Correlation vs. Causation Error

Correlation vs. Causation Error

Describe the error in the conclusion. Given: There is a linear correlation between the number of cigarettes smoked and the pulse rate. As the number of cigarettes increases the pulse rate increases. Conclusion: Cigarettes cause the pulse rate to increase.

The error in conclusion in the above statement is drawing a conclusion based on an inappropriate statistical technique in analysis. For instance, correlation is a statistical technique that is used to ascertain the nature of a relationship between two variables. Linearly, it tells the degree of the strength of the relationship. Therefore, it does not explain or imply the causation relationship – in that the changes in the dependent variable are caused by changes in the independent variable (Pearl, 2010). In other words, correlation does not indicate that changes in Variable A are due to changes in Variable B. However, it is an indication of close causational relationship but does not prove the causation. Therefore, any conclusion made that a correlation translates to causation is termed as a cause logic fallacy, indicating a false cause, hence may lead to an error as in the above case.  

Therefore, in the above example, it is erroneous to conclude that the number of cigars smoked explains the increase in pulse rate. As much as these two variables may be related theoretically it does not that the cigarette smoking causes the pulse rate to increase.  Actually, there may be the issue of third factor C influencing the pulse rate more than the cigars smoked, or that the correlation is just by coincidence (Pearl, 2010).  Possibly, it might appear like the person has health factors which are affecting his or her pulse rate more than the cigars smoked which in the real sense affect the pulse rate. From such a scenario, for such a person, smoking cigars will indicate a correlation but in the real sense, it does not cause the change in pulse rate. Therefore making the conclusion to be flawed – a more powerful technique such as regression is appropriate to assess causal relationship. 

References

Pearl, J. (2010). Causality: Models, Reasoning, and Inference. Cambridge University Press.