Paired T-Test Calculator

A paired t-test compares two measurements taken on the same subjects. Paste the before and after columns and this works out the within-pair differences for you, then returns their mean and standard deviation, t on n minus one degrees of freedom, both p-values, a confidence interval for the mean difference and Cohen's d_z. It also prints the correlation between the pairs and what the same numbers would have given as an unpaired test, which is where the power of pairing shows up. A third tab rebuilds the test from published means, standard deviations and a correlation when you have a table rather than the data.

Calculator Type

Paste the before and after readings in matching order. The differences, their mean and standard deviation, t, the p-values, the confidence interval and the effect size all come from the raw numbers.

First Measurement of Each Pair (before) *

Commas, spaces, tabs or new lines all work.

Second Measurement of Each Pair (after) *

Must hold the same number of values, in the same order.

Mean Difference You Are Testing Against

Almost always zero, meaning no change.

Confidence Level (%)

Which Test

Widget

When you have a table rather than the data. The standard deviation of the differences is rebuilt from the two column standard deviations and the correlation between the pairs.

Mean of the First Measurement *

Standard Deviation of the First *

Mean of the Second Measurement *

Standard Deviation of the Second *

Number of Pairs *

Correlation Between the Pairs *

The Pearson r between the two columns. Without it the paired test cannot be reconstructed from a summary table at all.

Mean Difference You Are Testing Against

Confidence Level (%)

Which Test

Widget