Posters · Glossary

What is a p value?

A p value is the probability, under a specified statistical model such as no difference between groups, of a result at least as extreme as the one observed. A small p value says the data fit that model poorly. It is not the probability the hypothesis is true, and it does not measure how large an effect is.

The p value is the most reported and most misread number in research, and a poster is where misreadings travel fastest because there is no room for caveats. Reporting it well takes one exact number, an effect size and an interval beside it.

Indunil Asanka · Co-founder

6 min read · Published 13 September 2026

Ways of reporting a p value on a poster
Style	Example	What it tells the reader	Use it when
Exact value	p = 0.037	How incompatible the data are with the null model	Almost always; the preferred form
Very small value	p < 0.001	The value is below what is useful to print	The exact value would need many zeros
Threshold only	p < 0.05	Only which side of a cut off it fell	Avoid; it hides the actual value
Not significant	NS or p > 0.05	Almost nothing	Avoid; give the exact value and the interval
Asterisks	* p < 0.05, ** p < 0.01	A threshold category on a chart	Only with a legend, and never instead of intervals

What the number actually measures

The American Statistical Association's 2016 statement is the clearest short guide. Its informal definition is the probability under a specified statistical model that a statistical summary of the data would be equal to or more extreme than its observed value. Its six principles follow from that. P values can indicate how incompatible the data are with a model. They do not measure the probability that the studied hypothesis is true, or that the data were produced by random chance alone. Conclusions should not rest only on whether a threshold is passed. Proper inference needs full reporting. A p value does not measure the size or importance of an effect. And by itself it is not a good measure of evidence.

Where 0.05 came from and why it is contested

The 0.05 threshold is a convention, not a law of nature, and treating results just below and just above it as opposites is the core problem. A comment in Nature in 2019, signed by more than 800 researchers, called for retiring statistical significance as a yes or no verdict, arguing that a result of p = 0.049 and p = 0.051 carry almost the same evidence. The practical upshot for anyone presenting is simple: report the exact value, show the effect size with its confidence interval, and describe the finding by what the interval includes rather than by the word significant alone.

Why a p value needs an effect size beside it

With a large enough sample almost any difference produces a tiny p value, and with a small sample an important difference can produce a large one. Sullivan and Feinn's paper in the Journal of Graduate Medical Education gives the example of the Physicians Health Study, where aspirin's effect on heart attacks was highly significant yet very small in size. That is why the number a reader needs first is the effect: the difference, ratio or standardised size, with its interval. The p value then adds how surprising that result would be if there were no effect at all.

How to show it on a poster

Put the effect first and the p value last: difference 4.7 points (95% CI 0.3 to 9.1), p = 0.037. Give p values to two or three significant figures, and use p < 0.001 below that. On charts, prefer intervals to asterisks, and if asterisks are used, define them in the caption. When several outcomes are tested, say so and say whether any correction was made, because a board with twenty p values will contain a small one by chance. Keep the same format everywhere on the board so the key number, the results text and the figure caption match.

Common mistakes

Writing that p = 0.03 means a three percent chance the result is due to chance is the most common misreading. Calling p = 0.08 a trend towards significance implies the result is on its way somewhere. Reporting only p values without the effect or interval leaves readers unable to judge importance. Treating p > 0.05 as proof of no effect ignores the interval, which may include effects large enough to matter. And changing the analysis after seeing results until p drops below 0.05 invalidates the number entirely.

Where it shows up in the poster builder

A p value usually lives in two places on a board: under a key number and in the results text. Key number blocks, on 27 of the 40 layouts, have a value, a label and a smaller detail line designed for the interval, a p value or the sample size. The generator fills key numbers, chart values and results text from the brief, and if the brief contains no numbers it still produces a complete looking poster with invented figures that nothing on the board marks as illustrative. Put the effect, the interval and the exact p value in the brief and check every block against the analysis. The clinical trial example prints its difference with the interval and p < 0.001.

Questions people ask

Is a p value of 0.05 significant?

By the most common convention, results with p below 0.05 are called statistically significant, and exactly 0.05 sits on the line depending on the rule used. The label adds little. A p value of 0.05 means the data would be fairly unusual if there were no effect, but it says nothing about how big the effect is. Report the exact value and the interval instead of relying on the word.

What does p < 0.001 mean?

It means the calculated p value was smaller than one in a thousand, so data this extreme would be very unusual if the null model were true. It is a reporting shortcut rather than a separate category. Very small p values often come from large samples, so the effect size and its interval remain essential for judging whether the finding matters in practice.

Can a p value be exactly zero?

No. Software sometimes displays 0.000 because it rounds to three decimal places, but a p value from real data is never exactly zero. Report it as p < 0.001 rather than p = 0.000. Printing zero implies certainty that no statistical test can provide, and reviewers treat it as a sign that output was copied without checking.

What is the difference between one sided and two sided p values?

A two sided test asks whether the effect differs from zero in either direction, while a one sided test asks only about one direction. A one sided p value is roughly half the two sided value for the same data. Use two sided tests unless a one sided hypothesis was stated before the data were collected, and say which was used on the board.

Should I adjust p values for multiple comparisons?

It depends on the question. When many outcomes or subgroups are tested and any one positive result would be treated as a finding, some adjustment or a stated primary outcome protects against false positives. When outcomes were prespecified and are reported in full, many authors present unadjusted values with that context. Either way, state the number of tests on the poster.

Do Bayesian analyses use p values?

Not usually. Bayesian analyses report posterior probabilities, credible intervals or Bayes factors, which directly express how strongly the data support one hypothesis over another given prior assumptions. If a poster uses a Bayesian approach, report those quantities and the prior used, rather than converting to p values, and explain the key terms briefly for a mixed audience.

Make one with posters

The button opens the generator with this use case already described. Change the wording to match your own.

Create a poster with OneCraft

Related questions

Step by step in the builder: Add charts and diagrams.

Sources

Written and checked by the OneCraft team. Last checked 13 September 2026.