When Should Recruitment Difficulties Lead to Changes in Research Design Rather Than Extended Recruitment?

Should You Keep Recruiting Participants, or Is It Time to Reconsider the Study?

Suppose you are conducting a quantitative study examining whether access to digital learning resources is associated with academic persistence among adults returning to higher education. Your original sample size calculation indicates that approximately 300 participants are needed under the assumptions of your planned analysis.

After three months of recruitment, only 58 eligible participants have enrolled. You have distributed additional invitations, contacted community organizations, and extended the survey deadline. Recruitment has improved slightly, but the remaining target appears increasingly difficult to achieve.

A colleague suggests extending recruitment for another six months. Another recommends reducing the required sample size to whatever number you can obtain. A third proposes changing the research design entirely.

Which response is methodologically defensible?

The answer depends on why recruitment is failing and whether additional recruitment can realistically provide the evidence needed to answer the original research question.

Recruitment difficulty is not merely an administrative inconvenience. It may reveal problems with population accessibility, sampling assumptions, participant burden, eligibility criteria, statistical precision, or the feasibility of the proposed design.

Researchers therefore need to distinguish problems that can reasonably be resolved through additional recruitment from problems requiring a substantive reconsideration of the study.

Change the Design When More Recruitment Is Unlikely to Resolve the Underlying Methodological Problem

How to Decide Whether Recruitment Problems Require a Different Research Design

Why should recruitment feasibility influence research design?

A research design must be capable of producing evidence sufficient to address its intended question.

Recruitment feasibility affects whether the study can obtain the required observations, represent relevant population segments, and achieve the planned statistical precision or analytical depth.

For example, a study requiring 300 participants may be impractical when only a small fraction of the eligible population can be contacted during the available period.

However, a small sample does not automatically invalidate a study. Its adequacy depends on the research question, design, estimand, expected effect, and analytical approach.

The methodological problem arises when the achievable sample cannot support the conclusions the study intends to make.

Recruitment planning should therefore be considered part of research design rather than a task that begins only after the protocol is finalized.

What is the difference between a temporary recruitment delay and a structural feasibility problem?

A temporary recruitment delay occurs when enrollment is slower than expected but the underlying sampling strategy remains feasible.

For example, recruitment may slow during an academic examination period or because an institutional approval takes longer than anticipated.

A structural feasibility problem occurs when the original recruitment assumptions are unlikely to hold even after reasonable procedural improvements.

Examples include a substantially smaller eligible population than anticipated, inaccessible sampling frames, persistent nonresponse among essential population groups, or eligibility requirements that exclude most otherwise relevant individuals.

The distinction matters because additional time may resolve a temporary delay without addressing a structural limitation.

Researchers should identify the mechanism responsible for low enrollment before choosing between extension and redesign.

What recruitment indicators should researchers monitor?

Recruitment monitoring should distinguish the number of individuals approached from the number who are eligible, consent, and provide usable data.

These stages reveal different feasibility problems.

Researchers should use these indicators together. A low enrollment count alone does not reveal whether the main problem is contact, eligibility, willingness, or data completion.

How can researchers determine whether recruitment is realistically achievable?

A useful starting point is to compare the remaining recruitment target with the observed enrollment rate and the time available.

For example, a study that has recruited 60 participants over three months has averaged 20 participants per month.

If another 240 participants are required, a simple projection suggests approximately twelve additional months at the same rate.

However, this calculation is only a preliminary feasibility estimate.

Recruitment rates may change because of seasonal conditions, recruitment-channel exhaustion, new partnerships, or changes in eligibility.

Researchers should examine recent enrollment trends and the number of remaining accessible eligible individuals rather than assume that the initial rate will continue indefinitely.

When is extending recruitment appropriate?

Recruitment extension may be appropriate when the original research design remains scientifically valid and the remaining participants can realistically be enrolled.

For example, an additional recruitment site may have already agreed to participate, with a credible pool of eligible individuals.

Extension may also be reasonable when a temporary administrative delay prevented recruitment from beginning on schedule.

However, researchers should consider whether the extension changes the study population, observation period, or contextual conditions.

In longitudinal research, extended recruitment may introduce differences between early and late cohorts.

The decision should therefore consider scientific comparability as well as operational feasibility.

When does continued recruitment become methodologically unproductive?

Continued recruitment may become unproductive when additional effort repeatedly reaches the same limited population segments without meaningfully improving sample adequacy.

For example, repeatedly posting an online survey in similar social media groups may generate additional responses while continuing to exclude people with limited internet access.

Recruitment numbers may increase, but the underlying coverage problem remains.

Similarly, extending recruitment for several months may provide little benefit when the eligible population is smaller than originally assumed.

Researchers should assess the marginal contribution of additional recruitment to the study's scientific objectives.

The relevant question is not simply whether more participants can be recruited, but whether those participants improve the evidence needed for the intended inference.

How can recruitment difficulties reveal problems with the target population definition?

Low recruitment may indicate that the original target population was defined too broadly, too narrowly, or using characteristics that are difficult to establish.

For example, researchers may define a population as all adults who discontinued university education within a specified period but discover that no accessible records identify them.

The problem may concern the sampling frame rather than the actual number of eligible individuals.

Researchers should distinguish the target population, accessible population, and achieved sample.

Redefining the target population may improve feasibility, but it can also change the research question and generalizability.

Such changes should be treated as substantive design decisions rather than administrative corrections.

Should researchers relax eligibility criteria to increase recruitment?

Eligibility criteria should be reconsidered when they impose unnecessary barriers or exclude individuals who are substantively relevant to the research question.

For example, requiring an official academic transcript may exclude adults who previously attended university but cannot obtain their records.

Alternative verification may improve inclusion without changing the substantive eligibility criterion.

However, removing a scientifically necessary criterion can alter the target population and the estimand.

Researchers should therefore distinguish changing verification procedures from changing who qualifies for the study.

See eligibility verification when participants lack documentation and balancing eligibility rigor with inclusion.

Can recruitment problems create selection bias?

Yes. Recruitment difficulties may produce systematic differences between individuals who participate and those who remain unreached or decline.

For example, a study of digital learning barriers may disproportionately recruit people with reliable internet access because recruitment occurs exclusively online.

Extending the same online recruitment process may increase sample size without resolving that imbalance.

Selection bias is particularly concerning when inclusion depends on characteristics related to the exposure, outcome, or relationship under investigation.

Researchers should distinguish inadequate sample size from inadequate sample composition.

A larger sample does not automatically correct a biased selection mechanism.

Can changing recruitment channels preserve the original research design?

Sometimes. Adding recruitment channels may improve access while retaining the original population, eligibility criteria, outcomes, and analytical objectives.

For example, a researcher may supplement online recruitment with community-based invitations to reach individuals who rarely use social media.

However, different recruitment channels may introduce different selection mechanisms.

Researchers should document channel-specific eligibility, participation, and sample characteristics.

They should also assess whether the new channels change the sampling design or require amendments to the research protocol.

For related considerations, see community recruitment and selection bias and digital recruitment for hard-to-reach populations.

What if the original sample size requirement cannot be met?

Researchers should revisit the assumptions and inferential objectives underlying the sample size justification.

For example, a power calculation may have assumed an effect size larger than the effect considered scientifically meaningful, or a precision calculation may have used an unnecessarily narrow target interval.

However, revising assumptions simply to make the achieved sample appear sufficient is not defensible.

Researchers should distinguish justified corrections to the original planning assumptions from retrospective adjustments motivated by recruitment failure.

Where the attainable sample remains inadequate, the study may need to adopt a more limited objective, alternative design, or explicitly exploratory analysis.

These decisions should be made transparently and preferably without examining outcome relationships that could influence the choice.

Should researchers calculate post hoc statistical power?

Observed or post hoc power calculated using the effect estimated from the completed study generally provides little additional information beyond the observed test statistic and p-value.

It should not be used to declare an under-recruited study adequate merely because the observed effect happened to be large.

Researchers should instead examine the precision of effect estimates, uncertainty intervals, and the range of scientifically plausible effects compatible with the data.

Prospective power or precision calculations may still be useful when considering a revised design, provided the assumptions are independently justified.

Sample size adequacy should be evaluated against the intended scientific question rather than retrospectively justified by the observed results.

Can the research question be narrowed?

Yes. Narrowing the research question may be appropriate when the original objective requires information that the achievable sample cannot provide.

For example, a study initially designed to estimate differences across several demographic subgroups may lack sufficient observations for reliable subgroup comparisons.

The researcher might instead focus on a prespecified primary association or a more limited descriptive objective.

However, narrowing the question after inspecting results can create selective reporting concerns.

Researchers should document the timing and rationale for changes and distinguish confirmatory analyses from exploratory analyses.

A narrower but answerable question may be scientifically preferable to a broad question supported by inadequate evidence.

Should a quantitative study be converted into qualitative research?

Not automatically. Qualitative research is not a fallback method for obtaining a smaller sample when quantitative recruitment becomes difficult.

Qualitative and quantitative designs answer different kinds of questions and require different sampling, data collection, and analytical procedures.

For example, a quantitative study intended to estimate the prevalence of educational barriers cannot simply become a qualitative study without changing its objective.

However, if the emerging feasibility evidence indicates that understanding experiences and mechanisms is more achievable and scientifically valuable, a separately justified qualitative design may be appropriate.

Such a change should involve a revised research question, sampling rationale, instruments, analysis plan, and relevant approvals.

The methodological justification should be substantive rather than based solely on the belief that qualitative research requires fewer participants.

Can a mixed-methods design address recruitment difficulties?

Mixed-methods research may be appropriate when integrating quantitative and qualitative evidence helps answer a clearly defined research question.

For example, a limited quantitative survey could describe observed patterns while interviews investigate why some eligible individuals remain difficult to recruit.

However, adding interviews does not automatically repair selection bias in the quantitative sample.

Nor does qualitative evidence necessarily provide the missing statistical precision required for a population estimate.

Researchers should specify the contribution of each component and explain how integration supports the revised objective.

Mixed methods should be selected because the combined evidence is substantively useful, not simply because the original sample size target was missed.

Can a different quantitative design reduce recruitment requirements?

Potentially. Alternative designs may use available information more efficiently, depending on the research question.

For example, repeated-measures designs may improve precision for certain within-person comparisons when measurements are sufficiently reliable and appropriately modeled.

Case-control designs may be efficient for studying rare outcomes, while cohort designs may be advantageous for particular exposure structures.

However, changing designs can alter the estimand, introduce different biases, or require information not available in the original study.

Researchers should not assume that a more complex design automatically compensates for fewer participants.

Any proposed alternative requires its own feasibility and validity assessment.

Can Bayesian analysis solve a small-sample recruitment problem?

Bayesian methods can incorporate prior information and express uncertainty through posterior distributions.

In some settings, informative and defensible prior evidence may improve estimation when new data are limited.

However, Bayesian analysis does not create information that was never collected, eliminate selection bias, or guarantee precise conclusions from an inadequate sample.

Results may be sensitive to prior assumptions, particularly when the observed data are sparse.

Researchers should assess whether prior information is appropriate, transparent, and relevant to the intended estimand.

See Bayesian research designs for small samples for related considerations.

Should researchers stop recruitment when the target becomes unattainable?

Stopping recruitment may be appropriate when credible evidence indicates that continuing would impose disproportionate costs or burdens without producing sufficient scientific value.

However, stopping should follow an explicit assessment of the study's objectives, available resources, participant commitments, and possible alternatives.

For example, terminating a study after recruiting a small number of participants may be preferable to extending recruitment indefinitely when the target population is demonstrably inaccessible.

Researchers should consider whether existing data can support a narrower descriptive or exploratory objective.

Any change in study status or analysis should be documented and communicated according to applicable institutional and ethical requirements.

Stopping recruitment does not automatically require discarding all previously collected information.

What happens to data already collected if the design changes?

Existing data may remain useful, but their suitability depends on the revised question and design.

For example, survey responses collected under the original protocol may support descriptive analysis of the recruited sample even if the intended population comparison is no longer feasible.

However, researchers should not assume that data collected for one design automatically satisfy the measurement requirements of another.

Consent conditions, data-use permissions, measurement comparability, and recruitment differences must be considered.

When the study changes substantially, researchers may need additional approval before using previously collected data for the revised purpose.

The final report should distinguish original and revised objectives and explain which data support each analysis.

How can researchers avoid outcome-driven redesign?

Research design changes become particularly problematic when decisions are influenced by knowledge of the observed effects or statistical significance.

For example, a researcher might stop recruiting after an initially significant result or change the primary outcome because another variable appears more favorable.

Such decisions can distort error rates and undermine the interpretation of confirmatory analyses.

Where feasible, recruitment feasibility should be evaluated using information such as enrollment rates, eligibility, missingness, and operational costs without examining treatment effects or outcome associations.

When outcome-informed adaptations are scientifically justified, the design and analysis must explicitly account for them.

Researchers should document decisions, distinguish planned adaptations from post hoc changes, and avoid presenting exploratory findings as though they were prespecified.

What is an internal pilot or feasibility study?

A feasibility study evaluates whether key research procedures can operate as intended, including recruitment, eligibility screening, measurement, retention, and intervention delivery.

An external pilot is generally conducted separately from the definitive study, whereas an internal pilot forms part of a larger study under an appropriate design.

Internal pilot data may contribute to the final analysis when the protocol and methodological assumptions permit.

These approaches can help researchers identify recruitment problems before committing all available resources.

However, pilot studies should not be treated as miniature definitive studies whose primary purpose is to establish intervention effectiveness.

Feasibility objectives and progression criteria should be established clearly.

Should recruitment progression criteria be prespecified?

Yes, where feasible. Prespecified progression criteria can make decisions about continuation, modification, or termination more transparent.

For example, a pilot study may establish recruitment-rate ranges that indicate whether the proposed definitive study is feasible, requires modification, or is unlikely to proceed successfully.

However, numerical thresholds should be justified by the actual study context rather than copied mechanically from unrelated projects.

Progression criteria may consider recruitment, retention, data completeness, and implementation feasibility together.

Researchers should also allow for documented contextual judgment when unexpected circumstances arise.

The objective is to reduce arbitrary decisions, not to replace methodological reasoning with rigid administrative rules.

How should researchers decide between extension and redesign?

The decision should consider whether the original study remains capable of answering its question within realistic constraints.

Recruitment extension is more defensible when additional eligible participants are accessible, the expected enrollment rate is credible, and the original sampling and analysis remain appropriate.

Redesign becomes more defensible when the attainable sample cannot support the intended inference, recruitment channels systematically exclude essential population groups, or the original design rests on assumptions contradicted by feasibility evidence.

Researchers should also consider whether the proposed redesign introduces new biases or changes the estimand.

There is no universal recruitment percentage at which redesign becomes mandatory.

The decision should follow a documented assessment of scientific adequacy, feasibility, and participant welfare.

How should research design changes be reported?

The final report should describe the original recruitment plan, observed recruitment difficulties, and the evidence supporting any changes.

Researchers should identify whether the modifications affected eligibility, sampling, outcomes, statistical analysis, or the target population.

The timing of changes matters, particularly whether decisions were made before or after examining substantive results.

Required protocol amendments, ethics approvals, and registry updates should be completed according to the study's applicable requirements.

Researchers should distinguish analyses that remain confirmatory from those that became exploratory because of redesign.

Transparent reporting allows readers to assess what the revised study can establish rather than assume that the original design was implemented unchanged.

Deciding Whether to Extend Recruitment for a Study of Adult Online Learners

The hypothetical example illustrates why a recruitment extension should not be the automatic response to low enrollment.

If the original study remains feasible through credible recruitment improvements, extension may be appropriate. If the required evidence cannot realistically be obtained, redesign may produce a more scientifically useful and ethically defensible study.

Common Misconceptions About Recruitment Failure and Research Redesign

Decide Whether Additional Recruitment Will Improve the Evidence, Not Just the Sample Count

Before extending recruitment, determine why enrollment is below expectations and whether the original research question remains answerable.

Then evaluate the feasibility of additional recruitment against the scientific requirements of the study.

The key distinction is between recruitment difficulty that can be resolved operationally and recruitment difficulty that reveals a fundamental mismatch between the research design and the available population.

Researchers should preserve the original question when feasible, but they should not continue an unattainable design merely because it was initially approved.

Before Extending Recruitment or Changing the Research Design

Frequently Asked Questions About Recruitment Difficulties and Research Design Changes

Extend Recruitment When the Original Study Remains Feasible; Redesign When It Does Not

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.