Introduction
Correlation analysis is the group of methods used to measure how strongly two variables are associated, and in which direction they change together. It is widely used in social medicine and epidemiology to describe relationships such as income and life expectancy, or body mass index and blood pressure, before any causal conclusion is considered. This note covers the correlation coefficient, its Pearson and Spearman forms, the coefficient of determination, significance testing and the most common pitfalls in interpretation.
Basic idea of association
Two variables are correlated when knowing the value of one gives information about the value of the other. The association may be:
- Positive: both variables tend to increase together.
- Negative: one variable tends to increase as the other decreases.
- None: no consistent pattern, with a coefficient close to zero.
Correlation measures association, not cause. It also captures only the direction and strength of the relationship, so a strong curved (non-linear) relationship can still produce a coefficient close to zero.