When Does an Underrepresented Population Need Its Own Study?

When does missing representation become a research problem of its own?

You review the literature and notice something striking: a population that matters to your topic appears rarely, appears only in very small numbers, or is absent altogether. Perhaps most previous studies were conducted with university students from large cities, while you are interested in working adults in rural communities. Perhaps an intervention has been evaluated extensively, but people with disabilities were routinely excluded. Or perhaps a population appears in large datasets but is combined with other participants so that little can actually be learned about its experiences.

That looks like a research gap. But there is a harder question underneath it: does this population actually need a dedicated study?

The distinction matters. A population can be underrepresented in a literature without automatically generating a scientifically meaningful research question. At the same time, repeatedly studying only the populations that are easiest to recruit can leave important uncertainties hidden behind apparently mature evidence. The challenge is therefore not simply to count who has and has not been studied. It is to determine what we cannot adequately know because a population has been missing.

Underrepresentation matters when it creates a consequential evidence gap

The real question is what the existing evidence cannot tell you

Start by defining what you mean by underrepresented

"Underrepresented population" is a relational term. A group is not inherently underrepresented. It is underrepresented relative to something: the population affected by a problem, the population expected to use an intervention, the population to which conclusions are being applied, or the population whose experiences are relevant to the research question.

This distinction prevents a common conceptual shortcut. A population does not need to constitute a large percentage of the general population to matter scientifically. Conversely, a numerically large population may still be poorly represented in a particular body of research.

Representation can also be more complicated than whether participants were technically present. A population may have been included in previous samples but in numbers too small to support meaningful interpretation. It may have been absorbed into a broad category that obscures important variation. Researchers may also have collected demographic characteristics without asking questions that capture the population's particular circumstances.

Underrepresentation is evidence about the literature, not yet a justification for a study

Suppose you review 30 studies and discover that only two include the population you want to investigate. That is useful information. It establishes a representational gap. It does not yet establish why another study should be conducted.

The next question is causal in the broad scholarly sense: what does the missing representation prevent researchers, practitioners, policymakers, communities, or other relevant decision-makers from knowing?

If the answer is essentially "we have not studied them yet," the justification remains thin. This is why underrepresentation alone may not justify a new study. You need to connect the representational gap to a knowledge problem.

A dedicated study is easier to justify when existing evidence may not transfer well

Many studies implicitly or explicitly ask readers to generalize from a sample to a wider population. That inference becomes less secure when characteristics associated with the population could plausibly affect the phenomenon being studied.

The relevant differences need not be biological. They might involve language, access to technology, institutional treatment, educational opportunity, socioeconomic conditions, occupational context, cultural practices, environmental exposure, discrimination, infrastructure, or other features of the setting in which the phenomenon occurs.

Research on representation in health research illustrates the broader methodological principle. The REP-EQUITY framework, for example, treats representativeness in relation to the relevant population and research objectives rather than as a generic requirement that every sample mirror society. Its authors argue that systematic exclusion can limit generalizability and contribute to inequalities. Likewise, the SAGER guidelines ask researchers to consider whether sex and gender are relevant to the research question and, where appropriate, design studies capable of examining those factors.

The key word is relevant. You do not need to assert in advance that the population will produce a different result. Sometimes the uncertainty itself is important. The better question is whether there is a credible reason that the finding might differ or require qualification in another population, and whether the consequences of remaining uncertain are substantial enough to investigate.

The population may face a question that broader studies were never designed to answer

Sometimes generalizability is not the main issue at all. The population may experience a phenomenon in a way that creates a distinct research question.

Consider research on online learning. A large literature may already examine student satisfaction with online courses. A study involving students with visual impairments might not simply ask whether their average satisfaction score differs from everyone else's. It could investigate how learning-management-system accessibility, compatibility with assistive technologies, inaccessible assessment formats, or instructor practices shape participation. Those questions concern conditions that a generic satisfaction study may never have measured.

In such cases, the rationale is not "this group has been studied less." The rationale is that existing designs have not adequately investigated the phenomenon as it occurs for that population.

Representation and population-specific research are not the same thing

A broader study can sometimes solve the problem. If your only concern is that previous samples contained too few members of a relevant population, improved recruitment, stratified sampling, oversampling, or planned subgroup analysis may be more appropriate than creating an entirely separate study.

A dedicated study becomes more defensible when the research requires something broader studies are unlikely to provide: population-specific questions, tailored measures, sufficient depth, a particular sampling strategy, culturally or contextually appropriate procedures, or sustained attention to experiences that would otherwise remain peripheral.

You do not always need to predict a different effect

A common objection sounds reasonable: if there is no theoretical reason to expect a different result, why study the population separately?

That question deserves serious consideration, but it does not settle the issue. Research can be valuable for reasons other than demonstrating effect heterogeneity. Direct evidence can establish whether an intervention is feasible, acceptable, accessible, safe, interpretable, or implementable for a population. Inclusion can also matter when a group bears the consequences of decisions based on evidence from which it has repeatedly been absent.

In other words, inclusion may remain important even without an expected different effect. What matters is whether obtaining direct evidence serves a defensible scientific, practical, or ethical purpose.

Ask whether absence reduces confidence in the conclusion you want to make

One useful test is to write down the conclusion supported by the existing literature and then ask who that conclusion is intended to cover.

If researchers repeatedly make a broad claim such as "this intervention improves learning outcomes among university students," yet nearly all supporting studies involve a narrow subset of university students, the problem is not merely demographic imbalance. The scope of the claim may exceed the scope of the evidence.

This is an external-validity problem. A sample can support strong causal inference within the study while still leaving uncertainty about whether the result applies elsewhere. Determining when a missing population limits confidence in existing evidence therefore requires attention to both the intended target population and plausible sources of variation.

Consider consequences, not just uncertainty

Not every uncertainty deserves a new study. Research resources are finite, participants assume burdens, and repeatedly dividing populations into narrower categories can generate studies that are difficult to interpret.

The importance of an evidence gap partly depends on what happens if it remains unresolved. A small uncertainty about a low-stakes descriptive question may not justify a major population-specific project. The same degree of uncertainty may matter considerably more if the evidence informs healthcare, education, public policy, technology design, resource allocation, or access to services.

This is where scientific relevance and equity can intersect. Research may be warranted not because a population is statistically unusual but because decisions affecting that population are already being made using evidence that provides little direct information about them. In those circumstances, studying another population may improve fairness rather than merely add another subgroup to the literature.

Historical exclusion can strengthen the rationale, but it should not replace the research question

A history of exclusion is relevant. It can explain why an evidence gap exists, why conventional recruitment practices repeatedly fail to reach a population, and why the benefits and burdens of research have been distributed unevenly.

Still, historical exclusion should not become a generic sentence inserted into every proposal. A persuasive rationale explains how that history connects to the present research problem. What knowledge is missing? What decisions are affected? What aspects of the population's circumstances have been overlooked? What will the proposed research make possible that the existing literature cannot?

That connection is central when you need to justify research involving a historically excluded population.

Sometimes the case for a study changes how the study should be designed

If the rationale depends on a population's distinctive priorities, experiences, or barriers, researchers should be cautious about designing the entire project first and consulting the population afterward. A technically rigorous study can still ask the wrong question.

For some population-specific projects, early engagement can help determine whether the proposed problem is actually important, whether the terminology and categories make sense, whether recruitment procedures create unnecessary barriers, and whether the measures capture what researchers think they capture. The need for community involvement before study design will vary with the question and context, but it should be considered rather than assumed away.

From a demographic gap to a defensible research problem

Notice what changed. The study is no longer justified by a demographic sentence saying that "remote students are understudied." It is justified by an identifiable uncertainty linking the population, the intervention, the conditions under which it operates, and the claims researchers want to make.

There is another possible outcome. Suppose further investigation shows that connectivity, access, instructional conditions, and other theoretically relevant factors are essentially comparable, and the proposed study would simply administer the same questionnaire to another convenient sample. In that case, population underrepresentation alone may not provide a compelling reason for a separate study. The researcher may need a stronger question or a different design.

Weak arguments for population-specific research

Build the justification from the unanswered question, not the population label

If you are considering a study because a population is underrepresented, resist writing the justification immediately. First identify the chain of reasoning that connects representation to the research problem.

A useful starting point is to ask four questions: What evidence already exists? In what sense is the population inadequately represented? What important uncertainty follows from that absence? Why is a dedicated study an appropriate way to reduce that uncertainty?

Your final justification should therefore move beyond "Population X has received limited research attention." A stronger argument identifies what previous studies have established, who or what settings those findings represent, what remains uncertain for the proposed population, why that uncertainty matters, and how the proposed design addresses it.

There is also a language issue worth taking seriously. Population categories can be administratively convenient while remaining conceptually crude. Avoid implying that membership in a demographic category itself explains an outcome when the relevant mechanisms may instead involve institutions, resources, environments, discrimination, accessibility, policy, or other contextual conditions. Researchers should be particularly careful about defining a population primarily through deficit or problem.

Before proposing a dedicated population-specific study, check the logic

Questions about deciding when population-specific research is warranted

A population needs more than a blank space in the literature

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.