How to Find a P-Value in Excel

To find a p-value in Excel, use T.TEST for a two-sample or paired t-test, or use a distribution function such as T.DIST.2T when you already have a test statistic and degrees of freedom. First identify the statistical test your question requires: Excel can calculate the probability, but it cannot choose the right test for your study.

What You’ll Need

Microsoft Excel for Microsoft 365 on Windows, or another recent desktop version of Excel. The formulas below also work in Excel for Mac and Excel for the web, although the Analysis ToolPak setup differs.
Your observations arranged clearly in worksheet cells, with each sample in its own column for a two-sample test.
A decision about the test type, whether the test is one-tailed or two-tailed, and the significance level you plan to compare the result with, often 0.05.
For a test that is not directly covered by T.TEST, the relevant test statistic and degrees of freedom, or the appropriate Excel function for that statistical distribution.

Step-by-Step

Decide what comparison you are testing. A p-value is calculated under a specific statistical test and its assumptions; it is not a general score for a spreadsheet. Use a paired t-test when each value in one group is naturally matched with one value in the other, such as the same people measured before and after. Use an independent two-sample t-test for separate groups. If you are testing counts, proportions, correlations, or more than two groups, a t-test may not be appropriate. Confirm the method with your course materials, analysis plan, or a qualified statistician when the choice is not clear.
Arrange the data and inspect it. Put each sample in a separate column, use one observation per row, and keep headers outside the data ranges unless you account for them in the analysis. Check that the cells contain numbers rather than numbers stored as text, and decide how missing values should be handled. Do not remove an observation just because it makes the result less convenient; exclusions should follow a defensible rule established for the analysis.
Use T.TEST for a paired or two-sample t-test. The function takes two data ranges, the number of tails, and a test type. With Sample 1 in A2:A4 containing 1, 2, 3 and Sample 2 in B2:B4 containing 2, 3, 4, enter this in an empty cell: =T.TEST(A2:A4,B2:B4,2,2). The layout is three numeric observations in each of those two ranges; the final 2 requests a two-tailed test, and type 2 requests a two-sample test assuming equal variances. The result is approximately 0.2877. In this example, the group means are 2 and 3, but the small samples do not provide strong evidence of a difference at the 0.05 level. The p-value is not the probability that the null hypothesis is true.
Choose the correct test type and tails. In T.TEST, use type 1 for paired observations, type 2 for independent samples with equal variances, or type 3 for independent samples without assuming equal variances. For type 3, keep the same ranges and tails in the example above and enter =T.TEST(A2:A4,B2:B4,2,3); it returns approximately 0.2877 for these particular equal-sized samples. This result happens to match the equal-variance example, but the two choices can give different results in other data. Set tails to 2 for a two-sided question, or 1 only when your research question specified a direction in advance. Do not switch to a one-tailed test after seeing which direction the data went.
Calculate a p-value from a t statistic when needed. T.TEST does not perform a one-sample t-test against a stated reference mean. If you have already calculated the t statistic and its degrees of freedom, put the statistic in B2 and the degrees of freedom in B3. With B2 equal to 2.5 and B3 equal to 9, use =T.DIST.2T(ABS(B2),B3) to get a two-tailed p-value of approximately 0.0331. The absolute value makes the two-tailed result independent of whether the statistic is positive or negative. You must calculate the statistic and degrees of freedom using the method appropriate to your test before applying this function.
Use the matching distribution for other tests. A t-distribution function is not suitable for every analysis. For example, Excel includes CHISQ.TEST for a chi-square goodness-of-fit or independence test when you supply observed and expected counts, and F.TEST for comparing two variances. These functions have their own inputs and assumptions. Look up the exact function guidance for your test, and do not substitute one function simply because its name looks familiar. For a test where you already have a chi-square statistic and degrees of freedom, Excel has chi-square distribution functions such as CHISQ.DIST.RT.
Optionally use the Analysis ToolPak. In Excel for Microsoft 365 on Windows, open File > Options > Add-ins. At the bottom, set Manage to Excel Add-ins and select Go, then check Analysis ToolPak and select OK. Go to Data > Data Analysis, choose the suitable t-test entry, specify the input ranges and output location, and review the output. The table includes p-value fields such as P(T<=t) two-tail. For Mac, enable the add-in through Tools > Excel Add-ins and select Analysis ToolPak; then look for Data Analysis on the Data tab. The ToolPak is not available in Excel for the web, so use a worksheet function there.
Interpret the result in context. Compare the p-value with a significance level chosen before interpreting the analysis. If the p-value is below 0.05, for example, the result is commonly called statistically significant at the 5% level. If it is above 0.05, you have not established statistical significance at that threshold; that does not prove the groups are identical. Report the test, the p-value, and useful context such as sample sizes and an effect estimate. A p-value alone does not show how large or important a difference is.

Common Mistakes

Using the wrong test type. Paired measurements are not independent groups. Conversely, selecting a paired test for unrelated groups creates a misleading result. Match the test to how observations were collected.
Choosing one tail after looking at the data. A one-tailed test is justified only when the directional hypothesis was set in advance and an effect in the opposite direction would not count as evidence for the research claim. Otherwise, use a two-tailed test.
Including headers or unrelated cells. Selecting a title row, blank cells, totals, or extra observations can cause an error or change the analysis. Highlight the data ranges carefully and check their sizes.
Treating the p-value as the chance the result is random. More precisely, it is the probability, assuming the null hypothesis and test assumptions, of getting a result at least as extreme as the observed one. It does not measure the probability that the null hypothesis is true.
Reading 0.000 as exactly zero. A displayed value rounded to 0.000 is a very small value at the shown precision, not a probability of zero. Increase the displayed decimal places before reporting it.
Ignoring assumptions or practical importance. A small p-value does not correct biased sampling, dependent observations, outliers, or an unsuitable model. It also does not tell you whether a difference matters in practice.

Tips

Label columns with group names and record whether the test is paired, equal-variance, or unequal-variance. Clear labels make it easier to review the analysis later.
Keep the raw observations unchanged and put the formula in a separate cell. This makes the calculation easier to audit and reduces the chance of accidentally overwriting data.
Use a cell reference for a significance threshold if you will compare several results with the same cutoff. Record the threshold and the reason for choosing it rather than changing it to make a result significant.
Report a p-value with an appropriate number of digits, not excessive decimal places. Also report the sample sizes and an effect size or confidence interval where your analysis method allows it.
For a large set of tests, consider that repeated testing raises the chance of false positives. Decide whether a multiple-comparison adjustment is needed as part of the analysis plan; a single p-value function does not make that adjustment automatically.
Save a copy of the workbook or keep a dated analysis sheet with the test choice, ranges, and result. This helps another reader reproduce the calculation.

Troubleshooting

#VALUE! or an unexpected error: Check that the selected ranges contain numeric observations and that the arguments are in the correct order. In T.TEST, both ranges need data, tails must be 1 or 2, and type must be 1, 2, or 3. Remove text entries from the selected ranges or correct numbers stored as text.

The result is not what you expected: Confirm that the ranges refer to the intended groups, that you selected the right paired or independent test, and that the tail setting matches the hypothesis specified before analysis. Compare the formula result with the ToolPak output only when both are configured for the same test and assumptions.

The function name is not recognized: Check spelling and whether the workbook is being opened in a compatible spreadsheet program. Excel for the web supports worksheet statistical functions, but it does not provide the Analysis ToolPak interface. On Windows, check that the add-in is enabled if Data Analysis is missing; on Mac, check Tools > Excel Add-ins.

The p-value appears as a dash or zero: The cell may have a number format that rounds small values away. Select the result cell and use Home > Increase Decimal, or open Format Cells and choose a suitable number format. A displayed zero can mean the value is smaller than the displayed precision.

The output changes when you edit the sheet: Check whether the formula ranges accidentally include a changing total or extra rows. For a fixed analysis, select just the intended observation cells; when adding observations, expand both ranges consistently and verify that the test still matches the study design.

Frequently Asked Questions

Can Excel calculate a p-value directly?
Yes. Use T.TEST for paired or two-sample t-tests. For other tests, use the matching statistical function or calculate the appropriate test statistic and pass it to the corresponding distribution function.

What does a p-value of 0.05 mean?
Under the null hypothesis and the test assumptions, it represents a 5% probability of observing a result at least as extreme as the one obtained. It does not mean there is a 95% chance that the research hypothesis is true.

How do I get a one-tailed p-value in Excel?
For T.TEST, set the tails argument to 1 only for a prespecified directional test. If using a t statistic, a right-tail probability can be calculated with a t-distribution function, but make sure the statistic’s direction agrees with the alternative hypothesis.

Does T.TEST test one sample against a target value?
No. It compares two arrays, either as paired data or as two samples. For a one-sample t-test, calculate the one-sample t statistic and degrees of freedom, then use the appropriate t-distribution function to obtain the p-value.

Why does Excel show a different p-value from a website?
Check that both calculations use the same observations, test type, tail count, and variance assumption. Also check whether the website reports a one-tailed result while Excel reports a two-tailed result, and whether either value has been rounded.

Can I find a p-value in Excel for the web?
Yes, worksheet formulas such as T.TEST work in Excel for the web. The Analysis ToolPak is a desktop feature, so use Excel for Windows or Mac if you need its Data Analysis dialog.

Try it: the Date Difference Calculator: days, months, years and weekdays between two dates, with the DATEDIF formulas.

More in How-to basics

Sources: Function behaviour and menu paths: Microsoft Support documentation for Excel. Method & sources

Spotted an error in this guide? Let us know.