Critical Value Calculator

Most students reach for the Z-distribution when they see n ≥ 30. That is wrong. The choice between Z and t depends on whether you know the population standard deviation (σ) or estimate it from your sample (s). Sample size alone does not decide it. This critical value calculator gives you the exact threshold for Z, t, chi-square, and F distributions, no tables, no interpolation, no guessing which tail to use.

You pick your test, set alpha and degrees of freedom, and choose one-tailed or two-tailed. The calculator returns the critical value and tells you how to use it: reject the null if your test statistic falls beyond this boundary. It covers the four distributions you encounter in intro statistics and applied work, all on one page.

Open the calculator and start with the distribution selector. If you know σ, pick Z (Standard Normal). If you estimate σ from your sample, pick t (Student's t). For chi-square or F, the choice is driven by your test: goodness-of-fit or independence gets chi-square; ANOVA or variance comparison gets F.

Next, set your confidence level (for Z and t) or significance level (α, for chi-square and F). A 95% confidence level equals α = 0.05. The calculator uses α internally, so entering either one gives the same result. Choose test type: two-tailed splits α across both tails; right-tailed puts all α in the upper tail; left-tailed puts it in the lower tail. For chi-square and F, hypothesis tests almost always use right-tailed.

Enter degrees of freedom where needed. For the t-distribution, that is n − 1 for a one-sample test. For chi-square, it depends on the test: k − 1 for goodness-of-fit, (rows − 1) × (columns − 1) for independence. For the F-distribution, you need two degrees of freedom: numerator df (among groups) and denominator df (within groups).

Click Calculate. The result shows the critical value and a decision rule: reject the null if your test statistic exceeds this value (right-tailed), falls below it (left-tailed), or falls outside ± it (two-tailed). The interpretation panel writes the rule in plain English so you can copy it into your report.

A critical value is the boundary in the sampling distribution that separates the rejection region from the non-rejection region. If your test statistic crosses that boundary, you reject the null hypothesis at the chosen significance level. Think of it as the line you draw before you look at your sample data: “If my observed value is more extreme than this threshold, the result is statistically significant.”

How It Works in Practice

The picture below shows a two-tailed Z-test at α = 0.05. The shaded area in each tail holds 2.5% of the distribution. The critical values are ±1.96. Any test statistic more extreme than 1.96 or less than −1.96 falls in the rejection region.

The critical value changes with alpha, test type, and distribution. A smaller alpha moves the boundary further into the tail, making it harder to reject the null. A one-tailed test at α = 0.05 has its critical value at the 95th percentile (or 5th percentile), not the 97.5th. For asymmetric distributions like chi-square and F, the critical value for a right-tailed test is always positive and increases with degrees of freedom.

Where the Numbers Come From

Sources: NIST/SEMATECH e-Handbook of Statistical Methods, section 1.3.6.7, provides exact tables for all four distributions. OpenIntro Statistics (4th ed.) and Moore, McCabe & Craig, Introduction to the Practice of Statistics, define the standard decision rules.

Pick the distribution based on what you are testing, not on sample size.

Z (Standard Normal)

Use the Z (Standard Normal) distribution when the population standard deviation (σ) is known. This happens in textbook problems that give you σ, in quality control where σ is known from process history, and in large-sample tests of proportions (where the standard deviation is derived from the null proportion). Do not use Z just because n ≥ 30. If you have a sample size of 1000 but σ is unknown, you still use t.

t (Student's t)

Use the t (Student's t) distribution when σ is unknown and you estimate it from the sample standard deviation (s). This applies at any sample size, including n = 5. The t-distribution has heavier tails than the normal, meaning it produces larger critical values for small degrees of freedom. As df increases, the t critical value approaches the Z critical value. At df = 120, the difference is less than 0.01.

Chi-Square (χ²)

Use the chi-square (χ²) distribution for tests about a single variance, goodness-of-fit, or independence in contingency tables. The chi-square distribution is asymmetric: it is bounded at zero and has a long right tail. Hypothesis tests always use the right-tail critical value. A two-tailed chi-square test is unusual but available; it returns a lower critical value (near zero) and an upper critical value (the main threshold).

F-Distribution

Use the F-distribution for ANOVA and for comparing two variances (the F-test for equality of variances). The F-distribution has two degrees of freedom: numerator df (number of groups minus one) and denominator df (total sample size minus number of groups). Like chi-square, it is asymmetric and bounded at zero. The right-tail critical value is the one you want for hypothesis testing.

The critical value approach and the p-value approach are two ways to make the same decision. They never contradict each other if you apply them correctly.

In the critical value approach, you compute your test statistic and compare it to the threshold from the distribution. If the test statistic is more extreme than the critical value, you reject the null. This is the method this calculator supports. It is the method most textbooks teach first because it gives you a clear, reusable boundary.

In the p-value approach, you compute the probability of observing a test statistic as extreme as (or more extreme than) yours, assuming the null is true. If that p-value is less than alpha, you reject the null. The p-value is the exact tail area beyond your observed test statistic. It gives you a continuous measure of evidence rather than a binary reject/fail-to-reject decision.

The two approaches are mathematically equivalent. When your test statistic exactly equals the critical value, the p-value equals alpha. When your test statistic exceeds the critical value, the p-value is smaller than alpha. The ASA Statement on Statistical Significance and P-Values (Wasserstein & Lazar, 2016) reminds us that neither approach measures practical importance: statistical significance is not the same as a meaningful effect size.

Use the critical value approach when you want a quick decision rule for a fixed alpha. Use the p-value approach when you need to report the strength of evidence or when you are working with a non-standard alpha. A rejection region calculator like this one gives you the threshold; a p-value calculator gives you the tail area. Both are valid, and both depend on the same distribution, alpha, and degrees of freedom.