Everyone is familiar with the phrase “there are three kinds of lies: lies, damned lies and statistics”, attributed to Mark Twain, among other authors. This phrase has been used to illustrate how statistics can be twisted and misused to prove or disprove certain facts. This phenomenon was first described in the 1954 classic “How to Lie with Statistics” by Darrell Huff.
In “How to Lie with Statistics”, Huff explains how even seemingly innocent statistics can be manipulated to mean different things. He begins by pointing out that people tend to do two things when presented with numerical evidence: simplify the information, and unconsciously accept it as proven. Huff then moves on to give examples of how leaving out certain details can change the meaning of numerical information.
One of the main points that Huff outlines is that numbers alone don't give you the full picture and can contribute to inaccurate conclusions if context is not taken into consideration. He goes on to discuss underestimation, overestimation and miscellaneous statistical distortions. Huff also examines the changing meanings of terms that can drastically alter numerical results and he brings up how fractions are often better indicators of proportion than percentages.
In the closing of his book, Huff calls for better literacy among the public in relation to statistics. He was one of the first authors to make the public aware of how numbers can be manipulated, and this book still serves as a good teaching tool. It has since been updated to reflect newer technologies, but the basic principles and the message remain the same: statistics can provide valuable information if we learn how to interpret them accurately.
Overall, “How to Lie with Statistics” is an eye-opening book, and a must-read for anyone who is interested in becoming a more informed consumer of data. Huff quickly and easily explains how even the most basic of calculations can be manipulated to suit the needs of someone presenting the data. Though the book is nearly 70 years old, its relevance and value remain strong, especially in the digital age, when data collection and manipulation has never been easier.
