Critical Value Calculator

Compute critical values for hypothesis testing and confidence intervals. Choose from Z (normal), T (Student's t), Chi‑Square, or F distributions. Adjust significance level (α), degrees of freedom, and test direction (one‑tailed or two‑tailed). Visualize the distribution and critical region on an interactive plot.

df₂ (for F)
Enter your computed test statistic to get the p‑value.
Quick presets:
Z 95% two‑tailed
Z 99% two‑tailed
T 95% two‑tailed (df=10)
χ² 95% right‑tailed (df=5)
F 95% right‑tailed (df₁=3, df₂=10)
Privacy first: All calculations run locally in your browser. No data is sent to any server.

What Is a Critical Value in Statistics?

In statistical hypothesis testing, a critical value is a point on the scale of the test statistic that separates the rejection region (where the null hypothesis is rejected) from the non‑rejection region. It serves as a threshold: if the computed test statistic exceeds the critical value (in absolute value for two‑tailed tests), the result is deemed statistically significant at the chosen significance level α.

Critical values are derived from the cumulative distribution function (CDF) of the test statistic's distribution. For a given significance level α and test direction (one‑tailed or two‑tailed), the critical value c satisfies:

P(T > c) = α for a right‑tailed test,
P(T < c) = α for a left‑tailed test,
P(|T| > c) = α for a two‑tailed test.

The choice of distribution depends on the test statistic and the assumptions about the data. This calculator supports the four most common distributions used in applied statistics: the standard normal (Z), Student's t, chi‑square (χ²), and F distributions.

Understanding the Four Distributions

Z Distribution (Standard Normal)

The Z distribution is the standard normal distribution with mean 0 and variance 1. It is used when the population variance is known, or when the sample size is large enough (typically n ≥ 30) for the Central Limit Theorem to apply. Z‑tests are common in quality control, A/B testing, and large‑sample inference.

Common uses: Z‑tests for means and proportions, confidence intervals for large samples, and comparing two population means with known variances.

T Distribution (Student's t)

The t distribution, developed by William Sealy Gosset (under the pseudonym "Student"), is used when the population standard deviation is unknown and estimated from the sample. It has heavier tails than the normal distribution, accounting for the extra uncertainty from estimating the variance. As degrees of freedom increase, the t distribution approaches the normal distribution.

Common uses: One‑sample t‑tests, two‑sample t‑tests, paired t‑tests, and confidence intervals for small samples.

Chi‑Square Distribution (χ²)

The chi‑square distribution is the sum of the squares of k independent standard normal variables. It is right‑skewed and takes only non‑negative values. The chi‑square distribution is used in tests of independence, goodness‑of‑fit tests, and for testing the variance of a normal population.

Common uses: Chi‑square test of independence, goodness‑of‑fit test, test for a single variance, and in the construction of confidence intervals for the variance.

F Distribution

The F distribution, named after Sir Ronald Fisher, is the ratio of two independent chi‑square variables divided by their respective degrees of freedom. It is used primarily in analysis of variance (ANOVA) and regression analysis to compare variances or test the equality of multiple group means.

Common uses: ANOVA, tests of overall significance in regression, test for equality of two variances (F‑test), and model comparison.

How to Use This Calculator

  1. Select the distribution that matches your test statistic (Z, T, χ², or F).
  2. Choose the significance level α — common values are 0.05 (95% confidence), 0.01 (99% confidence), etc. You can also enter a custom α.
  3. Select the test direction: one‑tailed (left or right) or two‑tailed.
  4. Enter the degrees of freedom (df). For the F distribution, you need both df₁ (numerator) and df₂ (denominator).
  5. Click "Compute Critical Value" to obtain the threshold and view the distribution plot with the critical region shaded.
  6. Optional: Enter your computed test statistic and click "Compute p‑value" to obtain the exact p‑value and interpret the result.
  7. Interpret the sign: For left‑tailed tests, the critical value is negative (except for χ² and F, which are non‑negative); for right‑tailed tests, it is positive. The calculator automatically handles these cases based on your selected direction.

Step‑by‑Step Example: Z‑Test for a Population Mean

Example: Two‑Tailed Z‑Test at α = 0.05

Suppose you are testing whether a new manufacturing process has changed the mean diameter of ball bearings from the known historical mean of 10 mm. You take a sample of 50 bearings, compute the sample mean, and calculate a Z‑statistic. To determine if the result is statistically significant, you compare your Z‑statistic to the critical value.

Settings: Distribution = Z, α = 0.05, two‑tailed test.
Critical value: The calculator returns ±1.95996. This means that if your computed Z‑statistic is less than −1.96 or greater than +1.96, you reject the null hypothesis at the 5% significance level.

The shaded area in the distribution plot represents the rejection region: each tail has an area of α/2 = 0.025, totaling 0.05.

Example: One‑Way ANOVA with F‑Test

In a one‑way ANOVA comparing the means of three groups, the F‑statistic follows an F distribution with df₁ = (k − 1) = 2 and df₂ = (N − k) = 27. At α = 0.05, the critical F value is approximately 3.35. If the computed F‑statistic exceeds this value, the null hypothesis of equal group means is rejected.

Settings: Distribution = F, α = 0.05, right‑tailed test (ANOVA is always right‑tailed), df₁ = 2, df₂ = 27.
Critical value: The calculator returns the F critical value, which serves as the threshold for significance.

The Mathematics Behind the Critical Values

The critical value is the inverse of the cumulative distribution function (CDF) evaluated at a specific probability. For a given distribution with CDF F(x), the critical value c for a right‑tailed test is:

c = F−1(1 − α)

For a two‑tailed test, the critical values are:

clower = F−1(α/2),    cupper = F−1(1 − α/2)

This calculator uses robust numerical methods to compute these inverse CDF values. For the Z distribution, we use an efficient rational approximation (Moro's algorithm). For the T, χ², and F distributions, we use a combination of numerical integration and root‑finding (bisection method) to achieve high accuracy across a wide range of parameter values.

All algorithms have been validated against reference tables and statistical software (R, Python SciPy) to ensure accuracy. The source references are listed in the authority note below.

Common Misconceptions About Critical Values

  • "A smaller critical value always means a more significant result." — Not exactly. The critical value depends on α, df, and the distribution. A smaller critical value (in absolute terms) means the rejection region is larger, making it easier to reject the null (i.e., a less stringent test).
  • "Critical values are the same as p‑values." — No. A critical value is a fixed threshold determined by α; a p‑value is the probability of observing a test statistic as extreme as, or more extreme than, the one computed from the data. The p‑value is compared to α to make a decision.
  • "The t distribution is always used for small samples." — While the t distribution is preferred for small samples (n < 30) when σ is unknown, it can also be used for larger samples. The t distribution approaches the normal distribution as df → ∞.
  • "The F distribution is only for ANOVA." — ANOVA is the most common use, but the F distribution is also used in regression model comparison, testing equality of variances, and in many other contexts.
  • "Critical values are unrelated to confidence intervals." — In fact, the critical value is the boundary of the confidence interval. For a two‑tailed test, the confidence interval is [statistic ± c] (or its asymmetric counterpart for χ²/F), and the critical value directly determines the margin of error.

Practical Applications Across Fields

  • Medical Research: Comparing the effectiveness of treatments using t‑tests or ANOVA.
  • Quality Control: Z‑tests for monitoring production processes and detecting shifts in mean or variability.
  • Social Sciences: Chi‑square tests for survey data and categorical variable associations.
  • Finance: Testing whether a portfolio's returns differ from a benchmark using t‑tests.
  • Engineering: F‑tests for comparing variances in product testing.
  • Machine Learning: F‑tests for feature selection in ANOVA‑based ranking, or Z‑tests for comparing model performance metrics across different runs.

Rooted in statistical theory — This tool implements algorithms based on foundational works in mathematical statistics. The numerical methods are derived from Abramowitz & Stegun (1964) "Handbook of Mathematical Functions," Press et al. (2007) "Numerical Recipes," and the authoritative algorithms implemented in the R statistical environment and SciPy. Reviewed by the GetZenQuery tech team. Last updated June 2026.

Frequently Asked Questions

A one‑tailed test (left or right) examines whether a parameter is greater than or less than a certain value. A two‑tailed test examines whether a parameter is different from a certain value (either direction). For a given α, the critical value for a two‑tailed test is larger in magnitude than for a one‑tailed test because α is split between both tails.

Choose Z when the population variance is known or sample size is large (n ≥ 30). Choose T when the population variance is unknown and estimated from the sample (common in most real‑world scenarios). Choose χ² for goodness‑of‑fit, independence, or variance tests. Choose F for ANOVA, comparing variances, or testing multiple regression coefficients.

Degrees of freedom (df) represent the number of independent pieces of information available to estimate a parameter. For the t distribution, df = n − 1. For the χ² distribution, df = number of categories − 1 or (rows − 1)(columns − 1) in a contingency table. For the F distribution, df₁ is the numerator degrees of freedom (e.g., groups − 1) and df₂ is the denominator degrees of freedom (e.g., total observations − groups).

The calculations use double‑precision floating‑point arithmetic with high‑accuracy numerical methods. Typical errors are less than 1×10⁻⁹ for Z, 1×10⁻⁶ for T and χ², and 1×10⁻⁵ for F across a wide range of parameters. For most practical applications, this is more than sufficient.

This calculator is designed for parametric tests based on the Z, T, χ², and F distributions. Non‑parametric tests (e.g., Mann‑Whitney, Wilcoxon) use different distributions. However, some non‑parametric tests use approximations that rely on the normal distribution for large samples.

Visit authoritative resources like Khan Academy Statistics, Penn State STAT, or the textbook "Statistical Inference" by Casella & Berger. For a deeper dive, consult "Introduction to the Practice of Statistics" by Moore, McCabe, and Craig.

In that case, the p‑value equals α. This is a marginal case. Typically, researchers may reject the null hypothesis at the exact α level, but it is considered borderline. It is best practice to report the exact p‑value and state that the result is significant at the α level (if α is pre‑defined) or note the marginality.
References: MathWorld: Normal Distribution; Abramowitz, M. & Stegun, I.A. (1964). Handbook of Mathematical Functions; Press, W.H. et al. (2007). Numerical Recipes, 3rd ed.; Wikipedia: Critical Value.