Proofiness: The Dark Arts of Mathematical Deception

Proofiness: The Dark Art of Mathematical Deception by Charles Seife examines the misuse of data and statistics in today's world. The book divides the concept of proofiness into a number of categories such as disestimation and fruit-packing. Although few of the examples will be surprising, the book

Show More

does a nice job providing the "big picture" of this disturbing topic.

Seife defines proofiness as "the art of using bogus mathematical arguments to prove something that you know in your heart is true - even when it's not" (p. 4). He notes that proofiness is powerful because people are blind to the fact that every measurement is uncertain and impure in some respect.

Although the book contains many references to the use of proofiness in politics, he picks on both liberal and conservative groups to make his points. He uses examples from well-known scholarly journals as well as as advertising campaigns to show that proofiness is everywhere.

I'd recommend this book to anyone who wants to review the "dark art of mathematical deception" or enjoys a good book focusing on issues related to the application of mathematics in the "real world". If you like this book, there are many others that explore deception found in other content areas such as science and map making.

Show Less

Very Americentric -- US education, US economics, US elections, US courts, US you-name-it. The book's title and library categorization of 510 / QA99 notwithstanding, it is "mathematical" only in that it explores people's innumeracy, illogicality, and numerical malfeasance. At the outset, Seife says

Show More

that "if you want to get people to believe something really, really stupid, just stick a number on it." (p 7) I think an equally significant book could have resulted if the words "a number" had been replaced by "religion," "patriotism," or "military honor."

Show Less

Charles Seife’s book “Proofiness” is a description of the many ways in which mathematical claims can be deceptive. For example, he discusses “disestimation” in which one provides more precision to a measurement than is warranted. (Remember the rules for significant digits?) He introduces

Show More

this notion with the amusing example of a museum tour guide who asserts that a dinosaur skeleton is “Sixty five million and thirty eight years old”—because when she started working at the museum thirty-eight years before, she was told that it was sixty five million years old! But Seife goes on to give a detailed account of the 2008 Minnesota Senate race between Norm Coleman and Al Franken, which was so close that any claim that either of the candidates had actually won was necessarily based on disestimation.

Seife discusses many other examples of mathematical errors that appear in arguments, sometimes inadvertently, as in the case of the befuddled museum tour guide, but often intentionally, in an attempt to be misleading. These include cherry-picking data, comparing apples and oranges, and ignoring systematic biases in polling and instead focusing on the formal margin of error. While readers with a solid education in mathematics will already know of the problematic types of arguments he presents, the illustrations of their use that he provides--from advertising, journalism, and politics—make the book worth reading.

Show Less

The book starts with a fantastic quote: Dewey Defeats Truman.” It’s an iconic picture. Harry Truman, grinning with uncharacteristic glee, holds the Chicago Daily Tribune up for all to see. “That is one for the books,” he gloated. The headline was a colossal screw-up, even by newspaper

Show More

standards. Truman had trounced Dewey, but the first edition of the Chicago paper declared otherwise—scrawled across its front page in the biggest, boldest type it had available. It’s not unusual for first editions to have errors, as they’re usually rushed to press. The Tribune’s first edition was even more rushed than usual because of a printers’ strike—it had to be put to bed in the early evening. This was unfortunate, especially on an election night; the deadline was before even the East Coast returns were in. Without these results in hand, the editors had to make a judgment call about what to put in the election story on the front page. They chose to make a guess about the victor—a choice that was spectacularly risky and spectacularly wrong. It didn’t seem so risky at the time. All of the major polling agencies, including the big ones run by George Gallup and Elmo Roper, had long since concluded that Dewey would walk away with an easy victory. The results of their polls were so definitive that the pollsters closed up shop weeks before the election. Since the outcome was obvious, they reasoned, there was no need to continue collecting data.

This book is about the misuse and misunderstanding of number crunching. It tells about cherry picking, gerrymandering, packing & cracking, plurality voting, Borda voting, and instant runoff voting and Arrow’s theorem. The examples are very much US-centric, but not in the least now and then very surprising and shocking.

Nice read.

Show Less

Proofiness is a particularly enlightening book and a must read for any discerning citizen in our democracy. As author Charles Seife explains, proofiness is the art of using bogus mathematical arguments. This book is a virtual catalog of bogus argument techniques. All having the common thread of

Show More

using numbers of some sort for support.
Seife takes as his basic premise that the average person today is easily taken in by proofiness because of the old adage; " Numbers don't lie." After reading this book you will become a convinced skeptic numerical evidence and more easily spot proofiness.
Seife's catalog of proofiness includes:

Potemkin numbers - aka fabricated statistics.
Disestimation - intentionally understating or ignoring uncertainties of data
Fruit packing - presentation of valid numbers in a way to cause proofiness including apple polishing, cherry picking and apples-to-oranges comparisons.
Casuistry - making a misleading argument through seeming sound principles, for example, mixing correlation with causality.
Randumbness - insisting there is order where there is only chaos.

Taken as a whole this catalog of disinformation techniques provides quite an array tools for nearly any organization or personnel with an ax to grind to "prove" their case.
Because voting and how we elect or representatives is at the heart of a democracy, Seife spends a significant part of the book addressing proofiness in voting. Two elections, Bush - Gore presidential election and Franken - Coleman senate election, provide a litany of proofiness. In the context of the Franken - Coleman contest Seife revealingly illustrates that depending on how you count votes in a "fair" election any of the candidates (including the third candidate, Dean Barkley) could have won!! Seife goes on to briefly introduce what constitutes a "fair" election and the mathematical truism that all elections are flawed with more than 2 candidates. This known as Arrow's Theorem and attributed to Nobel economist, Dr. Kenneth Arrow.
In summary, Proofiness is a very interesting and revealing discussion on the state of disinformation today.

Show Less

Aside from inventing words for concepts that already exist, the author does a fair job of describing how people routinely exploit the general public's difficulty thinking critically about quantitative issues. In part, this difficulty stems from cognitive heuristics of the kind that Tversky &

Show More

Kahneman have documented, but not all of the difficulty can be traced to these biases. Part of the difficulty arises from a simple failure to be skeptical of other people's claims, particularly when the claims are dressed up with numbers that make things sound "precise." Part of the difficulty comes from failing to think through whether or not a claim is logical or plausible.

Show Less

Nice easy read on the misuses and manipultion of numbers. Polls, government and lawyers all have their fingers rapped for some of their manipulations. Eye-opening read. Recommended.

Charles Seife, a professor of journalism debunk the faulty math used in daily life, particularly in government and policy-making. In a world driven by data and numbers, math if used badly, could undermine our democracy. The term proofiness is considered a pseudoscience that uses bogus mathematics

Show More

to prove something true even when it is not. With an introduction that ominous and compelling, I am intrigued and want to read more. The cover itself give the impression that those who use "proofiness" are playing magic tricks on the unsuspecting public. Seife uses words such as "befuddle," "hoodwink," and "disarm" to tell us how dangerous it to bombard us with faulty mathematics. Perusing through the book, Seife works through the math involved election polls, sea level rise, global warming. It is disconcerting to see how shortsighted policies can come about when numbers are manipulated for a desired result.
Seife did his due diligence in research. His bibliography has approximately 300 sources, both primary and secondary. With a big as dense in text and content as this, a detailed index is included. The table of contents uses provocative chapter titles such as "propaganda by the numbers." This book could be used as a summer reading option in a social studies or higher level math class.

Show Less

Rather a good book, but it needed a better editor. "as many as 5% of the people who are on death row are not guilty of the crime they committed" is not what the author intended, surely.

He invents or formalizes some good catch phrases for the different kinds of deception he describes: Potemkin

Show More

numbers, cherry-picking, comparing apples and oranges, apple polishing, randumbness, causuistry, dis-estimation, regression to the moon, risk mismanagement. He also uses a few good technical terms invented by other people like absolute malice and pseudoevent. He distinguishes very nicely between systematic and statistical error in polling. He describes how poll questions may be chosen to manipulate rather than discover public opinion. He makes the important, and little made, point that the counting of votes is a kind of measurement with considerable built in inaccuracy, and that the contested 2000 presidential election and the contested 2008 Senate race in Minnesota were, by the same standard we would apply when, e.g., comparing the length of two pencils with a ruler, a draw, and should have been treated that way. He does not support this point w/ enough statistical evidence, he does not prove that the inherent inaccuracy in the vote counting swamped the difference between the counts for the opponents, but his argument that this is another very significant case of dis-estimation is a good one. He discusses the practice of gerrymandering, which probably needs a separate book all its own. I currently live right next to the very first ever gerrymander! He discusses the pathetic misuse of statistics and expert witnesses who use statistics in the courts of law and the grotesque lying of supreme court justices. He discusses statistical propaganda and warfare. He discusses the precise meaning of "margin of error" when that phrase is used by news media.

He is positive about electronic voting and he demands that all the software be open-source! Good for him!

The book has a bunch of endnotes, but alas, no further reading section.

I often feel a tremendous weariness with the news, since it is so full of this utterly vacuous proofiness. Seife has given me a name for what makes me so tired and at least a framework for objecting to it.

Not a partisan book; he seems to have made a very deliberate effort to pick on Republicans and Democrats equally and to point out very clearly with what faciilty opposing candidates swap principles when changing circumstances make it expedient.

Show Less

This title of this book is a play on Stephen Colbert's concept of "truthiness." The author defines proofiness as manipulating numbers to prove something that you know in your gut to be true. Numbers are generally associated with factual information, so it is an easy way to promote an agenda. It was

Show More

an interesting premise, but I was really disappointed in how it panned out.

The first half of the book was not very interesting or ground-breaking. The author gives real life examples of how numbers have been manipulated to serve a particular purpose. He spends a lot of time addressing concepts I think I learned about in grade school: the difference between correlation and causation, the importance of having an accurate sample in polling, and how people can cherry-pick data to prove a particular point.

There were some interesting parts in the middle of the book where the author discussed the Gore/Bush presidential election and the Franken/Coleman senatorial election. It is fascinating how the election process, which seems to be cut and dry, can actually be quite arbitrary, and the numbers are not really close to certain.

Towards the end of the book the author starting to take on a very liberal bias. While I agreed with much of what he said, I think that he was guilty of the crime he was accusing so many others of: proofiness. He only presented data that proved his point. (He acknowledged this a little bit, but only in the footnotes.)

The book did leave me a bit more skeptical about the numbers I hear on the news and other media, but other than that I didn't gain much from it.

Show Less

I would say this is my MUST READ of the year. It's a witty exploration of the many ways numbers mislead us. We are programmed to believe numbers. If someone tells you that x is faster than y, well, that could be debatable. But if they tell you that x is 3 times faster than y, we accept it. We even

Show More

accept people telling us that this group of people is 2 times happier than that group when happiness is something we don't even know how to measure. A spoonful of sugar may help the medicine go down, but a number will make us swallow anything.

Seife explores the many forms of mathematical fallacies that trap us and gives them clever names such as randumbness and causistry (a mathematical casuistry) and many others. He uses current and historical examples of proofiness and demonstrates again and again and again who the use and abuse of numbers and our credulous acceptance of numerical propaganda is damaging our lives, our health and our democracy.

I loved Seife's early book, Zero, Biography of an Idea. This book brought less joy as some of the examples are infuriating, but he still has a clever and light touch that makes books about math easy and interesting.

I must confess, though, that his information on the derailing of the Nuclear Test Ban Treaty in 1982 left me stunned and seething. To keep the story to its most basic level, fraudulent test results were knowingly cooked up and released by a hawkish neocon in the Reagan White House to a complicit and equally hawkish NYT reporter. The false data was successful in derailing the treaty (which remains stalled to this day) but subsequently it was proved that not only was the data false, but it was deliberately and knowingly falsified for the specific purpose of breaking down the peace talks.

I have to ask if it would have made a difference 20 years later when the case for the Iraq War was being made if it were widely known that the 1982 leaker was Richard Perle and the conniving and complicit reporter responsible for the false stories that derailed the treaty was Judith Miller. Why were they able to say anything that anyone anywhere would take seriously? Why, when there were questions about the accuracy of the WMD stories did not one say RICHARD PERLE AND JUDITH MILLER LIED IN 1982? Did no one think that was relevant?

Show Less

This is a book about the misuse of math, particularly statistics, in public life, covering politics, elections (voting and apportionment), criminal trials (including the death penalty), and various other topics. I found it quite readable—the whole point is that you don’t need to know much math

Show More

to identify when someone is fooling you, or themselves, with numbers, because there are predictable habits of distortion. The appendix explanation of the “prosecutor’s fallacy,” which I know more formally as an aspect of Bayesian logic (see conditional probability), is very well done. (I once was involved in a case where we easily spent six figures educating the judge about why the prediction that two names would be confused with a 99% likelihood was bunk, since it ignored the extremely low base rate of confusion.)

Annoying: Seife coins some overly cutesy terms such as “randumbness,” though this habit also produced the truly brilliant concept of “regression to the moon,” where someone cluelessly or maliciously uses an “if this goes on …” extrapolation to show that, for example, if top women runners keep their present rate of improvement, they will beat men within a few years (and then proceed to break the sound barrier). (The accepted term on which this is a play, regression to the mean, or technically regression towards the mean, is the tendency of extremes to moderate, meaning that, all else being equal, the children of highly intelligent parents are likely to be less intelligent, something my parents were quick to remind me of when I was growing up—which says a lot about me and my parents.)

I initially thought that his statement that “[t]he Ken Lays and the Bernie Madoffs of the world are able to make their money by hiding risk and moving it from place to place, deceiving their investors,” was just wrong, because Ponzi schemes are not about risk; there is a certainty that they cannot work. Nor did investors generally think that later investors would get screwed but they themselves would pull out in time. On the other hand, the risk-hiding and risk-increasing contortions that the larger financial system went into are on a continuum with Ponzi schemes. And it’s certainly possible that some of the people involved in the Ponzi schemes convinced themselves that things would magically work out, because the alternative was too painful to contemplate, so in that sense they were willfully denying reality, but it still seems odd to me to describe this as concealment of risk.

Show Less

A very disappointing book. The author gives us exceptionally long-winded, boring lead-ups to what seemed to be promising insights, only to provide the most cursory discussion of those points. Also, most of the points in this book are well known and well documented, e.g. that the scale of a graph

Show More

makes a big difference in the impact of the data being presented. Where were the editors???

Show Less