This relies on citing the authoritative body, and is identical to Science authority but with a focus on statistics. In certain cultures, there is a sort of ‘math phobia’ and people accept anything mathematical as semi-divine. This is the darling of every politician and commercial enterprise: “Nine out of ten negotiators surveyed agree.” But statistics may not support a position given. The mean, mode, and median are terms some use interchangeably, but are very different. The mean is determined by adding up the values in the data set and then dividing by the number of values added, also known as the “average.” The median is the value which appears in the middle of a list made of the data set in numerical order. This is often misleadingly called the “average” by politicians. The mode identifies which value in the data set occurs most often.
One may claim ingredients found in an ordinary food are toxic, such as lead in chocolate or arsenic in apricot pits. Or, for example, one could claim that allergic reactions to the HPV vaccine claim the lives of children every year at a rate which is much greater than other vaccines, according to a survey.
Distinguish in several ways: 1. The information supplied does not solve the problem. 2. The information is irrelevant. 3. The information is absurd when compared to other things.By applying these solutions to the examples given, we can defend the chocolate and peach pits by arguing that “Toxicity is not the issue, but rather risk to public health,” or “the allowable levels of lead in tap water are greater than those in a chocolate bar.” We could counter the HPV vaccine claim by arguing that “a survey is not a scientific clinical trial,” or that “your child is more likely to die as a convicted felon given the death penalty than from an allergic reaction to the HPV vaccine.” Use Cialdini’s influential communication tactics. Perhaps the most compelling way to undermine a statics-dropper is to look at the math. Focus on the statistical methodology itself: what is the standard deviation, how big was the sample set, when was the data taken, is it too old, and so on.