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

Start one
Contents

Academy · Research Data, Analytics & AI · Random errors in Clinical Trials

Random Error in Clinical Trials: Quantification, Interpretation and Control

A wider confidence interval does not indicate a design flaw. It represents an honest report of random error, which an apparently more precise but underpowered trial fails to provide.

Pubrica AcademyWritten by a biostatistics reviewer Reviewed bya clinical data manager Published10 October 2026 Last checked10 October 2026 Reading time8 minutes Sources1 peer-reviewed

Random error in a clinical trial is the variability in a result that arises purely from chance -- from sampling a finite number of participants rather than the whole population of interest -- and it is distinct from systematic error (bias), which arises from a flaw in design or conduct that would not be resolved by simply enrolling more participants.

Learning objectives

  • The distinction between random error and systematic error, and why conflating them misleads interpretation.
  • How confidence intervals and p-values both quantify random error, in different ways.
  • Why sample size calculation is the primary design tool for controlling random error.
  • Why a statistically non-significant result is not the same as proof of no effect.

Common misconceptions

  • “A wide confidence interval means the trial was poorly designed.” A wide confidence interval reflects the amount of random error given the trial's sample size; it can simply mean the trial was smaller than would be needed for a more precise estimate, which is a sample-size question, not necessarily a conduct flaw.
  • “A non-significant p-value proves there is no treatment effect.” A non-significant result is consistent with no effect, but is equally consistent with a real effect the trial was underpowered to detect with the random error present at its sample size -- the confidence interval's width, not the p-value alone, shows which interpretation the data support.

Distinction between random error and systematic error

The ICH E9 guideline on statistical principles for clinical trials, developed by the International Council for Harmonisation, frames a trial's statistical analysis around quantifying and controlling the uncertainty introduced by sampling variability -- what is conventionally called random error -- distinct from systematic error (bias), which arises from design or conduct problems such as inadequate allocation concealment or differential loss to follow-up. The distinction matters because the two call for different fixes: random error is addressed by sample size and precision; systematic error is addressed by design features such as randomisation, blinding, and complete follow-up.

Quantification of random error by confidence intervals and p-values

ICH E9 frames both the confidence interval around a treatment effect estimate and the p-value from a hypothesis test as ways of expressing the role chance could have played in producing the observed result, given the trial's sample size. A confidence interval shows the range of effect sizes reasonably compatible with the data; a p-value expresses how surprising the observed result would be if there were truly no effect. Neither tool corrects for systematic error -- a biased trial can produce a narrow confidence interval and a significant p-value that are both precise and wrong.

Sample size as the primary means of control

Because random error shrinks as the amount of information (broadly, sample size and event count) increases, ICH E9 requires that a trial's sample size be calculated in advance, based on the smallest treatment effect considered clinically relevant and the amount of random error (expressed through the trial's target significance level and power) the investigators are willing to accept. A trial that omits this calculation, or that is stopped early for reasons unrelated to its pre-specified design, risks a result dominated by random error in a way that is difficult to interpret credibly after the fact.

Key points

Key points

Random error is the variability a trial's result carries purely from sampling a finite number of participants, quantified by its confidence interval and p-value, and distinct from systematic error (bias), which arises from design or conduct problems -- sample size calculation in advance is the primary way to control how much random error a trial's result will carry.

Practical application

Practical application

  1. The sample size justification should be calculated and reported explicitly. The minimum clinically relevant effect size, the assumed variability, and the target significance level and power used to derive the sample size should be stated, rather than a sample size being given without a stated basis.
  2. The confidence interval should be reported, not only the p-value. The confidence interval around the effect estimate should be presented so that a reader can judge the range of effects compatible with the data, rather than relying solely on a binary significant/non-significant label.
  3. A non-significant result should not be equated with evidence of no effect. Where applicable, it should be stated explicitly that the confidence interval is wide enough to be compatible with both a clinically important effect and no effect.

Frequently asked questions

No; a larger sample size reduces random error, narrowing confidence intervals, but it does nothing to correct systematic error from a flawed design or conduct, and a larger sample may simply yield a more precise but still biased estimate.

Type I error is concluding there is an effect when there truly is none (a false positive); Type II error is concluding there is no effect when one truly exists (a false negative) -- both are forms of random error that the trial's design, through its significance level and power, is set up to control the probability of.

No; because any trial samples a finite population, some random error is unavoidable, which is why the aim is to control and quantify it to an acceptable level through sample size, not to eliminate it.

References

  1. International Council for Harmonisation. ICH Harmonised Guideline E9: Statistical Principles for Clinical Trials. read 10 October 2026. View ICH E9 guideline (PDF)

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). Random Error in Clinical Trials: Quantification, Interpretation and Control. Pubrica. https://pubrica.com/academy/concepts-definitions/random-errors-in-clinical-trials/

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.

Random-error reporting checklist

Use when reporting a trial's statistical results.

[ ] Sample size calculation basis stated (effect size, variability, significance level, power)
[ ] Confidence interval reported alongside point estimate
[ ] P-value reported alongside, not instead of, the confidence interval
[ ] Non-significant result interpreted against the confidence interval's width, not the p-value alone
[ ] Any early stopping or sample size deviation explained

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 develops sample size justifications and prepares statistical results sections that report confidence intervals and the implications of random error transparently, to the standard expected under ICH E9 and by journal statistical reviewers.

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.

Transparent representation of random error in trial statistical reporting

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.