, a visible meeting with a researcher, or a message sent through an employer can unintentionally disclose group membership or sensitive circumstances.

This is where vulnerability and sensitivity often intersect. Researchers need to consider whether recruitment and participation themselves create unnecessary risk, not merely whether the eventual dataset is secure.

Confidentiality may be particularly difficult in small or distinctive populations

A small population can make indirect identification surprisingly easy.

A researcher may remove every name from a report yet describe a "female principal in her early 40s at the only rural secondary school in the district." Anyone familiar with the setting may know exactly who that is.

The same problem arises with rare diagnoses, specialized occupations, small organizations, distinctive migration histories, unusual family structures, or detailed qualitative quotations.

Researchers should therefore assess identifiability in relation to what outsiders or insiders could infer from combinations of information. If the protections necessary to prevent identification remove variables essential to the analysis, privacy and confidentiality requirements may begin to affect the feasibility of the research question.

Accessibility is part of research design, not merely recruitment logistics

Researchers can inadvertently exclude people because the study itself is inaccessible.

Online-only recruitment can exclude people without reliable internet access. Written questionnaires can exclude participants who cannot use the format comfortably. A research site requiring long-distance travel can exclude people with mobility, financial, caregiving, or transportation constraints. Interviews offered only during working hours can systematically exclude certain workers.

These decisions affect both fairness and the evidence produced.

Researchers should therefore consider whether participation procedures unnecessarily favor certain segments of the target population. Where practical and scientifically appropriate, alternative formats, locations, communication methods, scheduling arrangements, or accessibility supports may improve inclusion.

Researcher safety and competence matter too

Hard-to-reach research can sometimes place researchers in unfamiliar or hazardous environments. Fieldwork may involve remote locations, volatile settings, illegal activities, violence, infectious disease exposure, or participants experiencing acute crises.

Protecting participants does not require ignoring researcher safety. Institutions may have occupational safety, fieldwork, safeguarding, insurance, travel, or security requirements separate from research ethics review.

Competence is equally important. A researcher conducting trauma-related interviews, assessing impaired decision-making capacity, working through interpreters, or entering culturally unfamiliar settings may need training or specialist support. Good intentions are not a methodology.

Feasibility includes whether you can reach the population ethically and well enough

A research question may be scientifically interesting yet infeasible because there is no defensible way to recruit an adequate sample, establish meaningful consent, protect participants, or obtain the necessary evidence.

This should be considered before the research question is finalized, particularly when the population is central to the question.

Difficulty does not automatically mean the population should be replaced. Instead, ask whether the barriers can be addressed through collaboration, a different sampling strategy, additional safeguards, revised procedures, more realistic scope, or a different method.

If they cannot, then the research question may need to change. A theoretically perfect population that you cannot responsibly access is not, unfortunately, made more recruitable by an impressive conceptual framework.

How Population Choice Can Change Both the Ethics and the Evidence

Notice that "hard to reach" and "vulnerable" produce different questions in this scenario. Recruitment difficulty affects who enters the sample and therefore what can be inferred. Vulnerability affects whether participation could expose people to harm or undermine voluntary decision-making.

A strong design has to address both.

Common Mistakes When Researching Vulnerable or Hard-to-Reach Populations

Decide Whether You Can Include the Population Fairly, Safely, and Scientifically

Before committing to research with a population that may be vulnerable or difficult to reach, map three things separately: why the population is necessary, what makes participants potentially vulnerable in this particular study, and what makes the population difficult to access.

That separation prevents a common mistake. A recruitment problem should not automatically be treated as an ethics problem, while an ethical vulnerability should not be reduced to a recruitment inconvenience.

You should also be realistic about resources. Ethical inclusion may require interpreters, accessible materials, community engagement, specialized training, additional recruitment time, travel, multiple recruitment channels, secure data procedures, or expertise your current project does not have.

A study can therefore be ethically possible in principle but infeasible for a particular researcher, thesis timeline, budget, or institution. That distinction matters. Choosing a more manageable question because you cannot provide the protections and resources the original study requires can be a responsible feasibility decision rather than unjustified exclusion.

Where the research remains feasible, however, the objective is not to make participants easier for the study to handle. It is to design a study capable of including them without exploiting the very circumstances that made their experiences worth understanding.

Before Planning Research With a Vulnerable or Hard-to-Reach Population

Frequently Asked Questions About Vulnerable and Hard-to-Reach Research Populations

Protect People Without Making Them Invisible to Research

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.