Chapter 9, “How to Statisticulate,” explores the topic of how to determine lies that have been presented as truths with the help of statisticulation. Statisticulation refers to the use of statistical material to cause manipulation, thus having an influence on a target population and the subjects that concern them (Huff, p. 102). The authors give examples of bluffs associated with using misleading information, such as wrong rates, percentages, and the misrepresentation of graphs and maps. One of the examples of statisticulation is the use of maps by companies. When making maps, which are the graphical representations of how organizations perform, marketers distribute the statistics for making either of the axes finer in order to make data seem steeper and one-sided.
Today, such a method is used when businesses compare their data or products to another company, to make theirs look better and more prominent. Therefore, while the graphs between two companies seem staggering in the maps, the difference in performance is only in 1 to 2% between each other. Such an example is evident statisticulation because a company wants to manipulate statistical information without lying about it. While the data may not be over-emphasized intentionally, the graphical representation works in favor of a company that manipulates the potential audience. In addition, statisticulation also implies the excessive use of percentages, which offer significant confusion for the target audience. Any percentage figure that is based on a limited number of cases is more likely to be incorrect and misleading. However, companies widely use percentages because they are perceived as reliable and favoring some companies over others in terms of their performance.
In Chapter 10, “How to Talk Back to a Statistic,” the author points out the ways in which statistical data can be challenged by simple questions and finding answers that would help to avoid being manipulated by statistics. The first question, “who says so?” implies looking for bias in the presentation of information. For example, if there is data showing that a cleaning product performs ten times better than other products, the statistical findings are most likely to be manipulative if the data comes from the company manufacturing the product. The second question, “how does he know?” means looking out for evidence of a biased sample. For example, when measuring the effectiveness of a skincare product, the sample should be large enough to provide reliable data. If the sample is ten individuals, it is highly likely that the sample is biased.
The third question, “what’s missing?” means watching out for unspecified means and medians that may substantially differ. For example, statements such as 79% of women found the product to be effective for clearing their skin. Data on the number of women, the period of time when the product was used, and other variables were not mentioned, making the percentage deceptive. The fourth question, “did somebody change the subject?” means that people should determine whether a statistic is real or reported by someone. For instance, this means that the effectiveness of a skin product should be based on real data measured by independent parties rather than the company manufacturing it. The fifth question, “does it make sense?” means limiting a statistic down instead of making conclusions on unproven assumptions. For example, when a company makes a generalized assumption about the effectiveness of a skincare product in marketing material, it will make statements such as “it has been proven,” while no substantial evidence to support that is given.
Work Cited
Huff, Darrell. How to Lie with Statistics. W.W. Norton & Company, 1954.