Guide seven of seven

When a study finds nothing

It means the study did not find an effect. It does not mean there is no effect. Those are different claims, and only one of them is usually supported by the paper in front of you.

Written and reviewed by Reza D, OTR/L, MOT, CPACC · Published 10 September 2026 · Free to read

"No significant difference" means the study did not find an effect. It does not mean there is no effect. Those are different claims, and only one of them is usually supported by the paper in front of you.

What does "no significant difference" actually mean?

It means the result was compatible with chance. The study compared two groups, found some difference, and could not rule out that a difference that size would appear by luck if the treatments were identical.

Three quite different situations produce that sentence, and the paper does not always tell you which one you are reading.

Comparison

Three reasons a study reports no significant difference.
What happened	How to spot it	What you can conclude
The treatment does nothing	A large study, a narrow confidence interval sitting close to zero, and the authors say so plainly.	Reasonable to stop offering it, alongside whatever else you know.
The study was too small to tell	A small sample, a wide confidence interval spanning useful benefit and real harm, and often a pilot or feasibility label.	Nothing about the treatment. The study answered a question about itself.
The difference is real but small	A difference in the expected direction that misses significance, in a study sized for a larger effect.	Nothing yet. Worth a bigger trial, not worth a change in practice.

Why does the size of the study decide this?

Because significance depends on how much noise there is relative to the effect. A small study is noisy, so a real effect can disappear into that noise. Researchers calculate in advance how many participants they need to detect an effect of a given size, and a study below that number is likely to miss a real effect even when there is one.

This is why a null result from a pilot means very little. A pilot is deliberately undersized. The guide on studies that cannot prove anything covers why, and it applies with particular force here: a pilot reporting no significant difference has told you almost nothing about the treatment.

Read the confidence interval, not the p value

The p value tells you whether the result crossed a line. The confidence interval tells you what the study actually ruled out, which is far more useful.

An interval running from a trivial benefit to a large one means the study was uninformative. An interval sitting tightly around zero means the study genuinely found nothing worth having. Both get reported as "no significant difference", and only one of them should change what you do.

What does a useful null result look like?

Worked example

A clean null at moderate certainty · free to readAcute effects of contract-relax PNF with and without neuromuscular electrical stimulation on hamstring flexibility and jump performance in adolescent soccer playersInternational Journal of Sports Physical Therapy, 2026 · randomised controlled trialAdding electrical stimulation to a PNF stretching routine produced no more flexibility than the same routine with a placebo tingle, and neither changed jump performance. A randomised design with a placebo comparison, reporting that the extra component added nothing. That is a result worth having: it tells you not to spend the clinic time.Read the summary

Worked example

Null, and the review says why · occupational therapySystematic review of hospital-based interventions to improve post-stroke independence in activities of daily livingAustralian Occupational Therapy Journal, 2026 · systematic reviewSix trials, no clear benefit of activity-based retraining over usual care. The review does the important thing and separates the two readings: the evidence is too small and too inconsistent to support the intervention, and that is not the same as showing it does not work. Its conclusion is to keep using clinical judgement, which is the honest answer.Read the summary

Worked example

The finding survived one adjustment and vanishedEnergy cost of walking in relation to impairment and functional independence after stroke: the role of walking speedPhysiotherapy Research International, 2026 · randomised controlled trialThe energy cost of walking correlated with impairment and independence in 30 stroke survivors. Every one of those correlations disappeared once walking speed was accounted for. The headline association was real and it was walking speed wearing a different name.Read the summary

Is "no difference" ever a positive finding?

Yes, but only when the study was designed to show it. A non-inferiority or equivalence trial sets out in advance how much worse a treatment could be while still counting as acceptable, and is sized to detect that. Those studies can legitimately conclude that two treatments are interchangeable.

An ordinary trial that finds nothing cannot make that claim, however much the discussion section would like to. If the paper did not state an equivalence margin before it started, "no difference" means "we did not find one", nothing more.

What should you do with a null result?

Check the size before anything else. If the study was small, stop there. It has not told you about the treatment.
Read the confidence interval. What did this study rule out, and does that range still include something you would care about?
Check what the primary outcome was. A null primary outcome with a positive secondary one is a null trial. The guide on certainty is relevant here: the design decides how much weight the paper carries, not which sentence the abstract leads with.
Do not treat it as permission. A treatment that has not been shown to work is not the same as one shown not to.

About 9% of the papers indexed here report a null or negative finding. That is lower than it should be, and the reason is worth knowing: studies that find nothing are published less often and later than studies that find something. Every index of published research, including this one, is tilted towards results that worked.

Other guides

01What the certainty rating on a research summary actually meansEvery paper on this site carries one of three words: moderate, low, or very low. Here is where each comes from, what it does not mean, and what it should change about how you read the summary underneath it.02MCID vs MDC: telling whether a change on an outcome measure is realTwo numbers decide whether a score has moved: one asks whether the change is bigger than the measurement error, the other whether it is big enough for the patient to care. They are not the same, and one is useless without the other.03Pilot, feasibility, protocol, case report: what a study that cannot prove anything still tells youA pilot, a feasibility study, a case report and a protocol each answer a real question. None of those questions is whether the treatment works — and reading them as if it were is the commonest way to misread the literature.04Scoping review, systematic review, meta-analysis: which one answers a clinical questionOnly one of them is designed to answer whether a treatment works. Telling them apart takes about ten seconds, and it changes how much weight the paper can carry.05What a qualitative study can tell you, and what it cannotIt does not measure whether a treatment works. It finds out what something is like, in enough detail that you recognise it. That is a different question, with answers a trial cannot produce.06Reading a validation study: the numbers that decide whether an assessment is usableA validation study asks whether an assessment measures what it claims, in the people you are using it on. It answers with a handful of numbers, and two of them decide whether the tool is usable at all.

Written and reviewed by Reza D, OTR/L, MOT, CPACC. Guides are the only pages on Applied Evidence written by a person. Paper summaries are generated by software and marked as such. Last reviewed 10 September 2026.