03 · What You Need to Know
When a Different Population Creates a Genuine Research Gap
Population Difference and Population Gap Are Not the Same Thing
A population is the group about which a study seeks to draw conclusions. Researchers can define populations using characteristics such as age, occupation, health status, educational stage, exposure, socioeconomic circumstances, geography, language, or many other attributes relevant to the question.
Because populations can be divided in countless ways, finding a group that has not appeared in an otherwise similar study is usually easy. The difficult part is demonstrating that the missing population creates an important limitation in existing knowledge.
Population novelty
No directly comparable study was identified in this particular group.
Population research gap
Existing evidence does not adequately answer the question for this group because relevant population differences create consequential uncertainty.
A useful population-gap argument therefore moves beyond representation alone and explains what researchers cannot safely conclude because the relevant population is missing or inadequately represented.
The Central Issue Is Whether Existing Evidence Applies
When research conducted in one group is used to inform conclusions about another, researchers face questions of external validity, generalizability, and applicability. These concepts are defined somewhat differently across methodological traditions, but all concern the limits of extending findings beyond the circumstances in which they were obtained.
Rothwell's influential discussion of external validity framed the practical question directly: to whom do the results of a study apply? Later methodological work has distinguished generalizability from applicability while emphasizing that evidence may become less direct when the population of interest differs meaningfully from the population studied.
The appropriate question is therefore not simply whether the participants are different. It is whether the difference is relevant to the relationship, intervention, exposure, mechanism, measurement, or outcome you are investigating.
Age Can Create a Meaningful Population Gap
Age is an obvious example of a population characteristic that can matter for some research questions. Biological development, disease prevalence, metabolism, cognition, social circumstances, exposure, service use, and responses to interventions can vary across the lifespan.
That does not mean every study involving adults must be repeated separately for children, middle-aged adults, and older adults. It means that age-specific evidence becomes important when age could plausibly change the answer or when the affected population has not been adequately represented.
NIH's Inclusion Across the Lifespan policy reflects this principle in the context of NIH-supported human-subjects research. It requires inclusion across age groups unless scientific or ethical reasons justify exclusion, with the stated purpose of helping ensure that knowledge gained from research is applicable to people affected by the disease or condition under study.
Sex, Race, and Ethnicity May Matter, but Categories Are Not Explanations
Other population characteristics can also be relevant. In U.S. clinical research, NIH inclusion policies require scientifically appropriate inclusion of women and members of racial and ethnic minority groups, with the goal of producing findings that can be generalized appropriately and, for certain clinical trials, examining whether important differences in intervention effects occur among groups.
These policies illustrate a broader research principle, not a universal rule that every demographic subgroup requires its own study. Researchers should examine whether a characteristic is relevant to the scientific question and whether existing evidence adequately represents the population affected.
Researchers should also avoid treating broad demographic categories as explanations by themselves. If you expect findings to differ, identify the plausible mechanisms or conditions involved rather than assuming that a label automatically causes a different outcome.
Health Status or Baseline Risk Can Change Applicability
Suppose an intervention has been studied extensively among generally healthy adults, but you want to know whether it benefits people with a particular chronic condition. The new population may differ in baseline risk, medication use, physiology, competing conditions, or ability to participate in the intervention.
Those differences could affect benefits, harms, feasibility, or interpretation. If the relevant population was excluded or inadequately represented in previous studies, additional evidence may be needed.
Here, the gap is not simply that “patients with Condition X have not been studied.” The stronger argument identifies which characteristics of Condition X make extrapolation uncertain and what information is therefore missing.
Occupation or Institutional Role Can Matter When It Changes the Exposure
Population differences are not limited to biological or demographic characteristics.
Imagine that workplace stress has been studied among office employees, while your proposed research concerns emergency responders. The occupational label itself is not enough to establish a gap. But emergency responders may experience different exposure patterns, schedules, risks, organizational structures, or demands directly relevant to the relationship being studied.
If those differences plausibly alter the phenomenon, studying the population can test an important limitation in the existing evidence.
Socioeconomic and Environmental Conditions Can Affect the Answer
Populations can also differ in resources, housing, education, income, infrastructure, environmental exposures, access to services, or other conditions that influence a research question.
For example, evidence about the effectiveness of an online intervention among populations with reliable internet access may not fully answer questions about its reach or implementation among populations with substantially different connectivity. The relevant difference is access to infrastructure, not merely the name of the population.
This is one reason population gaps can overlap with contextual gaps. The important characteristic may belong partly to individuals and partly to the environments in which they live or work.
Underrepresentation Can Limit What Researchers Can Conclude
A population does not have to be completely absent from the literature for a gap to exist. It may be present in numbers too small, or represented in ways that do not permit the question of interest to be answered adequately.
For example, a broad study may technically include older adults while enrolling very few of them. A clinical trial may include participants from several groups but lack enough relevant evidence to examine whether a clinically important difference exists.
NIH's current inclusion policies explicitly connect appropriate representation with producing evidence applicable to affected populations. For NIH-defined Phase 3 clinical trials, investigators must review prior evidence concerning whether clinically important differences in intervention effects by sex, race, or ethnicity may exist and plan analyses accordingly.
The broader lesson is that “included” and “adequately studied” are not necessarily the same thing.
A Missing Population Can Affect Measurement, Not Just Outcomes
Sometimes the main uncertainty concerns whether a measurement works appropriately in a new population.
A questionnaire developed among university students may not necessarily have the same interpretation among younger adolescents. A scale validated in people with one educational background may perform differently in another group. A diagnostic or predictive model may also behave differently when applied to populations whose characteristics differ from those used in its development.
If measurement validity or interpretation is uncertain, population-specific research may be justified even before researchers ask whether substantive relationships differ.
Different Populations Can Test Generalizability
Studying a different population can also contribute by testing the boundaries of an existing finding.
Suppose a relationship has been repeatedly observed among one type of participant. A study in a theoretically relevant different population can test whether the relationship is robust or whether it depends on particular conditions.
This is not merely “doing the same study again.” It can function as a test of generalizability or as a replication under meaningfully different conditions. The value comes from what the population difference allows researchers to learn.
In that situation, the proposed research may overlap with replication as a way of addressing uncertainty in existing evidence.
Population-Specific Decisions May Require Population-Specific Evidence
Sometimes the rationale for studying a population is practical rather than a prediction that a general scientific relationship will differ.
Decision-makers may need population-specific estimates of prevalence, risk, needs, service use, preferences, feasibility, harms, costs, or intervention uptake. Evidence from another population may not provide the quantities needed for that decision.
For example, a service planner deciding how much capacity is needed for a particular group requires information about that group's needs. An estimate from a substantially different population may be scientifically informative but insufficient for local planning.
The research rationale should state that decision need directly rather than claiming that the entire underlying phenomenon is unknown.
A Different Population Does Not Automatically Mean Existing Evidence Is Invalid
Researchers can make the opposite mistake by dismissing all previous evidence because its participants differ from their proposed population.
Evidence does not become irrelevant whenever a demographic characteristic changes. Some findings may generalize across broad populations, particularly when the relevant mechanisms and conditions are comparable.
Assess the differences that matter to the question. Do not assume either perfect transferability or complete non-transferability.
| Population difference |
When it may support a gap |
What you need to explain |
| Different age group |
Age plausibly affects the mechanism, response, risk, measurement, or applicability |
Why age is relevant to the particular research question |
| Different health status |
The condition may alter effects, harms, feasibility, or baseline risk |
Which characteristics make extrapolation uncertain |
| Different occupation |
Work conditions materially alter exposure or mechanisms |
Which occupational conditions affect the phenomenon |
| Different socioeconomic circumstances |
Resources or exposures could alter implementation or outcomes |
Which conditions matter and through what plausible pathway |
| Underrepresented demographic group |
Existing evidence cannot adequately establish applicability or relevant group differences |
What consequential uncertainty remains |
| Arbitrarily narrower subgroup |
Usually weak without further reasoning |
Why this subdivision changes the scientific or practical question |
Do Not Create Populations Solely to Manufacture Novelty
With enough demographic variables, almost any sample can be made unique. You might find that nobody has studied a specific relationship among first-year students aged 18–20 in one degree program at one institution.
That exact combination may indeed be absent from the literature. But unless those characteristics are relevant to the research question, the absence may tell you very little.
This is the same problem encountered when researchers claim that “no one has studied this before” as sufficient evidence of a gap. Novelty can be produced simply by making the population increasingly specific.
Watch Out
Do not define an extremely specific population merely because doing so produces a zero-result literature search. The population should follow from the research problem, theory, mechanism, evidence limitation, or decision need rather than being engineered to make the study appear original.
Country and Population Gaps Often Overlap
A country-based gap is frequently presented as a population gap because researchers want to study “people in Country X.” But country includes more than population characteristics. It can also involve different policies, institutions, infrastructure, environments, and systems.
If the relevant uncertainty arises from those broader contextual conditions, describe them. The stronger justification for studying the same topic in another country explains what about the context makes existing evidence insufficient.
A Population Gap Still Has to Matter
Even if a group is genuinely underrepresented and existing evidence does not completely resolve the question for that population, you still need to consider the importance of the uncertainty.
Would resolving it affect theory, understanding, clinical care, policy, professional practice, equity, implementation, resource allocation, or another meaningful decision? Can your proposed study produce evidence capable of reducing the uncertainty?
A population gap can be genuine without being a high-priority gap. After establishing that it exists, ask whether the research gap is actually worth filling.
