Pearson correlation

Pearson Correlation Coefficient

The standard linear correlation coefficient — measures the strength of the linear relationship between two continuous variables on a scale from -1 to +1.

When you'd see it: Exploratory data analysis, correlation matrices, feature selection in ML, and any quantitative analysis of whether two continuous variables move together. Named after Karl Pearson; often referred to simply as 'the correlation coefficient' in non-technical contexts.

Why it matters: Pearson captures linear relationships well — it assumes the relationship is approximately straight-line. When the relationship is monotonic but non-linear, or when outliers distort the relationship, Pearson can be misleading. Knowing when to use Pearson vs Spearman is a sign of statistical maturity.

Common mistakes: Using Pearson on ordinal data (survey scales, star ratings) or non-normally distributed data. Pearson assumes continuous, normally distributed variables. Spearman correlation is the appropriate alternative for ordinal data or when the relationship is monotonic but not linear.

Related Spearman correlation, correlation, covariance, linear regression