Is Underrepresentation Alone Enough to Justify a New Study?

You found an underrepresented population. Have you found a research problem?

Suppose your literature review shows that most previous studies involved one type of participant, while another relevant population appears rarely or not at all. It is tempting to write: "Population X is underrepresented in the literature; therefore, this study will investigate Population X."

The observation may be correct. The conclusion does not automatically follow.

Underrepresentation tells you something important about the distribution of existing evidence. What it does not tell you, by itself, is whether the missing representation creates an important uncertainty, whether existing findings are being applied beyond the populations that produced them, or whether studying the population would answer a question that matters. Those are the issues that turn a demographic gap into a defensible research problem.

Underrepresentation identifies a possible gap, not an automatic study justification

Move from "they were not studied" to "this is what we still do not know"

A population gap and a knowledge gap are related but different

A population gap describes who is missing or inadequately represented in a body of evidence. A knowledge gap identifies something important that the existing evidence cannot adequately establish.

The first can lead to the second, but it does not guarantee it. Imagine that a learning strategy has been studied in ten countries but never in an eleventh. That geographical absence is real. Yet "this country has never been studied" does not establish that another nearly identical study will contribute useful knowledge.

You need another step in the argument. Perhaps educational infrastructure differs in ways directly relevant to the intervention. Perhaps the language in which the intervention operates changes the task being studied. Perhaps policies, access conditions, or institutional arrangements make previous findings difficult to apply. Those propositions create researchable uncertainty.

Ask what claim the existing literature is making

One of the most useful ways to evaluate underrepresentation is to examine the scope of the conclusions being drawn from existing studies.

External validity concerns whether inferences from a study apply beyond the studied sample to a target population. Methodological work on generalizability and transportability therefore emphasizes defining the target population rather than treating a study result as universally applicable. A randomized study can provide a credible estimate for its participants while leaving uncertainty about the corresponding effect in another target population when relevant characteristics differ.

This gives underrepresentation considerably more significance when a population is expected to act on, receive, or be governed by conclusions derived largely from other populations. In that situation, the question is not merely whether everyone received equal representation. It is whether the evidence adequately supports the inference being made.

This is also why determining when a missing population limits confidence in existing evidence is more informative than simply counting how many studies included each group.

Not every difference between populations threatens generalizability

Researchers sometimes move too quickly in the opposite direction. After being told that underrepresentation alone is insufficient, they assemble a list of ways their proposed population differs from previous samples: culture, income, geography, language, age, educational system, technology use, and so forth.

Difference itself is still not the issue. The relevant question is whether a difference could matter to the phenomenon, intervention, measurement, mechanism, or outcome being studied.

In causal research, this issue is often discussed in terms of effect modifiers. If variables that modify an intervention's effect have different distributions in the study and target populations, the average effect observed in the study may not directly represent the target population's average effect. Generalizability and transportability methods have been developed precisely because the composition of the study and target populations can matter to the inference.

That does not mean you must demonstrate effect modification before proposing another study. It means your rationale should identify plausible pathways through which the population gap matters instead of treating demographic difference as self-explanatory.

You do not need to promise that the findings will be different

A particularly awkward justification is: "Previous research studied Population A, but my study will investigate Population B because the results may be different."

They may be. They may also be nearly identical.

If your study is worthwhile only if it produces a different result, you have built novelty around an outcome you do not yet know. Researchers should instead ask whether resolving the uncertainty is valuable regardless of whether the eventual findings converge with or diverge from previous evidence.

There are defensible reasons to obtain direct evidence without predicting a different effect. You may need to evaluate whether an intervention is accessible, acceptable, feasible, or implementable for the population. You may need population-specific estimates for decisions. You may be testing whether a previously observed relationship holds under materially different conditions. The reasoning behind whether you should expect a different finding before studying another population therefore deserves separate consideration.

Sometimes inclusion is valuable even when difference is not the hypothesis

Representation can have scientific and ethical significance beyond the search for heterogeneous effects. NIH policy in clinical research, for example, requires appropriate inclusion of women and members of racial and ethnic minority groups and explicitly connects inclusion to generalizability while also requiring applicable clinical trials to be designed to examine differences in outcomes. This is a policy for a particular research context, not a universal rule for every discipline, but it illustrates how inclusion and testing for difference are related without being identical objectives.

For other research contexts, direct evidence may matter because decisions affect a population that has historically received little opportunity to contribute to the evidence base. In such cases, inclusion can be important even without an expected different effect.

A new population does not automatically make an old question new

This is perhaps the most important test for a proposed population-based gap. Remove the population name from your research question. What intellectual problem remains?

If the proposal is essentially "the relationship between X and Y has been studied many times, but not among Population Z," you have established contextual novelty. Whether that novelty is sufficient depends on why Population Z changes what can be learned.

A stronger project may examine whether a proposed mechanism operates under different institutional conditions, whether an instrument retains its meaning across populations, whether implementation barriers alter an intervention's usefulness, or whether previous evidence can reasonably be extended to people who were systematically absent from it.

The goal is not to manufacture theoretical difference. It is to articulate the uncertainty honestly.

Underrepresentation may sometimes call for better inclusion rather than another separate study

There is another possibility researchers occasionally overlook: the appropriate response to underrepresentation may be to improve the next general study rather than create a parallel literature devoted to another subgroup.

Suppose a population is routinely excluded because researchers recruit from the same convenient locations. More inclusive sampling could address that problem. Depending on the question, oversampling, stratified recruitment, planned subgroup analysis, accessible study procedures, or disaggregated reporting may produce stronger evidence than repeatedly conducting separate studies.

A dedicated study is more compelling when the population has a question, context, mechanism, experience, or methodological requirement that cannot be adequately addressed as a small component of a broader design. That distinction is central to deciding when an underrepresented population needs its own study.

The stakes of getting the inference wrong should influence your judgment

Research gaps are not equally consequential. If existing evidence is used to allocate services, design educational systems, establish clinical practices, develop technologies, or make policy decisions, uncertainty about applicability may deserve more attention than it would for a low-stakes descriptive claim.

Transportability research makes this issue explicit by asking whether results obtained in one study population can support inference in a different target population. Recent methodological literature continues to treat differences between study and target populations as a substantive inferential problem rather than merely a demographic reporting issue.

The argument for another study can therefore become stronger when the population bears meaningful consequences from decisions based on evidence in which it was poorly represented.

Turning a weak population gap into a researchable justification

The statement accurately describes the literature, but it gives the reader no reason to believe the missing population changes what needs to be known.

Now the population matters for a reason. More importantly, the study can remain informative even if working adult students ultimately report outcomes similar to those found in previous research. Convergence would itself provide evidence about the applicability of earlier findings under the conditions examined.

There is no need to promise a dramatic population difference. Academia already has enough plot twists.

Common ways researchers overstate a population gap

Use underrepresentation as the beginning of the argument

When your literature review reveals an underrepresented population, do not discard the observation. Develop it.

Your justification should move from evidence about representation to evidence about uncertainty. Show who has been studied, identify the population to which the knowledge is relevant, determine what existing evidence cannot establish for that population, and explain why resolving that uncertainty matters.

A useful final test is simple: could you explain the contribution without using the phrase "has not been studied"? If you can identify a genuine uncertainty and explain how your design addresses it, the population gap has probably become a research problem. If you cannot, more conceptual work may be needed before data collection begins.

Before using underrepresentation to justify your study

Questions about using underrepresentation as a research gap

Underrepresentation is the observation; your research problem must explain why it matters

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.