B4RN: How To Lie With Statistics

Darrell Huff's excellent book How To Lie With Statistics was written over fifty years ago and yet is as applicable today as it was then.

I was reminded of the book when I was surfing the The B4RN project's web site where they have illustrated their progress using a barometer-style graphic.  It is a little deceptive as is it appears to have a non-linear scale, but having being alerted to this kind of skullduggery by Huff, I knew I had to delve deeper.

The barometer is currently showing an achievement of 368 against a target of 662 (about 56%) however the liquid in the thermometer is about 63% of the way up, exaggerating the success of the project by 46 sign-ups, if the scale were linear.

The lesson from Huff's book is that statistics are rarely presented under oath, and that an easy way to mislead the reader is to subvert our natural expectation that graphics depict objects that follow real-world physics - we naturally expect a thermometer to have a linear scale so we trust the height of the 'liquid' without looking at the actual numbers.

More on B4RN