The history of "How to Lie with Smoking Statistics"
Alex Reinhart
– October 4, 2014
refsmmat.com
Darrell Huff is best known as the author of How to Lie with Statistics, which was published in 1954 and has been the most popular statistics book ever since. It was a short, snappy book illustrated by cartoonist Irving Geis and containing numerous humorous examples and anecdotes, priming its readers to resist the misleading plots and figures used by marketers and politicians.
Less well-known is Huff’s other statistical book, How to Lie with Smoking Statistics. It was commissioned by the tobacco industry in the 1960s after the Surgeon General’s famous report that smoking causes lung cancer, but never reached print.
The full story of Huff’s book is now available in my article “Huff and puff” in Significance magazine (preprint PDF here). Here I’ll collate some links to source documents available in the Legacy Tobacco Documents Library.
The manuscript
The manuscript of Huff’s book is available only in fragments, but I believe I have located every chapter. Some chapters appear in multiple undated copies in the Library, so it is not clear which represents the final revision; I usually picked the copy with the highest-quality scan, rather than fussing over establishing the right precedence.
- Chapter 1: How a Sample May Be Biased
- Chapter 2: Sample Size and Significance, along with chapters 4 and 12
- Chapter 3: Epidemiological Studies vs. Controlled Experiments (numbered as 4 in the PDF, but renumbered according to this letter)
- Chapter 4: Averages and Ranges, along with chapters 2 and 11
- Chapter 5: Reporting and Classifying Data
- Chapter 6: The Omitted Figure, along with chapter 7
- Chapter 7: Overprecise and Unknowable Figures, along with chapter 6
- Chapter 8: Semi-Attached Figures, along with chapter 7
- Chapter 9: Graphic Techniques, along with chapters 6 and 11
- Chapter 10: From Figures into Words, also renumbered
- Chapter 11: Correlation and Causation, along with chapters 6 and 9, also renumbered
- Chapter 12: Genotypes, along with chapters 2 and 4
I believe this is the entire manuscript, apart from perhaps a preface and conclusion.
Correspondence
- Letter from Huff to Edwin Jacob, December 1964, discussing Huff’s plans.
- Outline of proposed book, December 1965.
- Memorandum from Edwin Jacob, January 1967, proposing paying Huff $1,500 more, for a total of $9,000.
- Letter from H.T. Austern to William Kloepfer, April 1968, discussing the legal issues in promoting Huff’s work (and other marketing efforts).
- Letter from William Kloepfer (Tobacco Institute) to Earle Clements, April 1968, suggesting caution in their advertising strategies after the True fiasco.
- Letter from H.T. Austern to the Tobacco Institute, April 1968, suggesting disclaimers to be attached to Institute marketing materials (including Huff’s book) and reviews to prevent further embarrassments.
- Review by K.A. Brownlee, in May 1968, of Huff’s manuscript. Doesn’t notice several glaring errors.
- Letter from John Lyden (Tiderock) to Macmillan, July 1968, noting that Huff has connections to the industry which should be acknowledged in any book.
- Letter from Seymour Henck (Tiderock) to William Kloepfer (Tobacco Institute), July 1968, complaining that Huff’s book is a “mass of verbiage” which “needs drastic editing before it will directly address itself to the needs of our industry.” (Misidentified by Andrew Gelman as being from Kloepfer.)
- Letter from Seymour Henck (Tiderock) to Huff, November 1968, saying that Macmillan would soon mail a revised contract but that Tiderock was no longer working with the Tobacco Institute.