Book Review: How to lie with Statistics by Darrell Huff

Sitting in a chair with a mug of coffee, like everyday, I was reading the editorial section of the newspaper, Prothom Alo. I, abruptly, wondered an article named “Mr. Finance Minister, balance the sum please” written by Dr. Mainul Islam, former UGC Professor of Economics, University of Chittagong, Bangladesh. It was a write-up describing ongoing controversy about GDP growth(5.28)—which is provided by Bangladesh Bueuro of Statistics(BBS). The controversy began when IMF and World Bank provided a Growth rate of 2.28 and 1.2 respectively. In accusing the BBS Dr. Mainul quoted from  Mark Twain: 

There are three kinds of lies- lies, damned lies and Statistics. 

Now you may wonder why Mark Twain regarded Statistics as more than damned lies. Here the book How to lie with Statistics can shed light on the argument made by Twain. Besides, what Darrell Huff did is to write the book to enlighten your mind in understanding the chicanery of Statistics. Let's come to the main part of the book with an example.

Suppose, Yale University, in their advertising, published that average monthly income of their alumni is $3000. Now what does it mean to you? Does it mean whoever get admitted to Yale will earn a salary of $3000? Wait! Wait! If you think the way they force you to think, then you are gullible. Now you may think why I said so. Patience dear! Here are a lot of deceptions and one of them is what the author depicted as The built-in Sample Bias(One of the chapters of the book).

You and I are not sure about how they collected the sample, who got the questionnaire and who answered the questionnaire. In general, higher income earning people are easily reachable and the people who have lower income tend to hide his income,  implying that they are unreachable.  That means most of the people from the sample have higher income and they are not the exact representatives of the population. 

Another trickery is The well chosen average(another chapter of the book). The term average has a loose meaning as maintained by Darrell huff. It can mean arithmetic mean, median, mode. Now the question is what type of average Yale authority used here? If it is median(middle number of the sample), then 50% of the alumni have an income lower than $3000 and 50% have higher than that. And if It is mode(most frequent number of the sample), then most of the alumni earn $3000. See, Yale authority didn’t mention the type of average.     

The author also explain how, by cutting the graph from below and by one dimensional picture illustrated below, Statistician can delude you. These things are described in Chapter The Gee-whiz curve and The one-dimensional picture respectively.  

If it comes about writing style, then I would say It is as simple as water. Readers might not encounter any esoteric statistical term what could make their time unworthy.   

At the end of the book, the author “explain how to look a phony statistic in the eye and face it down; and no less important, how to recognize sound and usable data in that wilderness of fraud to which the previous chapters have been largely devoted”.

Before sum up, I would like to mention a quote from Aaron Levenstein, former business professor, Baruch College: 

Statistics are like a bikini. What they reveal is interesting. But what they hide is vital. 

Now If you think statistics as a bikini and are interested about what It conceals😜, you must read the book. 

Happy Reading… 🙂

Rating: 4.5/5