How to Find P Value in Excel, Calculator, and by Hand

When analyzing data, one question sits at the heart of nearly every statistical test: could these results have occurred by chance? The p value answers that question. It quantifies the probability of observing results at least as extreme as your data, assuming the null hypothesis is true. A small p value signals that your results are unlikely under the null hypothesis — grounds for rejecting it. A large p value suggests the data are consistent with chance variation alone.

Understanding how to find a p value is an essential skill for students, researchers, and analysts across disciplines, from medicine and psychology to economics and data science. The calculation depends on your hypothesis test — whether you’re running a z-test, t-test, chi-square test, or F-test — but the underlying logic is the same.

What Is a P Value?

A p value (probability value) is a number between 0 and 1 that measures the strength of evidence against the null hypothesis. Formally, it is the probability of obtaining a test statistic as extreme as — or more extreme than — the one calculated from your sample, assuming the null hypothesis is true.

The null hypothesis (H₀) is the default assumption that there is no effect, no difference, or no relationship in the population. The alternative hypothesis (H₁) proposes the opposite. Your p value tells you how compatible your data are with H₀.

Interpreting the p value

A p value is always interpreted against a significance level (α), most commonly set at 0.05.

• p ≤ α: Reject the null hypothesis. The result is statistically significant.
• p > α: Fail to reject the null hypothesis. The result is not statistically significant.

For example, a p value of 0.03 means there is a 3% probability of observing results this extreme if the null hypothesis were true — low enough, at α = 0.05, to reject H₀.

What a p value is not

A common misconception is that the p value represents the probability that the null hypothesis is true, or that it measures the probability that results occurred by chance. It does neither. The p value assumes H₀ is true from the outset — it does not assign a probability to that assumption. It also does not measure effect size or practical significance; a result can be statistically significant yet have little real-world importance, particularly with large sample sizes.

Key Concepts

1. Null and alternative hypotheses

Every hypothesis test begins with two competing statements. The null hypothesis (H₀) asserts no effect or no difference — it is the claim you are testing against. The alternative hypothesis (H₁) asserts that an effect or difference exists. Your p value is computed under the assumption that H₀ is true, so defining both hypotheses clearly before collecting data is essential.

2. Significance level (α)

The significance level is the threshold at which you will reject the null hypothesis. The most widely used value is α = 0.05, meaning you accept a 5% risk of incorrectly rejecting a true null hypothesis (a Type I error). Other common choices are α = 0.01 (stricter) and α = 0.10 (more lenient), depending on the consequences of a false positive in your field.

3. Test statistic

A test statistic is a single number calculated from your sample data that summarizes how far your observed results deviate from what H₀ predicts. Different tests produce different statistics — z, t, χ², and F are the most common. The larger the deviation from H₀, the more extreme the test statistic, and the smaller the resulting p value.

4. One-tailed vs. two-tailed tests

The direction of your alternative hypothesis determines whether your test is one-tailed or two-tailed.

• Two-tailed test: H₁ states the parameter is simply different from the null value (≠). Evidence in either direction counts against H₀. This is the default choice for most research.
• One-tailed test: H₁ states the parameter is specifically greater than (>) or less than (<) the null value. Only evidence in that one direction counts. A one-tailed test is appropriate only when the direction of the effect is specified in advance and there is a strong theoretical justification.

Methods to Find P Value

There are three practical ways to find a p value: using a formula and statistical table, applying software or a calculator, or reading it directly from output. The method you choose depends on your context — an exam, a research workflow, or a quick check.

Method 1: Formula and statistical table

This is the foundational approach. You calculate a test statistic from your data, then use a distribution table to convert that statistic into a p value.

The general test statistic formula is:$$\text{Test Statistic} = \frac{\text{Observed value} – \text{Null hypothesis value}}{\text{Standard error}}$$Test Statistic=Standard errorObserved value−Null hypothesis value​

Once you have the test statistic, you locate it in the relevant distribution table — z, t, chi-square, or F — and read off the corresponding probability. The steps differ slightly by test type, covered in detail in the next section.

When to use it: Coursework, exams, or when you need to understand the mechanics of the calculation.

Method 2: Online calculator

Several reliable calculators compute p values instantly from your test statistic and degrees of freedom, without requiring manual table lookups.

Recommended tools include:

• GraphPad QuickCalcs — straightforward interface for t, z, chi-square, and F tests
• Social Science Statistics Calculator — supports a wide range of tests with clear input fields
• Statology P Value Calculators — organized by test type with plain-language output

Input your test statistic, select your distribution, specify degrees of freedom where required, and the calculator returns the exact p value.

When to use it: Quick verification, teaching demonstrations, or when working without statistical software.

Method 3: Statistical software

Statistical software computes p values automatically as part of full test output, eliminating manual calculation entirely.

• R: Functions such as t.test(), chisq.test(), and aov() return p values directly in their output.
• Python (SciPy): Functions including scipy.stats.ttest_ind(), scipy.stats.chi2_contingency(), and scipy.stats.f_oneway() return a test statistic and p value as a paired result.
• Excel: The functions T.TEST(), CHISQ.TEST(), and F.TEST() return p values for common hypothesis tests.

When to use it: Real research, large datasets, or any analysis where reproducibility and precision matter.

Step-by-Step Examples

Example 1: Z-test (large sample, known population standard deviation)

Scenario: A manufacturer claims its light bulbs last an average of 1,000 hours. You test a sample of 50 bulbs and find a mean of 980 hours. The population standard deviation is known to be 80 hours. Test at α = 0.05 whether the mean differs from the claimed value.

Step 1: State the hypotheses

• H₀: μ = 1,000
• H₁: μ ≠ 1,000 (two-tailed test)

Step 2: Calculate the test statistic$$z = \frac{\bar{x} – \mu_0}{\sigma / \sqrt{n}} = \frac{980 – 1000}{80 / \sqrt{50}} = \frac{-20}{11.31} = -1.77$$

Step 3: Find the p value

For a two-tailed z-test, look up |z| = 1.77 in the standard normal table. The area in one tail is 0.0384. Double it for two tails:$$p = 2 \times 0.0384 = 0.0768$$

Step 4: Interpret the result

p = 0.0768 > α = 0.05. Fail to reject H₀. There is insufficient evidence that the mean bulb life differs from 1,000 hours.

Example 2: T-test (small sample, unknown population standard deviation)

Scenario: A nutritionist believes a new diet reduces systolic blood pressure. Eight patients are tested before and after the diet. The mean reduction is 8 mmHg with a sample standard deviation of 6 mmHg. Test at α = 0.05 whether blood pressure decreased.

Step 1: State the hypotheses

• H₀: μ = 0 (no reduction)
• H₁: μ > 0 (one-tailed test)

Step 2: Calculate the test statistic$$t = \frac{\bar{x} – \mu_0}{s / \sqrt{n}} = \frac{8 – 0}{6 / \sqrt{8}} = \frac{8}{2.12} = 3.77$$

Step 3: Find the p value

Degrees of freedom: df = n − 1 = 7. Look up t = 3.77 in the t-distribution table at df = 7. The value falls beyond t = 3.499 (p = 0.005, one-tailed), so:$$p < 0.005$$

Step 4: Interpret the result

p < 0.005 < α = 0.05. Reject H₀. There is significant evidence that the diet reduces systolic blood pressure.

Example 3: Chi-square test (categorical data)

Scenario: A researcher surveys 200 people on their preferred social media platform — Twitter, Instagram, or Facebook — to determine whether preferences are evenly distributed. Test at α = 0.05.

Platform	Observed (O)	Expected (E)	(O − E)² / E
Twitter	60	66.67	0.67
Instagram	90	66.67	8.17
Facebook	50	66.67	4.17
Total	200	200	13.01

Step 1: State the hypotheses

• H₀: Social media preferences are evenly distributed
• H₁: Social media preferences are not evenly distributed

Step 2: Calculate the test statistic$$\chi^2 = \sum \frac{(O – E)^2}{E} = 13.01$$

Step 3: Find the p value

Degrees of freedom: df = k − 1 = 2. Look up χ² = 13.01 in the chi-square table at df = 2. The value exceeds 10.597 (p = 0.005), so:$$p < 0.005$$

Step 4: Interpret the result

p < 0.005 < α = 0.05. Reject H₀. Social media preferences are not evenly distributed across the three platforms.

Example 4: F-test (comparing variance across groups)

Scenario: A researcher tests whether three teaching methods produce different mean exam scores. Group A (n = 10) has a mean of 78, Group B (n = 10) a mean of 85, and Group C (n = 10) a mean of 82. The between-group variance is 136.67 and the within-group variance is 30.11. Test at α = 0.05.

Step 1: State the hypotheses

• H₀: μ_A = μ_B = μ_C (all group means are equal)
• H₁: At least one group mean differs

Step 2: Calculate the test statistic$$F = \frac{\text{Between-group variance}}{\text{Within-group variance}} = \frac{136.67}{30.11} = 4.54$$

Step 3: Find the p value

Degrees of freedom: df₁ = k − 1 = 2 (between groups), df₂ = N − k = 27 (within groups). Look up F = 4.54 in the F-distribution table. The value falls between the critical values for p = 0.05 (F = 3.35) and p = 0.01 (F = 5.49), so:$$0.01 < p < 0.05$$

Step 4: Interpret the result

p < 0.05 = α. Reject H₀. There is significant evidence that at least one teaching method produces a different mean exam score.

How to Interpret the P Value

Calculating a p value is only half the work. Interpreting it correctly — and avoiding common misreadings — is what makes the result meaningful.

The decision rule

Every p value is interpreted against a pre-specified significance level (α). The decision rule is straightforward:

Result	Conclusion
p ≤ α	Reject H₀. The result is statistically significant.
p > α	Fail to reject H₀. The result is not statistically significant.

Note the precise language: you never “accept” the null hypothesis. Failing to reject it simply means your data did not provide sufficient evidence against it.

Strength of evidence

Beyond the binary reject/fail-to-reject decision, the size of the p value conveys the strength of evidence against H₀. The table below offers a widely used informal guide:

P Value	Strength of Evidence Against H₀
p > 0.10	Little to none
0.05 < p ≤ 0.10	Weak
0.01 < p ≤ 0.05	Moderate
0.001 < p ≤ 0.01	Strong
p ≤ 0.001	Very strong

This scale is a guide, not a rule. Different fields apply different thresholds depending on the consequences of a Type I error.

Statistical significance vs. practical significance

A statistically significant result is not automatically a meaningful one. With a large enough sample size, even a trivially small difference can produce a very small p value. For example, a study of 50,000 people might find that a new drug reduces blood pressure by 1 mmHg with p < 0.001 — highly significant statistically, but clinically irrelevant.

Always pair a p value with a measure of effect size — such as Cohen’s d, Pearson’s r, or eta-squared (η²) — to assess whether a statistically significant result carries practical importance.

One-tailed vs. two-tailed interpretation

The tail direction affects the p value directly. A one-tailed test concentrates all of α in one direction, making it easier to achieve significance — but only if the effect falls in the predicted direction. A two-tailed test splits α across both tails and is the conservative, appropriate default for most research questions. Switching from two-tailed to one-tailed post hoc to push a borderline result below 0.05 is a form of p-hacking and invalidates the test.

What p values cannot tell you

Misinterpretation of p values is widespread. Keep the following boundaries in mind:

• The p value is not the probability that H₀ is true. It assumes H₀ is true in order to calculate the probability of your data.
• The p value is not the probability that your results occurred by chance. Chance is not a competing hypothesis with a measurable probability.
• A non-significant result does not prove H₀. Absence of evidence is not evidence of absence.
• The p value does not measure the size or importance of an effect. A p value of 0.001 does not mean the effect is larger than one with p = 0.04.

Applications of P Value

FAQs