Newsletter article – Correlation vs. causation
Correlation does not mean causation
In a recent Black Belt class we had great discussions concerning when two variables correlate with each other and how it does not necessarily mean one is causing the other. As the illustration above highlights a correlation between divorce rates in Maine with American’s consumption of margarine. If there was causationbetween them it would be logical to have a “margarine reduction policy”. In fact, the governor in Maine could lobby for the congress to pass laws restricting the sale of margarine in an effort to protect Maine families.
While numbers manipulation and false claims is common in politics, we must recognize the same actions may be happening in your organization. As Lean Six Sigma professionals we must ensure the claims we make represent the truth and are backed up with facts and data. Politicians can get away with making unsupported claims, we cannot. Data integrity is critical for the long term prosperity and growth of the companies/organizations we serve.
In these final days before the mid-term elections watch some of the television ads and see which ones are claiming causation by using correlation . If nothing else it will make the political commercials a little more bearable.
NOTE: Las Vegas Lean Six Sigma began teaching students in July of 2010, since that first class up to present day Nevada’s unemployment rate dropped from 13.5% to 4.5%. You’re welcome Nevada. 🙂
Erik