In statistics, correlation (also called dependence) is a statistical relationship between two random variables or sets of data. Correlation indicates the extent to which these variables change together. While it is often used to suggest potential cause-and-effect relationships, it's important to remember the adage: "Correlation does not imply causation."

Types of Correlation

There are three primary types of correlation:

  • Positive correlation: Both variables change in the same direction. As one variable increases, the other also increases (and vice-versa).
  • Negative correlation: Variables move in opposite directions. As one variable increases, the other decreases (and vice-versa).
  • Zero/No correlation: There is no discernible pattern or relationship between the variables. Changes in one are not consistently associated with changes in the other.

Correlation Coefficient

The degree of correlation is represented by the correlation coefficient (often symbolized by 'r'). Values of 'r' range from -1 to +1:

  • +1: Perfect positive correlation.
  • -1: Perfect negative correlation.
  • 0: No linear correlation.

Methods of Calculation

Several methods exist for calculating the correlation coefficient, the most common being:

  • Pearson correlation coefficient: Measures the strength and direction of a linear relationship between two continuous variables.
  • Spearman's rank correlation coefficient: Used for ordinal data, assessing the strength of a monotonic relationship (one in which the variables may not change at the same rate but consistently trend in the same direction).
  • Kendall's rank correlation coefficient: Another measure for ordinal data, often used as a more computationally friendly alternative to Spearman's.


Understanding correlation has wide-ranging applications:

  • Science: Identifying relationships between variables to design experiments, make predictions, and better understand natural phenomena.
  • Finance: Assessing relationships between assets, predicting portfolio risk, and making investment decisions.
  • Social Sciences: Studying trends, understanding social behavior, and testing hypotheses about human relationships.
  • Data Science and Machine Learning: Feature selection, dimensionality reduction, and developing predictive models.

Important Considerations

  • Correlation vs. Causation: Correlation indicates an association, but it doesn't guarantee cause-and-effect. Other factors, such as lurking variables or coincidence, might be responsible.
  • Outliers: Extreme values can significantly impact the correlation coefficient. It's important to screen data for outliers before analysis.
  • Nonlinear Relationships: Correlation coefficients primarily measure the strength of linear relationships. Complex, nonlinear relationships may not be adequately represented by 'r'.


A study finds a positive correlation between ice cream sales and the number of drownings during a summer. This does not imply that eating ice cream causes drownings. A more likely explanation is that both activities increase during hot weather.