How Should the Literature Change the Population You Plan to Study?

Are You Planning to Study the Population That Can Actually Answer the Research Question?

You may begin a project already knowing whom you want to study: university students, nurses, teachers, adolescents, older adults, employees, patients with a particular condition, or perhaps whoever is realistically accessible to you.

Then the literature complicates that choice.

You discover that previous studies concentrated heavily on one group. Findings differ by age or experience. A population routinely excluded from earlier research is directly affected by the problem. Evidence from one setting may not reasonably transfer to another. Or you learn that the population you planned to recruit is convenient but poorly aligned with the phenomenon you actually want to understand.

At that point, population selection is no longer merely a recruitment decision. The literature is telling you something about who needs to be studied for your research question to produce meaningful evidence.

Your Population Should Follow the Scientific Question, Not Just Access

What the Literature Can Tell You About Whom to Study

Start With the Population Your Question Is About

A useful distinction is the one between the population about which you ultimately want to make an inference and the participants you can realistically recruit.

Suppose your research question concerns burnout among hospital nurses, but the easiest participants available to you are nurses enrolled in a graduate program at one university. Those nurses may provide useful evidence, but accessibility does not automatically make them an adequate representation of the population your question implies.

The literature helps you identify which population characteristics may matter. Previous research might indicate that burnout differs according to clinical area, employment status, shift pattern, career stage, organizational conditions, or other factors relevant to the phenomenon. That does not mean every factor must appear in your eligibility criteria. It means you should understand which restrictions change the population to which your findings can reasonably speak.

The Literature May Show That Your Population Is Too Broad

A broad label can conceal substantively different groups.

“Teachers,” for example, may include early-childhood educators, elementary teachers, secondary teachers, vocational instructors, and university faculty working under very different professional conditions. “Older adults” can cover decades of life during which health, cognition, employment, caregiving, and technology use may differ considerably. Even apparently precise clinical populations can contain important variation in severity, treatment history, comorbidity, or disease subtype.

If the literature indicates that the phenomenon operates differently across these groups, combining them indiscriminately may produce an average that answers no particularly useful question.

Narrowing is therefore justified when a characteristic defines the phenomenon you intend to understand, marks an important boundary in existing evidence, or is necessary for a coherent test of the research question. Narrowing merely because a smaller group is easier to recruit is a different rationale and should not be disguised as a scientific one.

The Literature May Show That Your Population Is Too Narrow

Researchers can also inherit unnecessarily restrictive populations from previous studies.

Perhaps most studies included only younger adults even though the condition affects older people. Perhaps research on an educational intervention repeatedly sampled high-achieving students even though the intervention is intended for all learners. Perhaps clinical evidence excludes people with common comorbidities, leaving uncertainty about how well findings apply to the patients who would actually receive an intervention.

In clinical research, this concern is explicit in current policy. NIH's Inclusion Across the Lifespan policy requires inclusion of individuals of all ages in NIH-supported human-subjects research unless exclusion has a scientific or ethical justification, with the aim of generating knowledge applicable to populations affected by the condition under study. FDA guidance similarly encourages clinical-trial eligibility and enrollment practices that can produce study populations more representative of people likely to use the medical product.

Those are specific policies for particular research contexts, not universal rules for all disciplines. The broader lesson is still useful: exclusions should earn their place.

Underrepresentation Matters When It Creates an Important Evidence Problem

Discovering that a group is underrepresented in previous studies deserves attention, but the appropriate response requires thought.

Sometimes the rationale is inferential. Existing evidence may not establish whether findings apply to a population with meaningfully different characteristics. Sometimes it is practical: a group substantially affected by a policy, intervention, technology, or condition lacks evidence needed for decisions concerning them. Ethical considerations may also matter, particularly when groups bear the consequences of research-informed decisions while rarely being represented in the evidence used to make those decisions.

But “understudied” should not become another mechanical gap statement. Ask what uncertainty the missing population creates and what studying that population would allow you to learn.

A Different Population Does Not Automatically Create a New Study

“This relationship has been studied among university students, but not among senior high school students” identifies a population difference. It does not yet establish why the new study is necessary.

The change becomes more meaningful when developmental stage, institutional environment, curriculum, autonomy, technology access, social conditions, exposure, or another relevant characteristic could affect the phenomenon. The literature may provide theoretical or empirical grounds for expecting that difference to matter.

This is closely connected to the contribution you expect the study to make. If changing the population merely changes the address or age label without changing what can be learned, the resulting contribution may be limited.

Population Differences Can Limit Generalization Without Making Earlier Research Wrong

A study conducted in one population can be internally informative while having limited applicability elsewhere. These are separate issues.

Suppose an intervention has been tested rigorously among adults aged 18 to 40. The results do not become invalid because older adults were absent. Rather, the evidence may be insufficient to establish what happens among older adults, particularly if age could influence treatment response, risk, implementation, or another relevant feature.

The same reasoning applies outside health research. Findings from large urban universities may not automatically describe small rural institutions. Research involving experienced professionals may not characterize novices. Evidence from high-resource settings may not transfer straightforwardly to settings where implementation conditions differ.

Do not claim non-generalizability merely because populations are not identical. Identify which differences could plausibly matter and calibrate your inference accordingly.

Sometimes the Literature Justifies Restricting the Population

Broad inclusion is not always methodologically preferable.

If your question specifically concerns first-time mothers, novice teachers, adolescents with a defined condition, or employees during their first year of employment, restricting eligibility is integral to the question. A narrowly defined population can also reduce irrelevant heterogeneity when a study is intended to examine a process under specific conditions.

Inclusion policies themselves recognize this logic. NIH permits age-based exclusions when scientifically or ethically justified and requires the selected age range to be justified in relation to the scientific question.

The principle is not “include everyone.” It is “include and exclude deliberately.”

The Population You Need and the Population You Can Recruit May Not Match

The literature may reveal an ideal population that your resources cannot realistically reach.

You may lack access to multiple institutions, specialized clinical groups, geographically dispersed participants, or a sufficiently large subgroup for the analysis you hoped to conduct. This creates a design constraint, not permission to pretend that the accessible population answers the broader question equally well.

You may need to narrow the research question, redesign recruitment, collaborate with additional sites, reconsider the intended inference, or acknowledge the resulting limitation. What you should not do is quietly allow convenience to redefine the target population after the fact.

When the Literature Shows That the Convenient Population Is Not Enough

The literature has not proven that part-time instructors will adopt the technology differently. Instead, it has revealed a defensible reason why excluding them would leave an important question unresolved.

If adequate recruitment of the relevant groups is impossible, the appropriate response may be to narrow the research question rather than claim that a convenience sample represents faculty generally.

Common Mistakes When Choosing a Study Population From the Literature

Recheck Every Important Population Boundary Before Recruitment

Return to your inclusion and exclusion criteria after the literature review. For each consequential boundary, ask why it exists.

Population decisions may also expose weaknesses in the studies you are using as precedents. If the literature repeatedly excludes the people most affected by the phenomenon or uses samples poorly aligned with the claims being made, that may be one of the weaknesses your own study should deliberately avoid.

Finally, population and outcome decisions interact. Once you change whom you study, revisit what outcomes are appropriate to examine. An outcome meaningful, valid, or feasible in one population may not function identically in another.

Does Your Planned Population Still Make Sense?

Questions About Choosing a Study Population From the Literature

Study the Population Your Question Requires, Not Merely the One Within Reach

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.