Pearson Correlation Coefficient (r) Between Two Variables
Last reviewed: April 2026
A correlation calculator computes the Pearson correlation coefficient (r) between two data sets, measuring the strength and direction of their linear relationship. Values range from -1 (perfect negative correlation) to +1 (perfect positive correlation), with 0 indicating no linear relationship.
The Pearson correlation coefficient (r) measures the strength and direction of the linear relationship between two variables. It ranges from −1 (perfect negative correlation) to +1 (perfect positive correlation), with 0 indicating no linear relationship. This calculator computes r, r², and the t-statistic for significance testing. Understanding correlation is fundamental to data analysis, research, and decision-making. For broader statistical analysis, use our Statistics Calculator.
|r| ≥ 0.9 is very strong, 0.7–0.9 is strong, 0.5–0.7 is moderate, 0.3–0.5 is weak, and below 0.3 is very weak or negligible. The coefficient of determination (R²) tells you what percentage of the variation in Y is explained by X. An r of 0.8 means R² = 0.64, so 64% of the variability in Y is accounted for by the linear relationship with X.
This is the most important caveat in statistics. Two variables can be strongly correlated without one causing the other. Ice cream sales and drowning deaths are positively correlated — not because ice cream causes drowning, but because both increase with hot weather (a confounding variable). Always consider confounders, reverse causation, and coincidence before drawing causal conclusions from correlation data.
| r Value | Strength | Direction | Example |
|---|---|---|---|
| 0.90 to 1.00 | Very strong | Positive | Height vs arm span |
| 0.70 to 0.89 | Strong | Positive | Study hours vs grades |
| 0.40 to 0.69 | Moderate | Positive | Income vs happiness |
| 0.10 to 0.39 | Weak | Positive | Shoe size vs IQ |
| 0.00 | None | — | Birth month vs height |
| -0.70 to -1.00 | Strong | Negative | Exercise vs resting HR |
The Pearson correlation coefficient (r) measures the strength and direction of the linear relationship between two numerical variables on a scale from −1 to +1. A value of +1 means a perfect positive linear relationship — as one variable increases, the other increases proportionally. A value of −1 means a perfect negative linear relationship — one increases exactly as the other decreases. A value near 0 indicates no linear relationship, though nonlinear relationships may still exist. In practice, perfect correlations of ±1 are extremely rare outside of mathematical constructions; real-world correlations of ±0.7 or above are considered strong.
| r Value | Strength | Interpretation |
|---|---|---|
| 0.90 to 1.00 | Very strong positive | Variables move together very closely |
| 0.70 to 0.89 | Strong positive | Clear positive trend with some scatter |
| 0.40 to 0.69 | Moderate positive | Noticeable trend but significant scatter |
| 0.10 to 0.39 | Weak positive | Slight upward tendency |
| −0.10 to 0.10 | Negligible | Essentially no linear relationship |
| −0.39 to −0.10 | Weak negative | Slight downward tendency |
| −0.69 to −0.40 | Moderate negative | Noticeable inverse trend |
| −1.00 to −0.70 | Strong negative | Clear inverse relationship |
The Pearson correlation formula calculates the covariance of two variables divided by the product of their standard deviations: r = Σ[(xᵢ − x̄)(yᵢ − ȳ)] / √[Σ(xᵢ − x̄)² × Σ(yᵢ − ȳ)²]. The numerator captures how x and y co-vary — when both are above their means simultaneously, the products are positive, contributing to a positive r. When one is above its mean while the other is below, the products are negative, pulling r toward −1. The denominator standardizes the result to the −1 to +1 scale regardless of the original units, making correlations between inches and pounds directly comparable to correlations between dollars and years.
This principle is the most important caveat in statistical analysis. A strong correlation between two variables means they move together, but it does not establish that one causes the other. Ice cream sales and drowning deaths are positively correlated — not because ice cream causes drowning, but because both increase with temperature (the confounding variable). Spurious correlations abound: the divorce rate in Maine correlates with per-capita margarine consumption; Nicolas Cage films correlate with swimming pool drownings. These examples illustrate that with enough variables and enough data, coincidental patterns emerge with impressive-looking correlation coefficients.
Establishing causation requires controlled experiments, natural experiments, or careful application of causal inference methods like instrumental variables, difference-in-differences, or randomized controlled trials. Correlation is often the first step in identifying relationships worth investigating further, but it is never sufficient evidence for causal claims on its own.
R-squared (r²) is the square of the correlation coefficient and has a particularly intuitive interpretation: it represents the proportion of variance in one variable that is explained by the other. A correlation of r = 0.8 gives r² = 0.64, meaning 64% of the variation in y is explained by the linear relationship with x. The remaining 36% is due to other factors, measurement error, or random variation. This distinction matters — an r of 0.5 sounds moderate, but r² = 0.25 means x explains only 25% of the variation in y, which may be insufficient for reliable prediction.
Spearman's rank correlation (rₛ) measures the monotonic relationship between two variables — whether they tend to increase together, even if not linearly. It works by ranking both variables and then computing the Pearson correlation on the ranks. This makes it robust to outliers and applicable to ordinal data (rankings, Likert scales) and nonlinear-but-monotonic relationships. If income and education follow a curved relationship where each additional year of education adds progressively more income, Pearson's r might understate the relationship while Spearman's rₛ captures the consistent trend. Use Spearman when your data is ranked, contains outliers, or when the relationship is monotonic but not strictly linear.
Outliers can dramatically inflate or deflate correlation coefficients. A single extreme data point can shift r from near-zero to 0.8 or from 0.9 to 0.3, depending on its position. Always visualize your data with a scatter plot before computing correlation — the coefficient alone can be misleading. Restriction of range is another common pitfall: if you measure the correlation between SAT scores and college GPA only among students at a highly selective university, the correlation will appear weak because the range of SAT scores is narrow. In the full population, the correlation would be substantially stronger.
Simpson's paradox occurs when a trend that appears in several groups of data reverses when the groups are combined. A treatment might show positive correlation with recovery in men and in women separately, but negative correlation overall if the groups have different baseline rates. Aggregating data without considering subgroups can produce correlations that misrepresent the underlying relationships entirely.
Correlation analysis is foundational across many fields. In finance, correlations between asset returns drive portfolio diversification — combining assets with low or negative correlations reduces overall portfolio risk. In medicine, correlations between risk factors and disease outcomes guide screening recommendations and public health interventions. In psychology, correlation matrices reveal relationships between personality traits, cognitive abilities, and behavioral measures. In quality control, correlations between process variables and product defects identify which parameters to monitor and adjust. This calculator handles both Pearson and Spearman correlations, giving you the right tool for both linear and monotonic relationship analysis.
In large datasets, even very weak correlations can be statistically significant — meaning they are unlikely to have occurred by chance — while still being practically meaningless. A study of 10,000 people might find a correlation of r = 0.05 between shoe size and IQ that is statistically significant at p < 0.01, but r² = 0.0025 means shoe size explains only 0.25% of IQ variation — utterly useless for any practical purpose. Statistical significance tells you whether a relationship exists; effect size (the magnitude of r) tells you whether it matters. In research and decision-making, always report both the correlation coefficient and its r² value, and consider whether the effect size is large enough to be meaningful in the specific context.
Enter your paired data points (X,Y values) to instantly compute the Pearson correlation coefficient, R-squared value, regression equation, and statistical significance. The calculator handles datasets of any size and provides both the correlation strength and a p-value indicating whether the observed correlation is statistically significant or could have occurred by chance. Always pair this quantitative analysis with a visual scatter plot to check for nonlinear patterns, outliers, or subgroup effects that the correlation coefficient alone cannot reveal.
For best results, use datasets with at least 20 data points to ensure meaningful statistical power and reliable correlation estimates that support sound decision-making in research and business contexts.
See also: Statistics · Combinations · Permutations · Logarithm · Integral
→ Correlation does not imply causation. Ice cream sales and drowning deaths are highly correlated — both increase in summer. The heat causes both; ice cream doesn't cause drowning. Always look for confounding variables.
→ R-squared is more intuitive than r. R² tells you the percentage of variation in Y explained by X. An r of 0.7 means R² = 0.49, so X explains only 49% of Y's variation. The other 51% comes from other factors.
→ Outliers heavily influence correlation. A single extreme data point can dramatically change r. Always plot your data first — a scatter plot reveals patterns that a single number cannot. Remove or investigate obvious outliers.
→ Use for data-driven decisions. Marketing spend vs revenue, study hours vs grades, exercise minutes vs weight change — correlation quantifies relationships. Combine with our Statistics Calculator for complete analysis.
See also: Statistics Calculator · Standard Deviation · Average Calculator · Z-Score Calculator