Skip to content
Omnikalc

Point-biserial correlation calculator

Enter the mean score of each group, the standard deviation of all the scores and the group sizes to get the point-biserial correlation, r squared, the t statistic and the two-sided p-value.

Updated on
Your inputs

Value out of range

Value out of range

Value out of range

Value out of range

Value out of range

Results

Point-biserial correlation

0.396

Two-sided p-value

0.0007

r squared (share of variance explained)
15.7 %
t statistic
3.555
Degrees of freedom
68
Equivalent Cohen’s d
0.862

The point-biserial correlation calculator gives the correlation between a yes-no variable and a score from the mean of each group, the standard deviation of all the scores and the group sizes, with r squared, the t statistic and the two-sided p-value.

How the calculation works

  • Enter the mean score of group 1 and of group 0.
  • Enter the standard deviation of all the scores and the size of each group.
  • Read the correlation, the t statistic and the p-value.

Example

With means of 78 and 70, a standard deviation of 10 and groups of 30 and 40, the point-biserial correlation is 0.396, which explains 15.7 % of the variance. The t statistic is 3.555 with 68 degrees of freedom and the two-sided p-value is 0.0007. The equivalent Cohen’s d is 0.862.

Why it matters

The calculator helps students, researchers and test developers who relate a binary variable, such as pass or fail, treated or not, to a score, report the correlation with its significance, and convert between a correlation and a standardized mean difference.

Good to know

Use the standard deviation of the whole sample computed with the divisor N, not N − 1, in the formula; if you only have the usual sample standard deviation, multiply it by √((N − 1) / N). Group 1 is the group coded 1, and the sign of r follows the direction of the difference. The p-value assumes independent observations and similar variances in the groups. The size of the correlation is limited by unequal group sizes, so compare it with the largest value possible for your split when you interpret it.

The correlation significance calculator tests a correlation, the Cohen d calculator measures a standardized difference, the effect size converter links the indices and the item difficulty and discrimination calculator uses this correlation for test items.

Frequently asked questions

What is the point-biserial correlation?

The point-biserial correlation is the Pearson correlation between a continuous variable and a truly dichotomous one, coded 0 and 1. It is r = (M1 − M0) / s × √(n1 × n0) / N, where M1 and M0 are the mean scores of the two groups, s the standard deviation of all the scores with divisor N, and N the total number.

When is it used?

It is used when one variable has two real categories, such as pass or fail, male or female, treatment or control, and the other is a score. In test analysis it gives the item-total correlation of an item scored right or wrong. It is equivalent to the t test for two groups, with a different effect size measure.

How is the p-value found?

The t statistic is t = r × √(N − 2) / √(1 − r²), with N − 2 degrees of freedom. The two-sided p-value is the probability, under no relation, of a t at least as large in absolute value. This is the same t as the pooled-variance t test of the difference between the two means.

What are its limits?

The maximum value depends on the split of the groups: the correlation cannot reach 1 when the groups are unequal. If the dichotomy comes from cutting a continuous variable, such as a pass mark, the biserial correlation is more suitable. The scores should be roughly normal in each group for the p-value to hold.

Was this calculator helpful?

Thanks for your feedback!

Results are estimates provided for information purposes. They do not replace professional advice.