01 · The Question
What Happens When the Sample You Planned Is No Longer Attainable?
You calculated that the study needed 300 participants. Recruitment slows, the accessible population is exhausted, funding approaches its end, or the study reaches a practical deadline. You have 218 participants and little realistic prospect of reaching 300.
What now?
Researchers sometimes treat the planned sample size as a pass-or-fail threshold: reach it and the study is valid; miss it and the study has failed. The reality is more nuanced. Falling short can reduce statistical power, precision, or the amount and diversity of information available, but the consequences depend on why the target was chosen, how far short the study falls, the analysis being conducted, and what claims the researchers intend to make.
The right question is therefore not simply whether you reached the planned number. It is what the achieved sample allows you to estimate, test, describe, or understand credibly.
03 · What You Need to Know
First Understand What the Planned Sample Size Was Supposed to Achieve
A Sample-Size Target Is Usually Based on Assumptions
A planned sample size is not a magical number that separates valid studies from invalid ones. It is generally derived from a set of design objectives and assumptions.
In a conventional hypothesis-testing study, a calculation may depend on the target effect or difference, expected variability or event rate, significance level, desired statistical power, allocation ratio, anticipated attrition, clustering, or other features of the design. Studies focused on estimation may instead determine sample size according to the precision desired for an estimate. Qualitative, pilot, feasibility, and other research designs use different rationales for determining how much data are needed.
CONSORT 2025 recommends reporting how the sample size was determined and the assumptions supporting the calculation. It also states that differences between the planned and actual sample sizes should be explained.
Consequently, when the target becomes unattainable, return to the rationale behind it. Ask what property of the evidence was expected at the planned sample size and how that property changes at the achieved size.
Falling Short Usually Changes Power or Precision Gradually, Not at a Cliff Edge
Suppose a study was designed for 300 participants but recruits 290. That does not suddenly transform an otherwise sound study into meaningless research. Conversely, recruiting 300 does not guarantee that every assumption behind the original calculation was correct.
For many quantitative designs, smaller samples produce less information. Estimates may become less precise, confidence intervals may become wider, and the probability of detecting a prespecified effect may decrease. How much these properties change depends on the design and analysis.
This is why the magnitude of the shortfall matters. Missing a target by 3% and missing it by 50% should not be treated as methodologically equivalent merely because both technically failed to reach the planned number.
Do Not Automatically Interpret a Non-Significant Result as “We Needed More Participants”
A common temptation appears after analysis. The study falls short of its planned sample, the primary result is not statistically significant, and the researchers conclude that the finding would probably have been significant if only the target had been reached.
That conclusion does not follow automatically.
A non-significant result can arise because the true effect is small or absent, because the estimate is imprecise, because the study has limited power, or for other reasons related to the design and data. The result itself does not tell you which explanation is correct.
Watch Out
Do not use a post hoc power calculation based on the observed effect as a way to explain away a non-significant result. CONSORT guidance specifically cautions that post hoc calculations of statistical power using trial results can mislead. Report the estimate and its uncertainty, explain the sample-size shortfall, and interpret the evidence accordingly.
Do Not Simply Recalculate the Target Until It Matches the Sample You Have
Another problematic response is to change assumptions after recruitment so that a new calculation conveniently produces the achieved sample size.
For example, a researcher who planned for a relatively modest effect might replace it with a much larger assumed effect solely because fewer participants were recruited. The arithmetic may produce a smaller required sample, but the scientific question has changed: the revised calculation now provides adequate power only for a larger effect.
A legitimate sample-size reassessment is different. Some study designs allow prespecified interim reassessment using updated information about parameters such as variability or event rates. CONSORT and SPIRIT recognize formal sample-size re-estimation and interim procedures, but these approaches require appropriate statistical methods, planning, documentation, and safeguards against data-dependent manipulation.
Methodological reassessment
Revises the sample-size requirement using a defensible method and clearly documented information, with appropriate procedures for the design.
Retroactive justification
Changes assumptions mainly to make the number already recruited appear to have been sufficient all along.
Ask Why the Target Cannot Be Reached
The cause of the shortfall affects the appropriate response. If recruitment is simply slower than expected but the population remains available, extending the recruitment period may be feasible. If one site is failing while others recruit successfully, adding or replacing appropriate sites may help. If eligible participants exist but frequently decline, the recruitment process may need attention.
Those are fundamentally different from exhausting the accessible population or discovering that the population is much smaller than originally estimated.
If recruitment is still ongoing, diagnose the underlying participant recruitment problem before concluding that the target itself must change.
A Smaller Sample Can Affect More Than Statistical Power
Sample size matters differently across research designs. In quantitative research, the consequences may include lower power, poorer precision, unstable estimates, difficulties fitting complex models, or insufficient observations within important subgroups. Clustered, longitudinal, multilevel, survival, and other designs introduce additional considerations.
For qualitative research, simply importing statistical notions of power is usually inappropriate. Adequacy depends on the methodological approach, research question, sampling strategy, heterogeneity of participants or cases, richness of the material, and analytic purpose. A shortfall from a planned number should therefore be interpreted according to the logic of that methodology rather than through a generic numerical rule.
Likewise, pilot and feasibility studies may use sample sizes chosen to estimate feasibility parameters or inform a future study rather than to test a definitive effectiveness hypothesis.
Recruiting a Different Population Just to Reach the Number Can Be Worse Than Missing the Target
Suppose you need 50 additional participants and can obtain them easily by broadening recruitment to a substantially different population. Numerically, this solves the sample-size problem. Methodologically, it may create a new one.
A larger sample is not automatically better if it no longer represents the population relevant to the research question. Changing eligibility criteria, settings, or sampling procedures may alter the meaning of the resulting evidence.
If the proposed solution changes who can enter the study, consider whether changing the research sample turns the project into a different study rather than treating the change as a simple recruitment fix.
Sometimes the Smaller Study Can Still Answer a More Limited Question
A study that cannot support its originally intended claim may still produce useful evidence, but the revised interpretation should emerge from the properties of the achieved data rather than from a desire to rescue the project.
For example, estimates may remain informative but substantially less precise than intended. Some planned subgroup analyses may no longer be credible because too few observations fall within each subgroup. A definitive study might need to be interpreted as inconclusive rather than as evidence that there is no effect.
The appropriate response is to narrow claims where necessary, not to make the data carry the same inferential weight originally expected from a larger sample.
Stopping Below Target Must Be Reported Transparently
If recruitment ends before the planned sample size is reached, readers should be able to determine both the intended and achieved sample sizes and why they differ. CONSORT guidance explicitly calls for an explanation when actual sample size differs from the originally intended size because of poor recruitment or revision of the target.
For observational studies, STROBE likewise emphasizes transparent reporting of how the study size was arrived at and enough information about participants and study conduct for readers to assess the evidence.
If the shortfall represents a departure from the protocol or original methodological plan, consider whether the methodological deviation needs to be reported and preserve a contemporaneous record of what happened.
06 · What This Means for You
Decide What the Achieved Sample Can Still Support
Once it becomes clear that the original target may not be reached, move from wishful recruitment projections to an explicit methodological assessment. Preserve the original sample-size rationale and examine the consequences of realistic alternatives.
A simple decision framework
If the target remains realistically attainable
Address recruitment barriers or consider a justified extension before changing the sample-size target.
If the original calculation used assumptions that can legitimately be reassessed
Use an appropriate statistical procedure and follow any prespecified, protocol, oversight, or blinding requirements rather than informally changing the assumptions.
If the final sample will be moderately smaller than planned
Quantify and report the implications for precision, power, planned analyses, and interpretation where appropriate to the design.
If some secondary or subgroup analyses are no longer supported
Do not force them simply because they appeared in the original plan; reconsider their role and interpret any exploratory analyses accordingly.
If the shortfall is so substantial that the original question cannot be answered credibly
Consider whether continuing, redesigning, or stopping the study is more defensible than producing an answer the design can no longer support.
A smaller sample may require more cautious conclusions, but caution should not become post hoc reinvention. If the response involves changing fundamental features of the study, evaluate whether the adaptation remains methodologically defensible.
There is also a point at which repeated concessions cease to be minor adjustments. If recruitment, measurement, sampling, and analysis have all moved away from the original plan, the broader question becomes how much methodological compromise the study can tolerate.
07 · A Quick Checklist
When the Planned Sample Size Looks Unattainable, Check These Points
Before deciding how to proceed, check:
Return to the original sample-size calculation or rationale and identify the assumptions behind the target.
Determine why recruitment is falling short and whether the problem can realistically be corrected.
Estimate the realistic final sample rather than continuing to rely on the original recruitment projection.
Assess how the smaller sample affects precision, power, model stability, subgroup analyses, or other relevant properties of the design.
Do not change sample-size assumptions merely to make the achieved sample appear adequate.
Check whether extending recruitment, changing sites, revising the target, or stopping recruitment requires protocol, ethics, registration, sponsor, or funder action.
Preserve the original target and document the achieved sample size and reason for any difference.
Adjust the interpretation and strength of conclusions to the evidence the achieved sample can actually provide.
