Book review of How to Lie with Statistics by Darrell Huff

If you want to scare someone and shut them up, speak in numbers. That’s the kind of “stupor effect” numbers will have on many people. This largely impacts the way we make our decisions in buying a company’s stock or any product. The reason behind is that numbers always are associated with “facts” and hence, comes with more credibility. And, when this is backed by some known brand or research organization, we tend to blindly trust their claim and execute our decision. There is nothing wrong as the world largely works that way but Darrell Huff, through his book, “How to lie with statistics”, throws light on how this behaviour is largely exploited by people- journalists, ad copywriters, professionals, statisticians- who know how to wield these data points to make us fall prey to their chicaneries.

This book claims that this is a primer on statistics and true to its claim, the first few chapters brush the readers up on the basics including concepts on the averages – the difference between mean, median, mode and how each one of these could be used in different context to deceive a common man, importance of the sample size and its range values, and how graphs are employed by the crooks in sending out an incorrect message to the recipient. So, it’s all about basics and if you are shopping for advanced concepts, you will find little material in this book.

The author, in the subsequent chapters, delves a bit deeper by building on these foundational concepts by providing good number of examples on how percentages could be misleading when the sample size is intentionally kept smaller. For instance, if your sample size for a survey to establish a result is 3, even if one of them votes in your favour, it would mean that 33% of the sample population have corroborated the results. This seems ridiculous but the fact is that everyday we are encountering such “semi-attached” figures where the person publishing such figures twist the outcome to lure the readers and attain their desired objective. He cites the importance of understanding the gimmicks about causal relationships, which in plain reading makes a lot of sense. When you see someone buy more stocks and also witness his/her wealth increase, it is natural to come to a conclusion that one when buys more stock, one’s wealth will also increase. But this needn’t necessarily be the case. It could be vice versa also; where, with increase in wealth, one buys more stock. Or there could be no relation at all between these two variables. Darrell warns the reader to look through the correlations, to see if there are any unwarranted conclusions made by establishing positive causal connections between the trends, especially by the interested party to gain out of the conclusion drawn.

Similarly, he illustrates the effect on readers when decimal numbers are used. Consider the two phrases- “around 40% , “at 40.3%”; though, there is little significance with the decimal usage, the precision that comes with decimal figures makes us feel the latter is more credible than the former.

The book also has a dedicated chapter consolidating all the crooked ways in which statistical material could be employed to misinform people and manipulate them, which he calls “Statisculation”, and the word is a combination (a portmanteau) of Statistics + Manipulation. Darrell also emphasizes throughout the book that these distortions of statistical data are not always the work of professional statisticians but of the salesman, public relations experts, journalist or ad copywriters who twist, exaggerate, over-simplify the figures as a means to achieve their goal.

As we read through the book, you will notice that Darrell not only points out the fallacies but also prescribes tools and techniques to guard yourself off from falling victim to the trap. He strongly suggests that whenever one encounters statistical information, one has to “talk back to statistics” by raising the below questions:

1. Who says so- This will help the reader analyse whether the sample is a biased one due to potential association with the person who publish such results.

2. How does he know- The reader should scrutinize how the sample is selected and also figure out if some misleading correlation is establish to twist the results.

3. What’s missing-  As the adage goes, “Devil is in the details”, it is important that the reader should look for finer details such as indications on the nature of the average used, range of the sampling and look for disclosure of raw figure when percentages are used. To quote the author, “Sometimes what is missing is the factor that caused a change to occur. This omission leaves the implication that some other, more desired factor is responsible”

4. Did somebody change the subject- Darrell cautions the reader to look out for a switch somewhere between the raw figure and the conclusion. For instance, more reported cases of a disease are not always the same thing as more cases of the disease.

5. Does it make sense- Last but not the least, reader should exercise his/her common sense in digesting a statistical information, as Darrell claims many a statistics is false on its face. And, one should not suspend the diligent application of their common sense because of the precise use of figures or based on extrapolation of a trend, as most of them have an underlying premise that the projection holds good subject to everything else being equal, which rarely is the case in reality.

The book, at times, sounds more accusatory of the actions of the people-facing professionals but I personally felt it is the need of the hour, as the overload of information on an individual is increasing day by day, with random figures and conclusions floating all around. This book, also, highlights the fact that despite its mathematical base statistics is as much as an art as it is a science, thus, leaving room for many manipulations and distortions. Reading this book will, therefore, arm you with required skills so that you can take a sharp second look at the statistical material before accepting any of them and not arbitrarily accept/reject them. So, when someone says that vaccine X is more effective than vaccine Y, you now know what all aspects to look for before you accept the claim/news.

Apart from the principles and techniques on identifying statistical misinformation, I liked the cartoons in the book- thanks to the cartoonist, Irving Geis- which are satirical and makes you yearn for more. Surprising fact is that this book, written in 1954, almost seven decades ago, still holds good in the age of Artificial Intelligence and Data Science, emphasizing the fact that human minds haven’t changed much over time and less likely that it will in the future! Looks like even the famous American Writer Mark Twain understood the fallacies with facts, which is evident from the below statement from his book ‘Life on the Mississippi’ written in 1883, “There is something fascinating about science. One gets such wholesale returns of conjecture out of such a trifling investment of fact” (the whole paragraph from which this statement is taken from is cited in the book).