03 · What You Need to Know
How to Design Eligibility Criteria That Are Both Rigorous and Inclusive
What are inclusion and exclusion criteria?
Inclusion criteria specify the characteristics individuals must possess to participate in a study.
Exclusion criteria identify conditions that prevent participation even when some inclusion requirements are met.
For example, a study of adults who discontinued higher education may require prior enrollment in a qualifying program and current nonenrollment.
An intervention study may additionally exclude individuals for whom the experimental procedure presents a specific unacceptable risk.
Eligibility criteria should be connected to the research question, study design, outcome measurement, or participant protection.
They should not be added merely because narrower criteria appear more academically rigorous.
What does marginalized mean in the context of research inclusion?
Marginalization refers to social, institutional, economic, political, or structural processes that limit people's access to resources, opportunities, representation, or participation.
Depending on the context, relevant barriers may involve disability, socioeconomic disadvantage, geographic isolation, language access, insecure housing, or exclusion from formal institutions.
Marginalized populations are not necessarily homogeneous, and individuals may experience overlapping forms of disadvantage.
Researchers should avoid treating marginalization as an inherent deficiency or assuming that all members of a broadly defined group face identical barriers.
The relevant question is which structural and procedural conditions affect participation in the particular study.
Why can narrow eligibility criteria threaten external validity?
External validity concerns whether study findings apply beyond the participants and conditions directly examined.
Strict eligibility criteria may produce a sample that differs substantially from the population to which researchers intend to generalize.
For example, excluding adults without reliable internet access from a study of online education barriers may remove an important subgroup from the target population.
Even a carefully conducted study may then provide limited evidence about the experiences of digitally excluded adults.
However, broader eligibility does not automatically guarantee external validity. Recruitment coverage, participation, measurement, and contextual differences also matter.
Researchers should assess whether exclusions remove people for whom the findings are intended to be relevant.
Can restrictive eligibility criteria also affect internal validity?
Eligibility restrictions may sometimes improve internal validity by excluding conditions that make an intervention unsafe, prevent valid measurement, or create serious interpretive difficulties.
However, restrictions are not universally necessary for causal identification.
For example, excluding participants with different baseline characteristics does not automatically eliminate confounding in an observational study.
Likewise, a randomized experiment can sometimes include heterogeneous participants while preserving internal validity through appropriate randomization and analysis.
Researchers should distinguish restrictions that protect a valid comparison from those intended merely to reduce observed variation.
Heterogeneity can complicate estimation, but it may also provide valuable information about differences in intervention response.
How do scientifically necessary exclusions differ from administrative convenience?
| Proposed requirement |
Possible justification |
Question to examine |
| Relevant prior educational experience |
Defines the intended population. |
Is the criterion necessary to answer the research question? |
| Specific clinical safety condition |
Protects participants from a defined intervention risk. |
Is the exclusion supported by appropriate safety evidence? |
| Personal computer ownership |
May simplify online participation. |
Could an accessible alternative participation mode be provided? |
| Fluency in one language |
May reflect instrument availability. |
Can valid translation or language support be provided? |
| Official identification document |
May support eligibility verification. |
Is strong identity verification necessary for this study? |
| Ability to attend during business hours |
May simplify scheduling. |
Could flexible scheduling reduce unnecessary exclusion? |
A criterion may be justified in one study and unnecessary in another. Researchers should evaluate its function rather than classify restrictions as inherently appropriate or inappropriate.
Why does the research question determine eligibility?
The target population should be derived from the phenomenon or causal relationship being investigated.
For example, a study specifically examining the experiences of adults who have never used an online learning platform may reasonably exclude experienced platform users.
That exclusion is related to the research question.
By contrast, requiring those adults to own a personal laptop may be difficult to justify when device ownership is not essential to participation or the phenomenon being studied.
Researchers should formulate the target population before designing convenient recruitment procedures.
Otherwise, the accessible sample may gradually replace the intended population without an explicit methodological decision.
Should researchers include everyone to maximize diversity?
No. Maximum inclusion without reference to the research question can produce a poorly defined study population.
For example, a study of adults who discontinued higher education should not automatically include individuals who never enrolled in higher education merely to increase diversity.
Doing so could change the phenomenon being investigated.
Inclusive design requires that all eligible individuals have a fair and reasonable opportunity to participate, not that every person qualifies for every study.
Researchers should preserve the conceptual boundaries necessary for meaningful analysis.
Where several related populations are scientifically relevant, separate strata or explicitly defined subgroups may be more appropriate than removing all distinctions.
How can accessibility barriers become hidden exclusion criteria?
Some studies formally use broad eligibility criteria but impose participation procedures that effectively exclude particular individuals.
For example, an online questionnaire may technically permit adults with disabilities to participate while remaining incompatible with assistive technologies.
Similarly, a study may accept participants from remote communities but require attendance at a distant research site.
These procedural barriers can function like eligibility restrictions even when they do not appear in the written criteria.
Researchers should therefore examine the entire recruitment and data collection process, not only the eligibility statement.
Accessibility should be considered during design planning rather than addressed only after an eligible participant encounters a barrier.
How should language requirements be evaluated?
Language requirements may be necessary when the research depends on a particular linguistic competence or when valid measurement cannot otherwise be established.
However, excluding participants solely because the research team lacks translation resources can narrow the accessible population.
Researchers should assess whether translated materials, interpreters, or multilingual data collection are feasible and scientifically appropriate.
Translation is not merely a matter of replacing words. Instruments may require cultural adaptation and evidence that they measure comparable constructs across languages.
Where language restrictions remain necessary, the methodology should explain their rationale and implications for generalizability.
Researchers should avoid claiming applicability to linguistic groups that were systematically excluded without supporting evidence.
How should disability-related eligibility criteria be assessed?
Disability should not automatically be treated as a reason for exclusion.
Researchers should identify whether a specific impairment or support need is genuinely incompatible with the study's scientific requirements or creates an unacceptable risk.
For example, a participant who cannot use a standard mouse may still be able to complete a digital learning study using an accessible input device.
Providing an appropriate accommodation may preserve eligibility without changing the substantive research question.
However, adaptations may affect measurement comparability and should be evaluated accordingly.
Researchers should distinguish disability-related safety or measurement concerns from barriers created by inaccessible procedures.
The United Nations Convention on the Rights of Persons with Disabilities provides an important international rights framework, although specific legal obligations depend on jurisdiction.
What if participants lack formal documentation?
Documentation requirements can exclude eligible individuals who cannot obtain records or who reasonably hesitate to disclose identifying information.
Researchers should distinguish substantive eligibility from the evidence used to establish it.
For example, prior educational enrollment may be established through a structured self-report procedure in some low-risk studies, whereas other designs may require stronger verification.
The appropriate choice depends on classification accuracy, participant safety, and applicable requirements.
See eligibility verification when participants lack documentation for a focused examination of that decision.
How can socioeconomic disadvantage affect participation?
Participation may involve transportation costs, internet expenses, lost working time, childcare responsibilities, or other burdens.
For example, requiring an in-person interview during ordinary working hours may disproportionately exclude individuals with inflexible employment schedules.
Researchers should consider flexible scheduling, accessible participation modes, and proportionate reimbursement where appropriate.
However, incentives should not create undue influence or pressure participants to accept risks they would otherwise decline.
Reducing participation costs can improve access without necessarily changing the substantive eligibility criteria.
Researchers should document accommodations that materially affect measurement or participation conditions.
Can broad eligibility increase heterogeneity?
Yes. Broader eligibility may introduce greater variation in participant characteristics, baseline outcomes, or responses to an intervention.
This can complicate statistical precision and interpretation, particularly in small studies.
However, heterogeneity is not inherently a methodological defect.
If the research objective concerns a diverse real-world population, excluding variation may reduce the relevance of the findings.
Researchers should determine whether important differences can be measured and addressed through design or analysis.
For example, prespecified subgroup analyses or effect-modification assessments may be useful when adequately supported by sample size and theory.
Broad eligibility should be accompanied by appropriate analytical planning rather than an assumption that all participants will respond identically.
Should researchers use stratified recruitment?
Stratified recruitment may help ensure that important population subgroups are included in the study.
For example, researchers might establish recruitment targets for participants with different levels of digital access or geographic circumstances.
However, stratified recruitment is not necessarily the same as probability-based stratified sampling.
Recruitment quotas can improve subgroup coverage while leaving inclusion probabilities unknown.
Researchers should distinguish purposeful subgroup representation from statistically representative population sampling.
The sampling method and intended inference should be reported accurately.
Can community organizations support inclusive eligibility?
Community organizations may help researchers identify unnecessary screening barriers and communicate with populations that are difficult to reach through institutional channels.
For example, organizations supporting adult learners may explain why official records are difficult to obtain or why certain recruitment messages are ineffective.
However, organizations may represent only selected segments of a broader population.
Researchers should avoid assuming that one organization speaks for every marginalized individual.
Community involvement can improve accessibility and contextual understanding, but recruitment selection mechanisms still require evaluation.
See community recruitment and selection bias.
How can gatekeepers undermine inclusive recruitment?
Gatekeepers may control access to potential participants through institutional or community relationships.
For example, a school administrator may recommend only students considered cooperative or academically successful.
Such selection can exclude individuals with more substantial difficulties or unfavorable institutional experiences.
Researchers should establish transparent eligibility and invitation procedures that reduce unnecessary discretionary filtering.
However, legitimate privacy and safety responsibilities should still be respected.
The distinction is examined in managing gatekeepers in hard-to-reach participant recruitment.
How should researchers protect participants from coercion?
Marginalized individuals may depend on institutions, employers, service providers, or community organizations that facilitate recruitment.
Such relationships can create perceived pressure to participate even when no explicit coercion occurs.
Researchers should separate participation decisions from ordinary services, benefits, academic evaluation, or employment consequences.
For example, students should not reasonably believe that declining research participation will affect their grades.
Recruitment and consent procedures should communicate voluntariness clearly.
However, ethical inclusion does not mean encouraging participation at any cost. Participants must remain free to decline or withdraw according to the applicable consent arrangements.
How does inclusion relate to research ethics and justice?
The Belmont Report identifies justice as a fundamental principle concerning the fair distribution of research burdens and benefits.
Unjustified exclusion may prevent some populations from contributing to or benefiting from research evidence.
Conversely, including marginalized individuals without adequate safeguards can expose them to disproportionate burdens.
Ethical inclusion therefore requires more than increasing participant numbers.
Researchers should examine who is invited, who can realistically participate, what risks participation creates, and whose circumstances the findings may inform.
These questions should be considered alongside respect for persons and beneficence rather than treated as separate administrative concerns.
Should eligibility criteria be justified individually?
Yes. Each substantive inclusion or exclusion criterion should have a defensible scientific, ethical, or operational rationale.
For example, researchers might classify criteria according to whether they define the target population, protect safety, support valid measurement, or address a necessary design requirement.
Restrictions based solely on convenience should be reconsidered when they exclude relevant participants.
However, researchers should also examine interactions among criteria.
Several individually modest restrictions may collectively exclude a substantial portion of the intended population.
An eligibility review should therefore consider both individual justifications and their cumulative consequences.
Can eligibility criteria be changed during recruitment?
Changes may be appropriate when the original criteria create unintended exclusion or when new evidence shows that a restriction is unnecessary.
However, changing eligibility can alter the target population, sample composition, or comparability of participants recruited at different times.
Researchers should determine whether the change preserves the original research question and estimand.
Required ethics or protocol approvals should be obtained before implementing substantive changes.
The timing and rationale should be documented.
Researchers should not broaden eligibility retrospectively simply to obtain enough participants without considering the consequences for interpretation.
How should inclusive eligibility be evaluated before recruitment?
Researchers should review the proposed criteria against the intended population and identify groups likely to be excluded.
They should then determine whether each exclusion is necessary and whether accessible alternatives exist.
Consultation with relevant community representatives and accessibility specialists may identify barriers that the research team overlooked.
Where feasible, pilot screening procedures can reveal ambiguous questions or disproportionate documentation burdens.
However, consultation should not be treated as proof that all perspectives have been represented.
The final protocol should document the justification for restrictions and the procedures used to support equitable access.
How should eligibility and exclusion be reported?
Research reports should describe the inclusion and exclusion criteria, recruitment settings, screening procedures, and relevant changes during the study.
Where feasible, researchers should report the number of individuals screened and excluded, together with meaningful reasons for exclusion.
However, reporting should protect confidentiality, especially when small subgroup counts could identify participants.
Researchers should explain how eligibility restrictions affect the population to which findings may apply.
For observational studies, the STROBE reporting framework provides relevant guidance concerning participant selection and study populations.
Transparent reporting allows readers to assess whether the study's claims extend beyond the individuals who could realistically participate.
