When I saw this maths infographic over at SEOmoz I knew i had to share it! Infographics are a great way to get information across in a simple, digestible form, and this one has an important lesson for anyone involved in statistical analysis. Correlation does not imply causation.
There is always room for anxiety in analytics and A/B or MVT testing about what is causing an effect you are seeing in your stats. You will often see relationships between different variables. You’ll probably be surprised at how many weird coincidences you come across in your stats! In many cases, there is a relationship between two or more trends, but it’s a common mistake to infer that one thing is causing another. Our human minds are hard-wired to spot patterns, and it helps us make sense of the world. Mathematically speaking, it’s a mistake to imply causation from correlation. There may in fact be a 3rd or more factors affecting all the variables you are tracking.
This infographic explains the difference simply and effectively.
That’s it! I’ll be sure
to share any more appropriate infographics I come across, but if you haven’t already, you should really check out SEOMoz. It’s got a lot of great posts around analytics topics as well as being a definitive source for SEO.