DataTerminology.comNEW
A statistical relationship measuring the strength and direction of association between two variables.
DESCRIPTION
Correlation quantifies how two variables move together, often via Pearson’s coefficient for linear associations or Spearman’s for monotonic relationships. Values range from −1 to 1, indicating negative to positive association, with 0 implying no linear correlation. Correlation does not imply causation, and confounding variables can produce spurious correlations. In modeling, high correlations can cause multicollinearity, affecting estimates. Profiling correlations helps feature selection and data understanding.
ALTERNATIVE DEFINITIONS
Wikipedia
In statistics, correlation or dependence is any statistical relationship, whether causal or not, between two random variables or bivariate data.
Frames correlation in the broader context of statistical dependence.
NIST Engineering Statistics Handbook
Correlation coefficients measure the strength of association between two variables and the direction of the relationship.
Defines correlation specifically via coefficients, emphasizing strength and direction.
Penn State Eberly College of Science
Correlation is a measure of the linear relationship between two quantitative variables.
Highlights linear correlation typically quantified by Pearson’s r.
RELATED TERMS
Pearson Correlation Spearman Correlation Covariance Dependence Multicollinearity Autocorrelation Partial Correlation
COMPETING TERMS
RELATIONS
KNOWLEDGE AREAS
Mathematics Data Science Data Quality
TARGET AUDIENCE
Data Scientist Data Analyst Information Analyst BI Specialist Data Steward
EXTERNAL SOURCES
- Wikipedia - Correlation and dependence
- NIST Engineering Statistics Handbook - Authoritative overview of correlation
- SciPy - Pearson correlation function docs
- Penn State - Lesson on correlation and regression
TOP QUESTIONS
Does Correlation imply causation?
No. Correlation measures association; causal inference requires additional assumptions, experiments, or identification strategies.
Category: Misunderstanding
What is Pearson vs. Spearman Correlation?
Pearson measures linear associations; Spearman measures monotonic relationships using ranks, robust to outliers and nonlinearity.
Category: Explanation
How does multicollinearity involve Correlation?
High predictor correlation inflates variance of estimates, destabilizing regression coefficients and inference.
Category: Risks
When is zero Correlation misleading?
Nonlinear relationships can have zero Pearson correlation yet strong dependence; inspect plots or use rank-based or mutual information.
Category: Mistakes
How to test for significant Correlation?
Use hypothesis tests on correlation coefficients or permutation tests; report confidence intervals alongside p-values.
Category: Testing
How to visualize Correlation effectively?
Scatterplots, correlation matrices with heatmaps, and pair plots reveal patterns, outliers, and nonlinearity.
Category: Best-Practices
How to handle highly correlated features?
Use regularization, remove redundant features, or apply dimensionality reduction like PCA.
Category: Guidance
What is autocorrelation?
Correlation of a variable with itself across time lags; crucial in time series analysis.
Category: Explanation
Can Correlation be used for data quality checks?
Yes, unexpected shifts in correlation structures can indicate data issues or pipeline problems.
Category: Application