Measure or calculation HL only
Correlation
Where you meet this term: IB Diploma
The extent to which two variables are linearly related: with positive correlation they tend to rise together, with negative correlation one tends to fall as the other rises, and with no correlation there is no straight-line pattern.
In context
Correlation is described by its direction (positive or negative) and its strength (strong when the points lie close to a line, weak when they are widely scattered), and can be measured by a correlation coefficient between −1 and +1. Correlation does not prove causation: two variables may move together by coincidence or because a third factor drives both. A strong correlation makes forecasts from a line of best fit more reliable, but never certain.
Example
Sales of umbrellas show a strong positive correlation with rainfall, while sales of ice cream show a negative correlation with it.
Frameworks, tools and diagrams
- Simple linear regression IB toolkit
Related terms
Suggested next
- Cyclical variation Also in 4.3
- Random variation Also in 4.3
- Seasonal variation Also in 4.3
- Trend Also in 4.3