Free scoping read. A written assessment of your manuscript in two business days, at no cost.

Start one
Contents

Academy · Research Data, Analytics & AI · Sample Size and Interval Width

Sample Size and Confidence Interval Width: The Square-Root Relationship and Its Implications for Study Planning

The width of a confidence interval decreases approximately in proportion to the square root of the sample size; consequently, the number of participants must be quadrupled, not merely doubled, to halve the width of the reported interval.

Pubrica AcademyWritten by a biostatistics analyst Reviewed bya research data methodologist Published10 October 2026 Last checked10 October 2026 Reading time7 minutes Sources10 peer-reviewed

Sample size affects a confidence interval through the standard error of the estimate, which is a function of the variance in the data divided by the sample size, and then, through a square-root relationship, carried into the width of the interval itself. Because the standard error scales with the square root of the sample size rather than with the sample size directly, interval width does not shrink in a straight line as more participants are added; doubling the sample size reduces the standard error, and with it the interval's half-width, by a factor of roughly the square root of two, not by half1. This single relationship explains both why early, small trials tend to report very wide intervals and why a meta-analysis pooling many studies can achieve a narrower interval than any individual study it includes.

Learning objectives

  • The square-root relationship between sample size and confidence interval width
  • Why quadrupling the sample size, rather than doubling it, is required to halve interval width
  • How pooling studies in a meta-analysis narrows intervals beyond what any single study achieves

Common misconceptions

  • “Doubling a study's sample size halves the width of its confidence interval.” Because the standard error scales with the square root of the sample size, doubling the sample size narrows the half-width of the interval by a factor of about 1.41, not 2; halving the interval width requires approximately quadrupling the sample size, a relationship derived directly from the standard formula for the margin of error of a confidence interval.

The square-root relationship between sample size and precision

A confidence interval's margin of error (the amount added to and subtracted from the point estimate to form the interval's bounds) is calculated as a critical value from the relevant distribution multiplied by the standard error of the estimate, and the standard error itself is calculated as the standard deviation of the data divided by the square root of the sample size1. This is the mechanical reason that sample size and interval width are related through a square root rather than linearly: because the sample size n appears under a square root in the denominator of the standard error formula, increasing n reduces the standard error, and therefore the interval width, but at a diminishing rate as n increases. The ICH (International Council for Harmonisation) E9 guideline on statistical principles for clinical trials relies on the same relationship in describing how a trial's planned sample size is chosen to achieve a target level of precision for its primary endpoint estimate2.

Quadrupling rather than doubling the sample to halve interval width

Because the standard error depends on the square root of the sample size, halving the interval width requires halving the standard error, which in turn requires multiplying the sample size by four, since the square root of 4 is 21. For example, if a trial of 100 participants produces a confidence interval of a given width, a trial of 200 participants studying the same population and outcome will narrow that interval by a factor of about 1.41 (the square root of 2), not by half; halving the interval width requires a trial of approximately 400 participants. This relationship is among the most practically important, and most frequently underestimated, considerations in determining how much additional recruitment is needed to achieve a meaningfully more precise estimate.

Narrowing of intervals through the pooling of studies

A meta-analysis narrows its pooled confidence interval beyond what any single contributing study could achieve by effectively combining the sample sizes (and, in an inverse-variance weighted model, the precision) of all included studies into one estimate3. DerSimonian and Laird's random-effects method weights each study inversely by its own variance, so that studies with larger sample sizes and smaller individual standard errors contribute more to the pooled estimate, and the combined effective sample size behind the pooled interval is larger than that of any one study; this is the central statistical rationale for conducting a meta-analysis3. This narrowing is tempered, however, by between-study heterogeneity: if the pooled studies disagree substantially, the additional between-study variance term in a random-effects model partially offsets the narrowing that the increased total sample size would otherwise produce.

Common errors in sample-size planning

Schulz and Grimes's methodological review of sample size calculation in randomised trials identifies several recurring planning errors that leave a study's reported confidence interval wider than investigators expected: underestimating the variability of the outcome measure used in the power calculation, failing to account for anticipated loss to follow-up when setting the recruitment target, and, in trials comparing proportions, not accounting for how the required sample size changes sharply as the expected event rate approaches either 0% or 100%4. Each of these errors has the same consequence: the effective sample size contributing to the final analysis, after dropouts, missing data or a lower-than-expected event rate, is smaller than planned, which, through the square-root relationship, produces a wider confidence interval than the original sample-size calculation anticipated.

Persistence of wide intervals for rare events

For a binary outcome such as a rare adverse event, the effective precision of an estimate depends on the number of events observed, not simply on the number of participants enrolled1. A trial of several thousand participants studying an outcome that occurs in only a few of them can still produce a very wide confidence interval around the estimated event rate, because the effective sample size determining the standard error is closer to the small number of observed events than to the large number of enrolled participants. This is the statistical reason that safety outcomes in clinical trials frequently have much wider confidence intervals than the trial's primary efficacy outcome, even though both are measured in the same adequately recruited population, and it is consistent with the finding of Zorzela and colleagues' systematic review of adverse-event reporting quality that harm outcomes are frequently reported with less statistical rigor and less precision than the benefit outcomes drawn from the same trials5.

A worked illustration of diminishing returns

Consider a pilot study of 50 participants that produces a confidence interval of a given width around a mean difference. On the square-root relationship, increasing the sample to 200 participants, a fourfold increase, is expected to halve the width of that interval1. A further halving, to a quarter of the original width, requires not another fourfold increase but a sixteenfold increase from the original sample, to 800 participants, because the square-root relationship compounds: each successive halving of width requires an increase in sample size four times as large as the preceding increase. This compounding explains why very large, costly confirmatory trials are needed to achieve a modest further reduction in the width of an already reasonably precise estimate, and why investigators planning a follow-up to a promising pilot study should calculate the sample size needed for a specific target width rather than assume that an unspecified increase in participants will deliver a proportionally narrower interval.

Implications for planning a new study or recruitment target

For investigators planning a new study, three practical consequences follow from the square-root relationship. First, a target confidence interval width should be specified before the sample size is calculated, since working backward from a specific precision target, rather than selecting a sample size that appears large, is the only way to establish whether a planned study can answer the question it is designed to address2. Second, because of diminishing returns, combining several adequately conducted smaller studies through meta-analysis is often more efficient than funding one very large trial to obtain the final increment of precision, provided that the studies are sufficiently similar for pooling to be appropriate3. Third, when a harm outcome is of particular interest, sample-size planning should be based on the expected rate of that harm outcome, not on the rate or variability of the primary efficacy outcome, since the two may require very different sample sizes to achieve comparably useful precision.

A further worked planning scenario illustrates the square-root relationship for a specific recruitment decision. Suppose a funding body has budgeted for 300 participants, and a protocol committee wishes to know whether that budget can achieve an interval half as wide as that observed in a prior pilot study of 75 participants. Because halving the interval width requires quadrupling the sample size, the pilot's 75 participants would need to increase to 300 to reach this target; the budgeted sample size therefore corresponds exactly to the precision goal, neither falling short of it nor exceeding it1. If the committee instead wished to narrow the interval to a quarter of the pilot's original width, the required sample size would be 1,200 participants, sixteen times the size of the pilot, which illustrates how steeply the cost of additional precision rises once a reasonably informative estimate has been obtained. This explicit reconciliation of a stated precision target with the implied sample size, conducted before a protocol is finalized rather than after a trial has been designed around a round-number enrollment target, is the calculation that ICH E9 guidance expects a statistical analysis plan to document, and it is often the single calculation that reveals whether a funding body's budget and a study's precision objectives are compatible2.

Key points

Key points

Confidence interval width narrows with the square root of the sample size, so the number of participants must be quadrupled, not doubled, to halve interval width. Dropouts, lower-than-expected event rates and rare outcomes all reduce the effective sample size underlying the interval even when enrolment targets are met, and a meta-analysis narrows its pooled interval by combining the precision of every included study into a single estimate.

Practical application

Practical application

  1. Recruitment targets should be planned using the square-root relationship rather than a linear assumption. If a target interval width requires a specific reduction from the width observed in a pilot study, the required sample size should be calculated using the square-root formula rather than by assuming a proportional relationship.
  2. Diminishing returns should be anticipated from adding participants at larger sample sizes. The precision gained by adding 100 participants to a trial of 1,000 is smaller than that gained by adding 100 participants to a trial of 100, as a consequence of the same square-root relationship.
  3. When a meta-analysis reports a pooled interval, the sources of its width should be examined rather than attributed to sample size alone. In a random-effects meta-analysis, between-study heterogeneity contributes to uncertainty in the pooled estimate. Confidence interval width also depends on study precision, weighting and the estimation method, so the heterogeneity statistics should be examined alongside the interval.

Frequently asked questions

In general, yes, for a fixed level of underlying variability, but the relationship follows a square root, not a straight line, so the practical gain in precision shrinks as sample size grows larger, which is why very large trials see smaller marginal improvements in interval width from additional participants.

Because the standard error term in the interval formula is inversely related to the square root of sample size, and small early trials have correspondingly small sample sizes, their intervals are mechanically wider; this is expected and is one reason early trial results are often described as preliminary rather than definitive.

References

  1. Deeks JJ, Higgins JPT, Altman DG (editors). Chapter 10: Analysing data and undertaking meta-analyses. In: Cochrane Handbook for Systematic Reviews of Interventions. Cochrane, 2024. read 10 October 2026. View on Cochrane Training
  2. International Council for Harmonisation. E9: Statistical Principles for Clinical Trials. read 10 October 2026. View on ICH
  3. DerSimonian R, Laird N. Meta-analysis in clinical trials. Controlled Clinical Trials. 1986;7(3):177-188. read 10 October 2026. View on ScienceDirect
  4. Schulz KF, Grimes DA. Sample size calculations in randomised trials: mandatory and mystical. The Lancet. 2005;365(9467):1348-1353. read 10 October 2026. View on The Lancet
  5. Zorzela L, Golder S, Liu Y, Pilkington K, Hartling L, Joffe A, Loke Y, Vohra S. Quality of reporting in systematic reviews of adverse events: systematic review. BMJ. 2014;348:f7668. read 10 October 2026. View on The BMJ
  6. Gardner MJ, Altman DG. Confidence intervals rather than P values: estimation rather than hypothesis testing. BMJ. 1986;292(6522):746-750. read 10 October 2026. View on The BMJ
  7. Greenland S, Senn SJ, Rothman KJ, Carlin JB, Poole C, Goodman SN, Altman DG. Statistical tests, P values, confidence intervals, and power: a guide to misinterpretations. European Journal of Epidemiology. 2016;31(4):337-350. read 10 October 2026. View on European Journal of Epidemiology
  8. Guyatt GH, Oxman AD, Kunz R, et al. GRADE guidelines 6. Rating the quality of evidence - imprecision. Journal of Clinical Epidemiology. 2011;64(12):1283-1293. read 10 October 2026. View on Journal of Clinical Epidemiology
  9. Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-analysis. Statistics in Medicine. 2002;21(11):1539-1558. read 10 October 2026. View on PubMed
  10. Riley RD, Higgins JPT, Deeks JJ. Interpretation of random effects meta-analyses. BMJ. 2011;342:d549. read 10 October 2026. View on The BMJ

Every source above is a peer-reviewed article, a publisher’s record of one, or a standard issued by the body that owns it. Each carries the date it was last read. Where a standard moves, this article is revised rather than left up with a note.

How to cite this article

Pick your referencing style and copy it. Your university or target journal may vary the punctuation — check its guide if it has one.

Pubrica Academy. (2026, October 10). Sample Size and Confidence Interval Width: The Square-Root Relationship and Its Implications for Study Planning. Pubrica. https://pubrica.com/academy/sample-sizes-in-confidence-intervals/

This article is written to be learned from. Cite it if you are referring to what it says; if you are using the evidence behind it, cite the sources listed above instead.

Sample-Size-to-Precision Planning Note

This checklist may be used to estimate the additional recruitment needed to narrow a confidence interval.

[ ] Recorded the pilot or prior interval's width and the sample size that produced it
[ ] Calculated the target width reduction needed
[ ] Applied the square-root relationship (target sample size is approximately proportional to the square of the desired precision improvement) rather than a linear assumption
[ ] Checked expected diminishing returns if the study is already large
[ ] If pooling across studies, separately noted the expected contribution of heterogeneity to the pooled interval's width

Pubrica AcademyWritten and maintained by Pubrica’s subject matter experts. Articles are revised when the underlying standards change, and each article states the date on which it was last reviewed.

How Pubrica can help

Pubrica’s biostatistics team performs sample-size and precision calculations against a target confidence interval width, so that recruitment plans are based on the correct square-root relationship rather than a linear approximation.

Find something else

Search the Pubrica Academy

Access is free and requires no account. Enter a topic, or open the full library to filter by category, sub-topic or resource type.

Pubrica can assess whether a recruitment target is sufficient to achieve the required confidence interval width.

A scoping assessment is free and takes two business days. It provides a written assessment of what the work requires and which specialist would undertake it.

Get a free quote

Regulations

Check your project against this rule

  1. 1Your work
  2. 2Your details
  3. 3Send
What do you need?

Material you send is held under signed confidentiality, with access limited to the assigned team and encrypted transfer. It is not used to train anything and is not passed to third-party tools. Ask for the agreement before you send anything you would rather cover first.