Correlation and Causation are not the same thing. This may be apparent to many as I say it, but in my experience it is the most common problem that I face when I deal with political or sociological data, research, or if someone is simply trying to make a point.
To highlight the problem for those of you who may be unfamiliar with it, here are several examples:
-> There is a correlation between girls watching soap operas, and developing eating disorders. But soap operas do not give or cause girls to have eating disorders.
-> There is strong statistical evidence that there is a correlation with smoking and lung cancer. But smoking does not cause lung cancer because not everyone who smokes gets lung cancer.
-> Sometimes, this can become foggy when correlations can be made, and the cause is not apparent. The rise in vaccinations correlated with the rise in autism. This led many to believe that vaccinations caused autism. However, studies show that it is most likely an increase in the diagnosis and awareness of autism and the decrease of the diagnoses of mental retardation (substituting the name with autism instead) that had caused for the rise in autism statistics.
Things can easily become very muddled, and we have to be careful while comparing data sets to make sure that we know the difference between correlation and causation.
---Rebecca
Awesome point Rebecca. I think this concept often gets lost within the average population who do not understand scientific studies. Simply because there is a correlation between red wine intake and lowered risk of heart disease does not mean that drinking red wine directly lowers one's chance of heart disease. Again, great point.
ReplyDeleteI like your use of the autism example to explain the difference between causation and correlation. We recently discussed the misconception that vaccines cause autism in my nutrition class. I had always heard about this controversy but had never been aware that the doctor who conducted the research showing a correlation between vaccination and autism actually fabricated his results. The results were disproven and the doctor credited with the study lost his medical license. I think this highlights the importance of not only understanding the difference between causation and correlation, but understanding the difference between valid and invalid research findings.
ReplyDeleteRebecca,
ReplyDeleteI don't think you could have answered this question any better!Personally, I like your example about the eating disorders. I see people make the mistake of using causation rather than coorelation a lot in my psychology classes. Overall, I thought you made a great point throughout your answer and your examples were perfect for someone that did not understand the difference before.
This was a very good post about the difference in correlation and causation. Many times people see the two as the same, but your examples perfectly show the differences in the two. I like how you show that external factors can play a part in the relationship between two events with the autism example. Numbers show if there is a relationship, but further research is required to see if there is any real causation.
ReplyDeleteFor the third major unit in this course we will look at the vaccine/autism controversy in depth, including reading the original research papers involved. Most of statistics is not actually concerned with causality, but with determining correlation in the first place.
ReplyDeleteRebecca, your smoking example is interesting to me because I would use a similar example to say that there are different definitions of "cause" in different contexts. Of course, we will talk more about causality later this semester. In the meantime, we know that correlation does not guarantee causation, but what can correlation tell us?