Skip to main content
v2026.11,986 entries · CC-BY 4.0

How to Show P Values on a Prism Graph (Pairwise Comparisons)

Add P values or asterisks to a GraphPad Prism graph with Pairwise Comparisons: steps, supported analyses and graph types, FDR caveat, and fixes.

Written and maintained by CASRAI Editorial Board

Published · Last updated

On this page

To show a P value on a GraphPad Prism graph, run the statistical analysis first (a t test, a one-way ANOVA with multiple comparisons, or a two-way ANOVA), open the graph, and click the Pairwise Comparisons button in the Draw section of the toolbar, or choose Insert › Pairwise Comparisons. Prism then draws a bracket for every comparison in the analysis and labels it with either a numeric P value or asterisks, and a second click opens the Format Pairwise Comparisons dialog where you choose which one. The labels are linked to the analysis, so they update if the data change. This works only for certain analyses and graph types, listed below.

What it is used for in research

A figure that compares groups usually has to carry its own evidence: readers expect to see which bars or columns differ and by how much. Prism calls the brackets with labels “pairwise comparisons”, and they put the result of a test directly on the graph instead of leaving it in a results table or a figure legend. The same feature is the standard way to show the output of a multiple-comparisons test after a one-way ANOVA, where several pairs are tested at once and the correct adjusted P value for each pair matters more than the overall F test.

The decision that matters for a methods section is not how to draw the bracket but which number belongs on it. Prism’s own guidance for unpaired t tests says the author recommends showing the exact P value rather than a summary symbol, and the Prism FAQ on asterisks assumes the conventional threshold of 0.05. For the reporting side, including how many digits to give and what to write in a legend, see our guide on how to report P values, and for the definition itself see the dictionary entry for p-value.

How to install or access it

Pairwise Comparisons is a built-in Prism feature, documented in the GraphPad Prism user guide at graphpad.com, so there is nothing separate to install. You need a copy of Prism obtained from GraphPad itself; our page on the GraphPad Prism student license explains the licensing routes. Analyses made in an older Prism version may need to be recalculated in a current one before the comparisons will appear on the graph, according to the GraphPad guide.

Step-by-step: add P values to a Prism graph

These steps follow the current GraphPad Prism user guide page on pairwise comparisons (read 2026-11-14). The menu and button names are quoted as the guide gives them.

  1. Run the analysis on the data table. Use a t test, a one-way ANOVA with one of its multiple-comparison options, or a two-way ANOVA with individual cell-mean comparisons. Open the results sheet at least once; if you made the analysis by duplicating it with the Wand tool and never opened its results, Prism alerts you and asks you to open the results sheet first.
  2. Open the graph. The graph must be built from that one data table. Prism does not draw comparisons when the graph contains data from more than one data table.
  3. Click Pairwise Comparisons. It is a button in the Draw section of the toolbar; the same command is under the Insert menu as Pairwise Comparisons. The first click adds lines (brackets) and P values or asterisks for all the comparisons in the analysis.
  4. Open the Format Pairwise Comparisons dialog. Click the button a second time, or double-click any comparison line.
  5. Choose how each result is labelled (Appearance tab). Show numeric P values or asterisks. For numeric values choose the prefix, “P =”, “p =” or a custom one. You can show all comparisons or only those below a cutoff you set between 0 and 1. The same tab sets the bracket style (three options), line thickness, colour and style, the size of the text and of the asterisks, and the distance between a bracket and its label.
  6. Choose which comparisons appear (Comparisons on Graph tab). Select or deselect individual comparisons. A dropdown on this tab selects which analysis the graph shows when more than one exists; Prism displays one analysis at a time.
  7. Fine-tune single lines. Right-click a line to change its font, colour and line style, or the position of its title. Shift-click or drag to select several lines and format them together.

Which analyses draw brackets

The data table below lists every analysis and graph type the Prism guide states a rule for. In short: all variants of t tests, all multiple-comparison options of a one-way ANOVA, and individual cell-mean comparisons from a two-way ANOVA are supported. One-sample t and Wilcoxon tests are supported only when the data have no subcolumns, or when the “Average the replicates in each row” option was selected. Multiple t tests are supported too, but the guide warns that many comparisons on one graph become hard to read. Comparing whole rows or whole columns in a two-way ANOVA is not supported.

Which graph types cannot show them

The guide lists these graph families as unsupported: XY, Survival, Nested, Parts of whole, Multiple variables, Heat map (part of grouped) and Three-way (part of grouped). For stacked grouped graphs the stacked groups share the same horizontal position, so the comparison lines would have zero length and are not drawn. If your design is one of these, use the manual method described under troubleshooting below.

Output and how to read it

Each bracket spans the two groups that were compared, and its label is the result of that pair from the analysis you chose. With asterisks, Prism’s FAQ on significance symbols gives the usual convention, which assumes an alpha of 0.05 (using a different alpha can produce a symbol that conflicts with your decision):

Symbol Meaning (alpha = 0.05)
ns P ≥ 0.05
* P below 0.05
** P below 0.01
*** P below 0.001
**** P below 0.0001

Exact P value or asterisks?

For an unpaired t test, the Prism documentation lists two ways to show the result on a graph, the exact P value or a summary (ns or asterisks), and its author says “I recommend you show the exact P value”. The same page notes that an estimation plot is a more informative display for an unpaired t test than a bracket alone. For several-group experiments, the context for choosing a test is explained in our guide on multiple comparisons and the Bonferroni correction.

Do not put FDR results on the graph as P values

If you control the false discovery rate, Prism compares q values with the chosen Q, not true P values. GraphPad’s FDR guidance says not to show asterisks or P values from FDR results on a graph: show the numeric q value, remove the lines, or switch to the P-threshold method. The worked example in that note uses Q = 0.01 and q = 0.0110: that comparison is not a discovery, yet a single asterisk would be displayed, which is exactly the contradiction to avoid.

Linked and unlinked labels

Labels made with the Pairwise Comparisons button are tied to the analysis. If you instead copy asterisks or P values from the results sheet and paste them on the graph, they snap into place and stay linked. P values or asterisks typed with the text tool are not linked to the analysis and will not change when the data do. That is the safest reason to prefer the button over typing the numbers.

Letters instead of brackets

When there are many groups, brackets crowd the figure. Prism’s compact letter display puts letters above or beside each group; groups sharing a letter did not differ at the chosen alpha. It is supported only for all-versus-all comparisons: a one-way ANOVA using “Compare the mean (rank) of each column with the mean (rank) of every other column”, or a two-way ANOVA using “Compare cell means regardless of rows and columns”. It is not recommended for t tests. Use the CLD button in the Draw section or Insert › Compact Letter Display, pick the analysis, the label location (above the data or outside), the direction, the identifier (upper-case letters, lower-case letters or numbers) and the font. As with brackets, it shows one analysis at a time.

Troubleshooting

No lines appear after clicking the button

Check, in this order, against the GraphPad guide: whether the graph family is one of the unsupported types; whether the graph uses more than one data table; whether the analysis is one of the supported kinds with the supported comparison option; whether the analysis was duplicated with the Wand and never opened; and whether it was created in an older Prism version and needs recalculating. If you deselected every comparison, clicking the toolbar button turns them back on.

Removing the lines

Click one line and press Delete, or deselect it on the Comparisons on Graph tab. To remove all of them, open the dropdown on the toolbar button and choose “Remove Pairwise Comparisons”, or use the Change menu › Pairwise Comparisons › Remove Pairwise Comparisons. Removing all lines erases them and their custom formatting; it does not simply hide them.

Marking significance manually

GraphPad’s FAQ describes a fallback for graphs where the automatic feature does not apply: copy the asterisks from the results page and paste them on the graph. It also describes a template method that adds a data set with Y = Y + K, displayed as an “*” symbol. For a t test background, see our guide to the GraphPad Prism t test, and if you only need the number before building the figure, the GraphPad P value calculator guide covers it.

Free alternatives

If you do not have Prism, the R package ggprism offers a comparable bracket function for ggplot2 figures. Its documentation describes add_pvalue() as adding p-values, with or without brackets, to a ggplot. The table below compares the options named in this guide.

Option How the P value gets onto the graph Linked to the analysis?
Prism Pairwise Comparisons button Automatic brackets from the chosen analysis, with numeric P values or asterisks Yes
Prism copy and paste from the results sheet Paste asterisks or P values onto the graph; they snap into place Yes
Prism text tool Type the value by hand No
R with ggprism Build a data frame with group1, group2, label and y.position, then add add_pvalue(df) to the plot No: the label is a value you computed, for example with t.test()

The ggprism vignette’s own example computes the value with base R, t.test(extra ~ group, data = sleep)$p.value, rounds it with signif(), places it in a data frame and adds it to the plot with p + add_pvalue(df_p_val). In code you are responsible for matching the label to the test you ran.

Frequently asked questions

How do I add a P value to a Prism graph?

Run a supported analysis, open the graph and click Pairwise Comparisons in the Draw section of the toolbar (or Insert › Pairwise Comparisons). A second click opens Format Pairwise Comparisons, where you choose numeric P values or asterisks.

Can Prism show the exact P value instead of asterisks?

Yes. On the Appearance tab you can show numeric P values, with a prefix such as “P =” or your own text, and you can limit the display to comparisons below a cutoff between 0 and 1.

Why does Prism not show pairwise comparisons on my XY or survival graph?

Those graph families are not supported by the feature. The unsupported families are XY, Survival, Nested, Parts of whole, Multiple variables, Heat map and Three-way. Use the manual copy-and-paste approach for a result you still need on the figure.

Does a P value pasted onto the graph update if the data change?

It does if it came from the Pairwise Comparisons button or was copied and pasted from the results sheet, since those stay linked. It does not if it was typed with the text tool.

Which stars mean which P value in Prism?

With alpha 0.05, ns means P ≥ 0.05, one asterisk P below 0.05, two P below 0.01, three P below 0.001 and four P below 0.0001.

Sources

  • GraphPad Prism user guide, “Pairwise comparisons on a graph”: graphpad.com (read 2026-11-14).
  • GraphPad support FAQ, “Can I make Prism mark certain bars or symbols with asterisks to indicate statistical significance?”: graphpad.com (read 2026-11-14).
  • GraphPad support FAQ, “FDR and significance” (KB 2231): graphpad.com (read 2026-11-14).
  • GraphPad Prism statistics guide, graphing the results of an unpaired t test: graphpad.com (read 2026-11-14).
  • GraphPad Prism user guide, “Compact letter display”: graphpad.com (read 2026-11-14).
  • ggprism vignette, “Adding p-values”: CRAN mirror (read 2026-11-14).

Data: GraphPad Prism pairwise-comparison brackets: supported analyses and graph types

Compiled from the GraphPad Prism user guide page on pairwise comparisons: for each analysis type or graph family, whether Prism draws P value brackets on the graph and the condition the guide states.

Period covered: 2026.

GraphPad Prism pairwise-comparison brackets: supported analyses and graph types
Analysis or graph typeCategoryBrackets drawnCondition stated in the guide
t tests (all variants)AnalysisYesNo extra condition stated
Multiple t testsAnalysisYesHard to read when many comparisons are shown
One-sample t test and Wilcoxon testAnalysisConditionalOnly with no subcolumns, or with 'Average the replicates in each row' selected
One-way ANOVAAnalysisYesAll multiple-comparison options
Two-way ANOVA: individual cell meansAnalysisYesFive options listed, including comparing cell means regardless of rows and columns
Two-way ANOVA: whole rows or columnsAnalysisNoOnly comparisons of individual cell means are supported
Stacked grouped graphGraph typeNoStacked groups share one horizontal position, so lines would have zero length
XY graphGraph typeNoUnsupported graph family
Survival graphGraph typeNoUnsupported graph family
Nested graphGraph typeNoUnsupported graph family
Parts of whole graphGraph typeNoUnsupported graph family
Multiple variables graphGraph typeNoUnsupported graph family
Heat map and three-way (grouped)Graph typeNoUnsupported graph families
Graph built from more than one data tableGraph setupNoComparisons do not appear

Sources

Compiled by CASRAI from the sources above. Free to reuse under CC BY 4.0 with attribution.

Follow CASRAI

Research-administration guidance, standards updates and independent tool reviews.

Ask CASRAI · free to try

Ask about How to Show P Values on a Prism Graph (Pairwise Comparisons)

2 free questions a day, no account.

An AI assistant for research: funder rules, compliance, methods and publishing. It links the sources behind its answers and labels answers that come from a web search.

Answers draw on CASRAI's guides and dictionary plus the federal and funder documents we index: Federal Register, Grants.gov, Regulations.gov and UKRI.

Works on this site and inside Claude, Cursor and the AI tools you already use.

Everything CASRAI publishes — this page, the dictionary, the guides and the news — stays open to read, with no account and no card.

Referenced across the research world

University of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logoUniversity of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logo
  • University of Cambridge logo
  • Columbia University logo
  • Crossref logo
  • University of Edinburgh logo
  • Harvard University logo
  • University of Oxford logo
  • Princeton University logo
  • Stanford School of Medicine logo
  • University College London logo
  • ORCID logo

View CASRAI adoption →