Test Statistic Calculator

This calculator works out z test statistics with their p-values, critical values and a reject or do-not-reject decision. Choose a one-sample test for a mean, a two-sample test for the difference of two means, both with known population standard deviations, or a test for one proportion or two proportions. Enter the summary figures, the tail and the significance level, and you get the z statistic, standard error and p-value with the formula worked through. If the standard deviation comes from the sample itself, the right statistic is a t, not a z.

Calculator Type

The one-sample z test. Use it only when sigma is genuinely known rather than estimated from the sample; if it came from the sample, the statistic is a t.

Fix this before looking at the data, not after.
Widget

The two-sample z test for independent groups. Like the one-sample version, it needs each group's sigma to be genuinely known; if they come from the samples, use a two-sample t test.

Usually 0, for no difference between the populations.
Fix this before looking at the data, not after.
Widget

The standard error uses the hypothesised proportion, not the observed one, because that is what the null hypothesis assumes.

As a decimal between 0 and 1, so 50 percent is 0.5.
Fix this before looking at the data, not after.
Widget

Two independent groups, each with a count of successes out of a total. The two counts are pooled for the standard error, as the null hypothesis that the groups share one proportion requires.

Fix this before looking at the data, not after.
Widget