Skip to content
Statculator
P-Values & Critical Values

P-Value Calculator

Free p-value calculator: turn a z, t, chi-square, F or r statistic into a p-value for a left, right or two-tailed test, with steps and the shaded curve.

Free, no sign-up Runs in your browserBy Hira Noor, BS MathematicsUpdated

Try

Pick the distribution your test statistic follows when H₀ is true.

Tail

Left for H₁: “less than”, right for H₁: “greater than”, two-tailed for H₁: “not equal to”.

Leave blank to get the p-value without a decision.

Results

p-value (two-tailed)
0.0251
p = 2 × P(Z ≥ |2.24|)
Critical z at α = 0.05
±1.96

Significant at α = 0.05: reject H₀

p = 0.0251 ≤ 0.05. A result this extreme would be unusual if H₀ were true.

Null distribution
Z ~ N(0, 1)
Left tail, P(Z ≤ z)
0.9875
Right tail, P(Z ≥ z)
0.0125
Two-tailed, 2 · P(Z ≥ |z|)
0.0251

Interpretation: Assuming H₀ is true, a z statistic at least as extreme as the one observed would occur about once in every 40 repetitions. That is moderate evidence against H₀.

normal distribution curve with the shaded probability region, markers at z = 2.24, -2.24-2.24z = 2.24-4-3-2-101234Standard normal distribution, Z ~ N(0, 1)
Shaded area = p-value = 0.0251 (both tails together). Dashed line: your statistic.

Decision at common significance levels (two-tailed)

αCritical zReject H₀?
0.1±1.6449Yes
0.05±1.96Yes
0.025±2.2414No
0.01±2.5758No
0.005±2.807No
0.001±3.2905No

p = 0.0251. “Yes” means p ≤ α, which is the same as the statistic landing in the rejection region.

Step-by-step solution

  1. 1

    Identify the null distribution

    When H₀ is true, the z statistic follows the standard normal distribution.

    Z ~ N(0, 1)
  2. 2

    Write the p-value as a probability

    The curve is symmetric, so the two-tailed p-value is twice the area beyond |statistic|.

    p = 2 × P(Z ≥ |2.24|)
  3. 3

    Find the tail areas and the p-value

    Each area comes from the cumulative distribution function (CDF) of the null distribution.

    P(Z ≥ 2.24) = 0.012545
    p = 2 × 0.012545 = 0.0251
  4. 4

    Decision: reject H₀

    p = 0.0251 ≤ α = 0.05
    Reject H₀ if |z| ≥ 1.96; observed z = 2.24

What is a P-Value Calculator?

A p-value calculator turns a test statistic into a p-value: the probability, if the null hypothesis H₀ is true, of a result at least as extreme as the one you observed. Enter a z, t, chi-square (χ²), F or Pearson r statistic, pick a left-tailed, right-tailed or two-tailed test, and the calculator returns the exact p-value, shades it on the curve and, if you give α, tells you whether to reject H₀.

What the P-Value Calculator does

Choose the type of statistic, enter its value and degrees of freedom, and pick the tail. The calculator finds the p-value from the correct null distribution, shades that area on the curve, lists the left, right and two-tailed areas, and, if you give α, returns the critical value and the reject or fail-to-reject decision. A table shows the decision at six common α levels at once, and the steps show every area used.

It is the last step of any hypothesis test. Use it to calculate a p-value from a statistic you worked out by hand, when software prints a statistic without a p-value, or to check a reported result such as “t(15) = 2.24, p = .041”. People also look for it as a p value finder, a p score calculator, or a test statistic to p-value calculator; all of them mean the same thing.

No test statistic yet? Start from the data instead: compare two group means with the two-sample t-test calculator, two rates or percentages with the two-proportion z-test calculator, or counts in a table with the chi-square test calculator. Each one reports its own p-value.

How to use the P-Value Calculator

  1. Select the Type of test statistic: z, t, χ², F or r. The table further down shows which one your test produces.
  2. Enter the statistic and the fields that appear for it: Degrees of freedom (df) for t and χ², Numerator df (df₁) and Denominator df (df₂) for F, or Sample size (n pairs) for r.
  3. Choose the Tail. z, t and r default to two-tailed; χ² and F default to right-tailed because that is how goodness-of-fit, independence, ANOVA and regression tests work.
  4. Optionally enter the Significance level (α). Read the p-value, the decision and the shaded curve, and open the steps to see the probability statement and the areas used.

How to find a p-value from a test statistic

Whether you use this calculator, a printed table or software, finding a p-value always takes the same four steps. The calculator does steps 2 to 4 for you.

  1. Compute the test statistic from your data, for example t = (x̄ − μ₀) / (s / √n) for a one-sample t-test.
  2. Identify its null distribution and degrees of freedom: standard normal for z, t(df) for t, χ²(df) for chi-square, F(df₁, df₂) for F.
  3. Pick the tail from the alternative hypothesis: “less than” is left-tailed, “greater than” is right-tailed, “not equal to” is two-tailed.
  4. Find the tail area beyond the statistic with the CDF. That area is the p-value; compare it with α to decide.

P-value formulas for each tail

Let X be the statistic’s null distribution, F its cumulative distribution function (CDF) and x the observed value. Every p-value is one of these areas.

Left-tailed
p = P(X ≤ x) = F(x)
Right-tailed
p = P(X ≥ x) = 1 − F(x)
Two-tailed, symmetric statistic (z, t, r)
p = 2 × P(X ≥ |x|)
Two-tailed, skewed statistic (χ², F)
p = 2 × min{ F(x), 1 − F(x) }
Correlation r converted to t
t = r √(n − 2) / √(1 − r²), df = n − 2
Formula symbols
SymbolMeaning
xobserved value of the test statistic
Xrandom variable with the statistic’s null distribution
F(x)cumulative distribution function, P(X ≤ x)
df, df₁, df₂degrees of freedom that pin down the t, χ² or F distribution

The p-value is not the probability that H₀ is true. It is the probability of data this extreme if H₀ were true. See What Is a P-Value? Meaning, Calculation and Common Mistakes for a full explanation.

Source for these definitions: the NIST/SEMATECH e-Handbook of Statistical Methods, section 7.1.3.1.

Worked example

Same number, two distributions: t = 2.24 or z = 2.24?

Problem. A regression with 17 observations and one predictor reports a slope t statistic of 2.24 with 15 residual degrees of freedom. Find the two-tailed p-value. What would it be if 2.24 were wrongly treated as a z statistic?

  1. 01The slope statistic follows a t distribution with df = 15 under H₀: slope = 0.
  2. 02Two-tailed statement: p = 2 × P(T ≥ 2.24) with T ~ t(15).
  3. 03Upper-tail area: P(T ≥ 2.24) = 0.020331, so p = 2 × 0.020331 = 0.0407.
  4. 04As a z statistic: P(Z ≥ 2.24) = 0.012545, so p = 2 × 0.012545 = 0.0251.

Answer. The correct p-value is 0.0407, which is significant at α = 0.05 (the critical t is ±2.1314). Using the normal curve would understate the p-value by more than a third because it ignores the extra uncertainty of estimating the standard error from 17 observations.

Which distribution does my test statistic follow?

The statistic’s name usually tells you the distribution, and the test tells you the degrees of freedom. Use this table to set the calculator up, or open the single-statistic page in the last column for more detail on one distribution.

StatisticNull distributionParametersUsual tailProduced byDedicated page
zStandard normalnoneanyz-tests for means with known σ, proportion tests, large-sample rank testsP-value from z
tStudent tdfanyt-tests, regression coefficientsP-value from t
χ²Chi-squaredfrightgoodness of fit, independence, one-variance tests (any tail)P-value from χ²
FFdf₁, df₂rightANOVA, overall regression F, two-variance test (any tail)P-value from F
rt after conversionn (df = n − 2)anyPearson correlation test of ρ = 0r to p-value

What z, t, χ², F and r statistics measure

Each statistic is a different way of measuring how far the data sit from what H₀ predicts, which is why each needs its own reference curve.

  • z score: the distance between an estimate and its hypothesized value in standard errors, when the standard error is known or the sample is large. See the z-score calculator.
  • t score: the same distance when the standard error is estimated from the sample. Its curve has heavier tails than the normal curve, and they shrink as df grows.
  • χ² statistic: a sum of squared, scaled gaps between observed and expected counts (or a scaled sample variance). It cannot be negative, and bigger means a worse fit to H₀.
  • F statistic: a ratio of two variance estimates, such as between-group over within-group variance in ANOVA. Values near 1 agree with H₀.
  • Pearson r: the strength of a straight-line relationship, from −1 to 1. Testing ρ = 0 converts r to a t statistic with n − 2 df.

P-value from a mean and standard deviation

A p-value needs a test statistic, so a mean and standard deviation on their own are not enough: you also need the sample size n and the value μ₀ that H₀ claims. Compute t = (x̄ − μ₀) / (s / √n) with df = n − 1, then enter t here. Use z instead when σ is the known population standard deviation.

Example: a sample of n = 36 has mean x̄ = 103 and standard deviation s = 8, and H₀ says μ = 100. Then t = 3 / (8 / √36) = 3 / 1.333 = 2.25 with df = 35, and the two-tailed p-value is 0.0308. The one-sample t-test calculator and one-sample z-test calculator do the whole test from raw numbers or summary statistics.

How strong is the evidence? A p-value chart

The calculator’s interpretation line uses the bands below. They are a reading aid, not a rule: the American Statistical Association’s 2016 statement on p-values warns that a single cut-off such as 0.05 should not decide whether a finding is real.

P-value strength chart
p-valueEvidence against H₀Often written as
p ≤ 0.001Very strongp < .001 (***)
0.001 < p ≤ 0.01Strongp < .01 (**)
0.01 < p ≤ 0.05Moderatep < .05 (*)
0.05 < p ≤ 0.10Weakmarginal, not significant at 0.05
p > 0.10Little or nonen.s.

P-values in Excel, Google Sheets and on a TI-84

The same areas are available as built-in functions. The Excel formulas work unchanged in Google Sheets. Replace z, t, x, F and df with your own cells or numbers.

Statistic and tailExcel or Google SheetsTI-83 / TI-84
z, two-tailed=2*(1-NORM.S.DIST(ABS(z),TRUE))2 × normalcdf(|z|, 1E99, 0, 1)
z, left-tailed=NORM.S.DIST(z,TRUE)normalcdf(−1E99, z, 0, 1)
t, two-tailed=T.DIST.2T(ABS(t),df)2 × tcdf(|t|, 1E99, df)
t, right-tailed=T.DIST.RT(t,df)tcdf(t, 1E99, df)
χ², right-tailed=CHISQ.DIST.RT(x,df)χ²cdf(x, 1E99, df)
F, right-tailed=F.DIST.RT(F,df1,df2)Fcdf(F, 1E99, df1, df2)

Two-tailed p-values for χ² and F

The z and t curves are symmetric around 0, so a two-tailed p-value is simply twice the area beyond |statistic|. The χ² and F curves are skewed to the right and start at 0, so there is no mirror-image point. The standard fix is the equal-tail rule: compute both tail areas, take the smaller, and double it. The calculator shades that tail plus an equal area in the opposite tail.

In practice, two-tailed χ² and F p-values are needed only for tests about variances: whether one variance equals a claimed value (χ²) or whether two variances are equal (F). Goodness-of-fit, independence, ANOVA and regression tests measure “how far from H₀” in only one direction, so they are always right-tailed.

Common use cases

  • Finishing any hypothesis test computed by hand, from a t-test to a chi-square test or an ANOVA.
  • Checking p-values printed in papers or software against the reported statistic and degrees of freedom.
  • Getting exact p-values where a printed t-table, chi-square table or F-table would only give a range.
  • Getting exact p-values for a fractional df, such as a Welch t-test. Excel’s T.DIST functions truncate df to a whole number; this calculator does not.
  • Teaching how the same statistic gives different p-values under different tails and distributions.

Assumptions and limitations

  • The p-value is only as good as the assumption that the statistic follows the chosen distribution under H₀. Check the conditions of the original test (independence, normality, expected counts, equal variances).
  • Choose the tail before looking at the data. Switching from two-tailed to one-tailed after seeing the result halves the p-value and inflates false positives.
  • Running many tests and reporting only the significant ones makes small p-values meaningless; adjust α for multiple comparisons.
  • A p-value measures evidence against H₀, not the size or importance of an effect. Report an effect size or a confidence interval as well.

Frequently asked questions

What is a p-value in simple terms?

A p-value is the probability of getting a result at least as extreme as yours if the null hypothesis were true. A small p-value, such as 0.01, means your data would be surprising under H₀, which counts as evidence against it. It is not the probability that H₀ is true.

How do you calculate the p-value?

Compute the test statistic, find the distribution it follows when H₀ is true, and take the area beyond the statistic in the direction of the alternative hypothesis. For a two-tailed z or t test, double the area beyond |statistic|. For example, t = 2.24 with 15 df gives p = 2 × 0.0203 = 0.0407.

How do you calculate a p-value from a chi-square (χ²) statistic?

Use the right-tail area of the chi-square distribution with the test’s degrees of freedom: p = P(χ² ≥ observed). For χ² = 12.59 with df = 6, p = 0.0500. In Excel or Google Sheets, use =CHISQ.DIST.RT(12.59, 6). Pick χ² in the calculator above and leave the tail on right-tailed.

How do I calculate a p-value on a calculator like the TI-84?

Use the cumulative distribution functions in the DISTR menu with 1E99 as the far bound. For a right-tailed t test, tcdf(t, 1E99, df) gives the p-value; double it for a two-tailed test. normalcdf, χ²cdf and Fcdf work the same way for z, chi-square and F statistics.

How do I know whether to use a one-tailed or two-tailed p-value?

Use a two-tailed p-value when the alternative hypothesis is “not equal to”, which is the default in most research. Use a one-tailed p-value only if you predicted the direction before collecting data and an effect in the other direction would be treated the same as no effect.

Is a p-value of 0.05 significant?

By the usual rule, a result is significant when p ≤ α, so p = 0.05 is significant at α = 0.05 but not at α = 0.01. Values close to the cut-off are borderline either way and are best reported exactly rather than as “significant” or “not significant”.

Why are chi-square and F p-values usually right-tailed?

Chi-square goodness-of-fit and independence statistics, and ANOVA and regression F statistics, grow larger the further the data are from H₀ in any direction. Only large values count as evidence, so the p-value is the right-tail area. Two-tailed versions are used only for tests about variances.

Can a p-value be greater than 0.5?

Yes. A one-tailed p-value exceeds 0.5 whenever the statistic falls on the opposite side from the alternative hypothesis, for example z = −1 in a right-tailed test (p = 0.8413). A two-tailed p-value can be as large as 1 when the statistic is exactly at the centre.

What does p < 0.001 mean?

It means that, if H₀ were true, a result this extreme would occur less than once in a thousand repetitions of the study. Very small p-values are often reported as “p < .001” rather than as an exact number; this calculator gives the exact value in scientific notation.

Cite this page

Statculator. (2026, November 14). P-Value Calculator. Retrieved November 14, 2026, from https://statculator.com/p-values-critical-values/p-value-calculator/