Point-Biserial r

Point-Biserial r — measure how strongly a yes/no group membership relates to a number. Type your two groups of values and instantly get the point-biserial correlation r, a significance test, and a clear two-group strip plot that shows the gap between the group means. It is the right correlation when one variable is binary (pass/fail, treated/untreated) and the other is continuous. Fully offline.

Does passing or failing relate to study hours? Does the treatment group score higher than the control? When one side is a two-category label and the other is a measured number, ordinary Pearson correlation still applies — it gets a special name, point-biserial r, and a neat interpretation: it is the standardized gap between the two group means, scaled by how the groups split.

No sign-up. No internet needed. No data collected. Just clear, instant statistics.

WHAT YOU CAN DO

• Data — name the groups and the value, then type each group's numbers into a clean editor (add or remove values with a tap). Everything recomputes as you type.

• Plot — see a two-group strip plot: every value is a dot, each group's mean is a thick tick, and the line joining the two means shows the gap that r measures. A steeper join means a stronger link.

• Concept — learn what r_pb measures, how big is "big" (effect size), how it becomes a t-test, and the lopsided-split warning — with offline, beautifully typeset formulas that never overlap.

• Simulate — set the mean difference, the within-group spread and each group's size with + / - steppered sliders, then watch how unequal group sizes shrink r even when Cohen's d stays put.

• Examples — load real group/number links (pass/fail vs hours, treatment vs outcome and more) with one tap.

WHAT YOU GET

• Point-biserial correlation r and its sign.
• A t-test with degrees of freedom and a p-value.
• Each group's mean and the standardized gap.
• A two-group strip plot with means joined.

WHY YOU'LL LIKE IT

• The right tool for binary vs continuous — not a workaround.
• See the gap — the strip plot makes the link concrete.
• Honest about splits — it shows how unequal group sizes pull r down (so it also reports Cohen's d).
• Fully offline — every statistic, plot and formula works with no connection.
• Hands-on — type exact data, or build groups with sliders and + / - steppers.
• Clean "Ink, Emerald & Coral" design, easy on the eyes.

A note on interpretation: because r_pb depends on how the cases split between the two groups, a very lopsided split (say 4 vs 25) shrinks r even when the real difference is the same — so the app shows Cohen's d alongside it. Whether you are a student learning effect sizes, a researcher comparing two groups, or anyone linking a label to a measurement, Point-Biserial r gives you the number, the test and the picture in seconds.

Download it, enter your two groups, and see how strong the link is.