Book Reaction: How To Lie With Statistics by Darrell Huff

How To Lie With Statistics is a best-selling critique on various weaknesses that may affect the validity of numbers used in statistics. Although Darrell Huff was not a statistician, he understood so much about numbers because of his journalism career. In the book, Darrell argues that many reports prepared by journalists or marketing organizations contain numbers because people are more likely to believe what they can quantify. The use of data has also assisted many companies to gain more insights into the market and modify their products to meet the needs of the consumers. Darrell’s primary argument is that statistics are good because they contain some truths that can help people especially managers to make effective decisions. However, statistics become vulnerable when some people want to use them to achieve malicious intentions. The problem is that statistics may not defend itself when people want to use them to support their diabolic agenda. Among the issues that Darrell Huff criticizes include sample biases, well-chosen averages, little figures that do not make a difference, incomplete data, and unintelligible graphs. With the mentioned challenges, it becomes difficult for people to rely on statistics even if they may be true. This essay discusses my reactions as well as lessons I have learnt from the book.

What surprised me after reading this book is that I may have believed statistics that are not true. I have always believed that the information presented by statisticians are factual and verifiable. For example, I have never doubted that about 50% of Americans are at risk of becoming obese. What I have never asked is the origin of such data and why it exists. According to Darrell Huff, I belong to a fact-minded culture where statistics are appealing regardless of their sources. Darrell says that the secret of statistics lies in the intentions of the pollster (Huff 10). What statistics say may be intended to persuade, “inflate, sensationalize, confuse or oversimplify” (Huff 10). The 50% obesity prevalence may be a sensationalized statistics to warn people against eating junk foods. On the other hand, the percentage may also assist grocery stores to make their products more popular. The hidden intentions of statistics are something that did not concern me until I read the book. The author provides more examples including the use of Gee Wizz Graph

to present information. The graph, according to Darrell, has hidden information, which observers may not notice. For instance, most of the graphs do not start at point zero the way many people expect. Again, the uses of three-dimensional figures create more confusion because people only see the numbers without a verifiable scale (Huff 15). In the future, I feel like I need to be more conscious about the source of data and verification before believing the contents.

What impressed me or ‘blew me away’ is the style of writing used by the author to express his arguments. The author has effectively used rhetorical styles such as ethos, pathos, and logos to make his arguments more appealing. Ethos is is the use of ethics or moral values to support an argument. For example, Darrell argues that manipulations or distortions of data to achieve a selfish goal may mislead people into making wrong decisions. For example, during elections statistical misrepresentations may lead people into choosing wrong candidates. By exposing the tricks, Darrell believes that he is enabling innocent people to defend themselves against manipulated data. The author also appeals to the emotions by using simple and illustrated texts that evoke a sense of reality. The author chose not to go so deep into statistics because it could have been difficult for many people to understand his points. His use of evidence from the news sources including Fox News makes his arguments more appealing to the logics. As a journalist, the author understands that a good argument is one that is factually true and verifiable. The use of evidence also provides insights that future researchers may use to develop detailed arguments against statistical misrepresentations.

I also learnt new phrases used by Darrell Huff to describe manipulated statistics. For instance, one of the new phrases I learnt is the Gee Wizz Graph. Although I had seen graphs being used in the media to present large data, I did not understand how ineffective they could be. The Gee Wizz Graph, according to Darrell, is an oversimplified representation of big data. The graphs appear ambiguous and decorative because they only have the large figures without details that people can use for verification purposes. Darrell gives an example of Obamacare Enrollment, which indicates that by March 27, only 6 million people had registered against a monthly target of 7 million. The data appears ambiguous because many people cannot relate the 6 million figure to the image. Another new phrase I obtained from the book is well-chosen average (Huff 28). According to the author, many pollsters pick one out of the three averages such as mean, median and mode to support their arguments. In a case where the mean differs significantly from mode and median, pollsters are likely to select the mean and discard the rest (Huff 121). In other words, what pollsters say about standardization may be a well-chosen average to support their thesis. Darrell also talks about semi-attached figures as a strategy that many marketing companies use to spread misleading information (Huff 122). A semi-attached figure is a type of data that does not relate to the conclusions made in statistics. For instance, when marketers use doctors to adverse products such as Marlboro, their intention is to portray the products as safe for consumption (Huff 132). However, the attached figure does not justify the number of lives that have been claimed by cancerous nicotine present in the cigarettes. The marketers may also give little or no considerations to the harmful effects of such messages to the consumers. According to the author, such misleading information is common in countries with weak consumer protection system (Huff 131). I plan to conduct more research on the new phrases to determine how they affect the validity of statistical figures.

What I like about the book is the simplicity, practicality and the ability to address issues that statisticians have chosen to ignore. Before Darrell wrote the book, many people did not understand the extent to which statistical data may be misleading. Many people do not listen to the silent voices of those who did not talk to the surveyors. For example, when statistics indicate that 58% of Americans would like Marijuana legalized, it does not talk about the silent voices. Some researchers may also pick a biased sample to ensure the results reflect their intentions (Huff 15). Therefore, there is a need for critical analysis of statistics and Darrell gives an effective formula that people can use to overcome manipulation of data. What I find discomforting is the fact that the author seems to ignore the rigorous process that goes into the collection of data. Many researchers spend years looking for information which may be necessary in the society. The fact that Darrell completely ignores the work done by researchers may be discouraging to the future statisticians. Darrell ought to have highlighted some of the strengths of statistical data to develop a more objective argument. What I find funny is the use of illustrations throughout the text. The cartoons that have been used throughout the text stimulates the visual appeal and entertains the readers.

This book should be used in high school or college because it trains students on how to conduct critical analysis of statistics they use daily or read in the media. The book highlights various weaknesses that may affect the credibility of statistics. Apart from the critical analysis, the book also trains students on how to avoid various mistakes when conducting research. For example, when choosing a sample for experiment or survey, students should ensure fair representation of the population to avoid biases. The book also talks about the silent voices which many researchers are likely to ignore. Apart from sample biases, the book also trains students on how to overcome the ambiguity when presenting information using visual tools such as graphs. According to the author, graphs without verifiable scale may provide misleading information to the readers. Another source of ambiguity is the use of figures that do not relate to the conclusions. The author’s argument is that statistics should contain simple information that readers can easily understand and verify.

After reading the book, I have gained useful skills that will shape my future analysis of statistical data. In Chapter 10, How to Talk Back to Statistics, the author gives five steps that people can use to critically analyze statistics. The first question to ask is about the source of information. The reason for analyzing the source is to determine if the source is official or bias. Sources like government institutions and the mainstream media are likely to present accurate and verifiable information. The second question examines the methodology used by researchers to collect and analyze data. A wrong methodology is likely to yield misleading information. The third questions examines missing information which may affect the arguments presented in the statistics. For example, if the researcher has used mean to describe data, readers should determine if the mode and median reflect the same information. Some researchers may omit information that contravenes their overall thesis. Therefore, if there is missing information, the author provides a fourth question which examines if the omission was deliberate or accidental. Deliberate omissions are likely to affect the overall outcomes presented by statistical data. The final question examines if the data makes sense or is unintelligible. According to the author, researchers should use statistical data to provide information needed by the society to solve a problem. Therefore, I shall use the critical thinking approach provided by the author to analyze information before making decisions.

Works Cited

Huff, Darrell. How to Lie with Statistics / by Darrell Huff ; Illustrated by Irving Geis. New York : Norton, 1993.