What "Corr" Means in Statistics: Pearson Correlation Coefficient
"Corr" is the standard abbreviation for Correlation — specifically the Pearson Correlation Coefficient (r).
Definition
A statistical measure (−1 ≤ r ≤ +1) quantifying the strength and direction of a linear relationship between two quantitative variables.
Interpretation
- r = +1: Perfect positive linear correlation (X↑ → Y↑).
- r = −1: Perfect negative linear correlation (X↑ → Y↓).
- r = 0: No linear correlation.
- |r| > 0.7: Strong; 0.3–0.7: Moderate; < 0.3: Weak.
Critical Caveats
- Correlation ≠ Causation. Third variables, coincidence, or reverse causality may explain it.
- Linear Only. Misses curved (quadratic, exponential) relationships.
- Outliers Sensitive. One extreme point can inflate/deflate r dramatically.
- Assumptions: Continuous variables, bivariate normality, homoscedasticity.
Common Notations Across Platforms
- CORR(X,Y) — SQL / SAS / SPSS function.
- np.corrcoef() — NumPy (Python).
- cor() — R / MATLAB.
- =CORREL() — Excel / Sheets.