01 · The Question
Can You Increase or Decrease Your Sample Size After Ethics Approval?
You planned to recruit 200 participants but recruitment is painfully slow. Could you stop at 150? Or perhaps an unexpectedly high response rate gives you the opposite idea: why not collect 300 responses while participants are available?
Sample size often feels like a statistical issue, but once a study involving human participants has been approved, it can also become an ethics issue. The number of participants may form part of the protocol reviewed by the ethics committee and can affect how many people are exposed to research procedures, burdens, or risks.
Changing that number therefore should not be treated as an automatic investigator-level decision simply because the statistical analysis can accommodate it.
03 · What You Need to Know
Why Sample Size Can Be an Ethical as Well as Statistical Question
Start With the Sample Size That Was Actually Approved
Before deciding that enrollment should change, check the current approved protocol. Some protocols specify an exact number. Others distinguish between the number expected to enroll, the number expected to complete the study, and a larger maximum enrollment ceiling designed to account for withdrawals, exclusions, or screen failures.
Those distinctions matter. If a protocol says that approximately 200 participants are expected but authorizes enrollment of up to 240 to account for attrition, enrolling participant 205 may not constitute a change at all. If the approved maximum is 200, however, enrolling participant 201 may move beyond the authorized protocol.
Read the actual wording rather than treating every number in a research proposal as equivalent.
Increasing Sample Size Can Increase Participant Exposure
When more people participate, more people may be exposed to whatever inconvenience, burden, privacy risk, psychological discomfort, physical intervention, or other research-related risk the study entails.
In a minimal-risk anonymous survey, adding 20 participants may have very different ethical implications from adding 200 participants to an interventional study. That is precisely why the significance of a sample-size change cannot be judged from the number alone.
For certain FDA-regulated investigations, U.S. regulations explicitly identify a significant increase in the number of subjects under study as an example of a protocol change that may require a protocol amendment at the regulatory level. This requirement is specific to the relevant FDA framework and should not be generalized to every form of human research, but it illustrates that enrollment can be a substantive protocol parameter rather than clerical metadata.
Decreasing Sample Size Can Also Have Ethical Consequences
Researchers sometimes assume that reducing enrollment cannot create an ethics problem because fewer participants means less exposure to research burden or risk. That captures only one side of the equation.
A study involving human participants should also have sufficient scientific value to justify asking people to contribute their time, information, specimens, or accept research-related risks. If enrollment is reduced so substantially that the study can no longer answer its research question with reasonable validity, the ethical justification may weaken.
In other words, “fewer participants” does not automatically mean “more ethical.” A severely underpowered study can still consume participants' contributions while becoming less capable of producing the knowledge that justified those contributions.
Statistical Power Is Relevant, but It Is Not the Whole Question
For studies based on inferential statistics, a revised sample size may alter statistical power, precision, detectable effect size, confidence intervals, or the assumptions underlying the original calculation. Researchers should therefore revisit the statistical rationale when proposing a substantial enrollment change.
But not every research design defines adequacy through a conventional power calculation. Qualitative studies may use information power, saturation-related reasoning, or another design-specific rationale. Pilot and feasibility studies may have objectives that differ from hypothesis-testing trials. Adaptive designs may have prospectively specified enrollment rules.
The ethical question should therefore be connected to the methodological logic of the particular study rather than reduced to “run another power analysis.”
Changing the Sample Size Is Different From Changing the Sample
Increasing an approved target from 200 to 250 members of the same eligible population is conceptually different from replacing one participant group with another.
Changing sample size
The number of participants changes while the approved population and eligibility criteria remain the same.
Changing the sample or population
The characteristics of eligible participants change, potentially introducing different risks, protections, consent considerations, or scientific questions.
If recruitment difficulties lead you to expand eligibility, the amendment involves more than sample size. The revised population itself must be considered.
A Higher Response Rate Does Not Automatically Authorize More Enrollment
Online surveys create a particularly easy trap. A researcher needs 300 responses, distributes the approved survey link, and discovers that 420 eligible people have responded before recruitment can be closed. This is different from intentionally deciding beforehand to continue recruiting beyond an approved maximum.
How excess responses should be handled depends on the approved protocol and applicable ethics procedures. Ideally, the protocol anticipates predictable operational issues such as over-recruitment, attrition, incomplete responses, or screen failures rather than leaving them to improvisation after approval.
If the researcher deliberately wants to raise the approved enrollment target, however, that is a proposed protocol change and should be assessed under the relevant modification procedure.
Stopping Early Can Be More Complicated Than Simply Revising a Number
Suppose a researcher planned for 500 participants but can recruit only 180. Reducing the target may require reconsideration of the analysis, research objectives, feasibility, and scientific value. In some studies, early termination may also have implications for participant follow-up, safety monitoring, sponsor obligations, or reporting.
A revised target should therefore have a defensible rationale. “The deadline is next month” explains the researcher's predicament but does not by itself establish that the modified study remains scientifically and ethically justified. Academic calendars, alas, are not a recognized statistical assumption.
Small Changes May Receive a Different Review Pathway
Not every enrollment adjustment necessarily requires full committee review. Under U.S. HHS rules, minor changes in previously approved research may be reviewed using expedited procedures. Advisory recommendations from SACHRP have also identified statistically small changes in participant numbers as possible examples of minor changes, while emphasizing that classification depends on whether the modification materially affects risk, benefit, or scientific merit.
Those examples do not create a universal percentage or numerical allowance. There is no general rule that researchers may independently increase an approved sample by 5%, 10%, or any other arbitrary figure.
The Amendment Threshold and the Statistical Threshold Are Different Questions
A statistician may conclude that increasing the target from 250 to 300 improves precision. That does not establish whether the additional enrollment is already covered by the ethics approval.
Conversely, an ethics committee may permit an enrollment modification, but that does not establish that the revised sample size is methodologically appropriate. Ethics authorization and statistical justification answer different questions, even though each can inform the other.
If the proposal differs from the approved enrollment plan, determine whether the change requires an ethics amendment. If it appears relatively limited, check how the responsible committee handles minor changes to approved research.
Watch Out
Do not invent a “10% rule,” “20% rule,” or other universal tolerance for exceeding an approved sample size unless that flexibility is actually provided by the protocol or the requirements governing your study. A convenient percentage is not an ethics policy.
04 · A Practical Example
When Recruitment Falls Short of the Approved Sample Size
Hypothetical Example
A survey study cannot reach its planned 400 participants
A researcher received approval for a cross-sectional survey with a target sample of 400 participants based on the study's original statistical plan. After exhausting the approved recruitment strategy, only 260 eligible participants have enrolled. The researcher considers closing recruitment and analyzing the available data.
Original plan
The approved protocol describes a target of 400 participants and explains the statistical rationale.
Practical problem
Recruitment has stalled at 260.
Methodological assessment
The researcher examines what the smaller sample would do to precision, power, planned analyses, and the study's ability to answer its research question.
Protocol assessment
The researcher determines whether stopping at the lower enrollment changes the approved study or analysis plan and what modification procedure applies.
Action
Any required amendment is submitted before intentionally adopting the revised study plan, with the new target and methodological rationale clearly explained.
The correct response is not automatically “keep recruiting until 400” or “260 is close enough.” The researcher needs to consider both the approved protocol and whether the smaller study remains capable of producing sufficiently meaningful evidence for its stated purpose.
06 · What This Means for You
Check Both the Approved Enrollment Limit and the Scientific Rationale
Before changing sample size, identify whether the approved protocol specifies a target, a range, a minimum, a maximum, or separate numbers for screening, enrollment, and completed participants. Then determine whether your proposed change actually moves outside those parameters.
A simple decision framework
If the proposed enrollment remains within an explicitly approved range or maximum
Follow the approved protocol and confirm that any other enrollment conditions remain satisfied.
If you want to exceed the approved maximum or formally change the approved target
Determine whether an amendment or other prospective ethics process is required before intentionally enrolling beyond the approved plan.
If you want to reduce the target substantially
Reassess whether the revised sample remains methodologically capable of addressing the research question and determine what ethics modification is required.
If the sample-size change also requires new eligibility criteria or a different participant group
Treat those as additional protocol changes rather than describing the proposal solely as a numerical adjustment.
If the change requires a revised statistical or analytical plan
Ideally, foreseeable enrollment flexibility is built into the protocol from the beginning and justified transparently. That does not mean inflating the requested maximum “just in case.” It means distinguishing thoughtfully among recruitment targets, maximum enrollment, expected attrition, exclusions, and the number of analyzable participants where those distinctions are relevant to the design.
07 · A Quick Checklist
Before Increasing or Decreasing an Approved Sample Size
Before changing enrollment, check:
Does the approved protocol specify a target, range, minimum, maximum, or separate screening and enrollment numbers?
Would the proposed enrollment remain within the flexibility already approved?
What methodological reason justifies increasing or decreasing the sample?
Would a larger sample expose additional people to research burden or risk?
Would a smaller sample materially weaken power, precision, information adequacy, or the study's ability to answer its question?
Does changing the number also change eligibility criteria, recruitment methods, study duration, or the analysis plan?
What amendment or modification process does the responsible ethics committee require?
Have you obtained any required prospective approval before intentionally implementing the revised enrollment plan?
