Shared movement, counted
For every dot, the calculation asks whether x and y sit on the same side of their averages. Same side adds positive evidence. Opposite sides add negative evidence.
Math turns a cloud of paired observations into one comparable score. Change the data below and watch exactly what that score notices.
For every dot, the calculation asks whether x and y sit on the same side of their averages. Same side adds positive evidence. Opposite sides add negative evidence.
A result near zero says a straight line is a poor summary. A strong curve can still be hiding in plain sight.
The sign tells you whether the variables usually rise together or move in opposite directions.
The distance from zero tells you how tightly the dots follow a straight-line pattern.
Scaling makes very different units comparable: inches and pounds can produce the same r as centimeters and kilograms.
Pearson's r is not magic. It is a repeatable compression of the picture you just made.
Find each average. Every point is now described by how far it sits above or below the middle.
Pair the deviations. Matching directions produce positive products; opposite directions produce negative ones.
Divide by the total spread. That keeps the answer in a shared language from −1 to +1.
Pearson's r listens for a straight-line relationship. A U-shape has a real pattern, but its rising and falling halves can cancel each other.
An extreme observation has leverage. Try the “One outlier” preset, then remove its final pair to see how much one dot can move the coefficient.
Two variables can travel together because of coincidence, a third variable, or reverse causality. Correlation describes association; study design supports causal claims.
Your export includes every x/y pair, the computed Pearson coefficient, its plain-language interpretation, and the calculation note behind it.