Correlation not causation

With the amount of data available today there is a great deal of research into trends. One example is looking at illnesses in an area by the amount of searches done, over a specific time period for specific terms. If a lot of people in a specific area are all of a sudden searching for flu symptoms it probably means people have the flu (or it could mean there is a big exam coming up and students are trying to figure out how to get out of the exam!). Another example is looking at the number of reservation cancellations at restaurant in a specific area as a sign that something is wrong in that area.

The problem with this type of research is sometimes there is a correlation but it’s not related to the cause. To use an example which was used in a recent article on, during the flu season there are more online searches for high school basketball schedules than other times of the year. This doesn’t mean that high school basketball (or their schedules) causes the flu but rather flu season happens to occur during high school basketball season.

Make sure you don’t react to correlation if there isn’t also causation.

Have a great day!


Leave a Reply

Your email address will not be published. Required fields are marked *