Scatter plot

Paste X and Y in pairs. Pearson r measures linear association, not causation. The line is ordinary least squares from the regression sheet's formula. A curved cloud can have a small r and still be a real relationship.

The example is six pairs close to a slope of 3. Read r with the picture. A high r with one point far from the rest is not a stable relationship.

Put the variable you control or suspect on X and the result on Y. Keep the pairs in the same row: shifting one column by a line pairs every reading with the wrong partner and the picture still looks plausible.

Check for groups before you trust r. Two machines, two shifts, or two lots can each show no trend and still line up into a strong one when plotted together. Plot each group alone if you have them. With fewer than about ten pairs, a single point can move r by 0.3 or more.