Correlation Is Not Causality, Again and Again

CRAIN'S COMMENTS

This gets so aggravating. We have a new article in Medscape based on a statistical analysis. The premise of the study report and article is that marijuana use among teens increases the frequency of depression and suicidal thoughts. The basis for this assertion is a statistical correlation between use of marijuana and a teen having depression and reporting suicidal thoughts.

“What’s wrong with that?” you may ask.

What about the notion that people who are depressed may seek relief by using marijuana? The statistical correlation would be the same, but causality reversed. With that hypothesis, marijuana use becomes a symptom of a problem, not a cause. If true, parents need to connect with their children and understand their feelings, rather than prattle about drug use.

Without getting into the math, a simple view of correlation is as follows:

  • If X always and only happens when Y happens, they have a…

View original post 449 more words