Correlation is used to identify both the degree and direction of a relationship, a scaled version of covariance which makes no temporal restrictions (Field and Hole, 2002). For this reason quasi-experiments can be conducted in a natural setting still making it clear that intervening variables may be acting on any results.