Critical Value Calculator

A critical value is a numerical value that defines the boundary of the rejection region in hypothesis testing. If the test statistic exceeds this value, then the results are statistically significant and the null hypothesis should be rejected. This calculator allows you to calculate critical values for the four most commonly used distributions in tests: the standard normal distribution (Z), Student's t-distribution, chi-square distribution, and F-distribution. It can handle both one-sided and two-sided tests.

What is a critical value?

Any hypothesis test begins with the significance level, usually denoted by the Greek letter alpha. This is the probability of incorrectly rejecting the null hypothesis. In practice it is almost always set to 0.05 or 0.01. The critical value converts this probability into a cut-off point on the distribution of the test statistic.

More precisely, a critical value is the point at which the test statistic would have to be in order for us to reject the null hypothesis. All results more extreme than that point are in the rejection region. So the critical value and significance level represent the same threshold from two different perspectives. One measured by distance and one by tail area.

Distribution selection:

The correct critical value depends on what distribution the test statistic follows. The four distributions listed here cover most tests at introductory and applied levels.

The Z-distribution and the t-distribution are symmetric about zero so the critical values for a two-sided test are mirror images of each other and correspond to the same values with opposite signs. The chi-square distribution and F-distribution are asymmetric and exist only in the positive half of the number line, therefore there will be two different bounds shown for a two-sided test: a lower bound and an upper bound.

One-tailed or two-tailed test?

The type of test determines how the significance level is distributed across the distribution. Q represents the quantile function, which is the inverse of the cumulative distribution function.

In a right-sided test, the entire area of α is placed in the upper region.

In a left-sided test, α is placed in the lower region.

In a two-sided test, α is split evenly between both tails, with each tail having half of it.

For symmetric distributions the lower bound is exactly the negative of the upper bound so often only one number with a sign is given, e.g., 1.96.

How to use this calculator:

Select the desired distribution, enter the significance level and select whether the test is one-sided (left or right) or two-sided. If degrees of freedom are required for a t-, chi-square- or F-distribution, then enter them. The critical values will be displayed immediately, as well as the lower bound for a two-sided test and the corresponding confidence level.

This tool can also be used in reverse. If the critical value field is left blank, this tool will calculate that value. If you enter a critical value and clear the significance level, you can back-calculate the corresponding alpha. This is the p-value for either a one-sided or two-sided hypothesis for this test statistic.

An example of a calculation:

Suppose you are doing a two-sided t-test and the sample contains 11 observations. So, the degrees of freedom is 10 and significance level is 0.05. As it's a two-sided test, each end value will be 0.025 and we need the 0.975 quantile of the t-distribution (with 10 degrees of freedom).

Thus, the two critical values are approximately at ±2.228. If the calculated t-value is greater than 2.228 or less than -2.228, then the null hypothesis should be rejected with a significance level of 5%. By switching to a Z-distribution, the limits decrease to ±1.96. This is the version of the same test for large samples.

Commonly used critical values:

These numbers come up very often and are worth memorizing. The row for Z also represents the limit of the t-distribution for large samples.

Application areas for critical values:

Critical values exist not only in textbooks but support real decisions. In quality control they are used to indicate whether a batch of product meets specifications. In medical research facilities they are used to indicate whether results are unusual. In finance and environmental monitoring critical values are used to assess whether an actual change has occurred or if it is just noise. In all cases, critical values represent the boundary between normal results and those where action is warranted.

Tips for correct calculation:

Before interpreting values, make sure the direction of your test matches your hypothesis. For claims that have a specific direction, like "the new method is faster", you need a one-sided test. If you're just claiming something is different, however, then you need a two-sided test. Also, make sure degrees of freedom match your sample size. In a t-test this is usually the sample size minus one. Finally, note that a smaller alpha means the statistic has to be more extreme, so a test for 0.01 will be harder to meet than a test for 0.05.

This calculator is for educational and general analysis purposes only. It provides critical values from standard distributions, but does not replace the judgment of a qualified statistician in risk management tasks.