Skip to section navigation Skip to main content

Sneaking Statistics

Now, look at this graph…

This is a line graph. The graph is titled United States of America. The horizontal axis is labeled year, and extends from 1978 to 2003, counting in increments of one. The vertical axis is labeled Obesity prevalence, and extends from 0% to 35%, counting in increments of 5%. There is a red line labeled obesity that is graphed. It begins at (1978, 15%) and extends straight to (1999, 30%), decreases to (2001, 28%), and increases to (2003, 33%). Below the red line is a blue line labeled motor vehicles and relatively parallel to the red line. It begins at (1978, 9%) and has a label 148 million. It ends at (2003, 28%) and has a label 231 million.

Source: Man's Greatest Mistake opens in new window
Licensed under a Creative Commons Attribution 2.0 UK: England & Wales License.
opens in new window

Can you spot what is misleading?

This graph comes from a statistical study aimed at finding causes for obesity. The graph clearly shows that as the number of motor vehicle rises, so does obesity. These two curves have almost the exact same rate of change!

The author uses this graph to make the claim that a rise in vehicles is causing the rise in obesity. This is another common misuse of statistics. There is an association between the rise in obesity and the rise in motor vehicles. However, nothing about this data shows or proves that the one causes the other.

Remember, association does not imply causation. From the graph, there appears to be an association between the rise in motor vehicles and a rise in obesity. They appear associated because the slopes of the lines are almost identical. Deeper research and experimentation must be done once an association is noticed to prove that the rise in motor vehicles caused the rise in obesity.



< Previous Next >