Correlation vs. Causation: What's the Difference?

Correlation does not mean causation. This statement means that the fact that two variables correlate with each other is not enough to assert that there is a causal relationship between them. Correlation between two independent variables or events can be as a result of a third effect called a confounder ( Altman & Krzywinski, 2015). For example, the correlation between smoking and lung cancer cannot be said that there is a causal relationship between the two. In short, one cannot conclusively say the smoking results in incidences of lung cancer. Other variables can affect this relationship such as diet, alcohol intake, and environmental conditions which together form the confounding effect. Therefore, to say that correlation means causation is a fallacy that has nothing to prove as far as a cause-and-effect relationship is concerned. 

However, there are few circumstances where correlation could be said to be a proof of causation. In an example where an experiment involving two twins who have been getting consistent grades in their exams is carried out, causation can be said to result from correlation. In the experiment, one of the twins would be taken to study for some hours while the other sent to a park to relax. When they are given a test and their grades diverged by a huge margin, then it can be concluded that there is a causal effect on their exam scores. Therefore it can confidently be said that the correlation exhibited between studying and the exam scores can mean that there was a causation effect. Three important steps can, therefore, be used to prove if truly correlation means causation. The first is to use the variables that can be used to quantify a relationship. Secondly, one should be in a relationship to explain the relationship, and thirdly, it should be able to be explained probably using the Hill's criteria. 

Reference  

Altman, N., & Krzywinski, M. (2015). Points of Significance: Association, correlation, and causation. Nature methods , 12 (10), 899-900.