Where two variables possess a correlation but the link between them is not direct but through another variable, often time or size. For example, the recorded number of burglaries and the number of teachers in towns in the UK show a strong positive correlation, but this is because both of them have a positive correlation with the size of the town.