This is a group of examples related to statistical chicanery or fallaciousness in general. Depending on your class, the readings below may or may not have been assigned, so check your syllabus.
Readings for Analysis
Usually, I bring up the following reading around Halloween, but it’s interesting to consider in terms of statistics:
The above article uses statistics to debunk vampires and zombies. Although zombies exist metaphorically (see Black Friday shoppers), they’re fictional representations making social commentary, so we can critique those narratives based on artistic as opposed to scientific merits. The same is true about vampires; however, I think the above article misses an important factor about vampires when stating, “if a vampire sucked one person’s blood each month–turning each victim into an equally hungry vampire–after a couple of years there would be no people left, just vampires” (“Vampires and Zombie,” 31 Oct. 2006).
- What if the vampire bite didn’t always turn a victim into a vampire but killed them?
- What if vampires could live off animals?
- What about feeding off of already dead people?
- How about if they robbed blood banks, so they drank blood but never bit a human?
Sometimes even science can’t answer the most important questions in life.
Al Gore’s Earth in the Balance Selection
Below I have a discussion about a statistic from Al Gore’s Earth in the Balance (1992)–on Canvas. The overall point isn’t to debunk evidence of the Earth warming; instead, it is about our (social, perhaps) commitment to facts, figures, and statistics.
The Al Gore excerpt is relevant for our next discussion on statistics, but I wanted to use it as an introduction for critically thinking about statistics and the false equivalence fallacy.
- Gore explains that the scientific community is overwhelmingly in agreement that Global Warming (he wrote this in 1992) is happening, yet some people want “equal time” devoted to the other side, the group that rejects the Earth is warming because of human activity. Gore’s point is that it’s misleading to give the opposition “equal time” because doing so implies there’s more doubt than there is or that both perspectives have the same amount of support. Again, overwhelming evidence, supported by the scientific community, demonstrates the Earth is warming. The small fringe group of scientists shouldn’t get “equal time” because that actually distorts the scientific view.
- Gore also does something that’s misleading, but it doesn’t really take away from the overwhelming scientific evidence of global warming. He states “when 98 percent of the scientists in a given field share one view and 2 percent disagree, both viewpoints are sometimes presented in a format in which each appears equally credible” (pp. 38-39; emphasis added).
- I’m not trying to debunk global warming or claim Gore is way off base with his claims. If anything, the fact that his source is nearly 30 years old, and we’re feeling the effects of global warming (now called “climate change”) proves his reporting was correct.
- However, Gore uses percentages to describe a scientific field’s agreement and disagreement regarding global warming. His percentages aren’t real. By “real” I mean they aren’t a survey of all scientists or a representative sample of scientists in climate science. If they do represent a survey or other enumeration, he doesn’t provide evidence.
- Again, I’m not refuting the idea that humans are causing the Earth to warm. I only want to point out that Gore is using 98 percent in ways that we might in everyday conversation, which doesn’t follow strict scientific verification.
- How often have you claimed, “99% of the time, I take an Uber if I go out drinking”; or “90% of the time, I’m happy with my spouse/partner/significant other”?
- We say “99% of the time…” quite often when we really mean the following: “My perception is that I overwhelmingly do something very often.” We can’t possibly quantify all our habits in such precise percentages, so our claims or “99%” can’t be taken literally.
The takeaway here is to be careful when presenting statistics or what you hope to pass off as “overwhelming evidence” by grabbing a random “98%, 99%, 99.9%, etc.” out of thin air when your audience expects quantifiable evidence gathered appropriately according to the field’s adherence to the scientific method.
Facts, Figures, and the Fine Print
- Height of Inequality
- This is a rather interesting representation of wealth based on percentile (not percentage) of the population.
- The size of the figure represents the amount of wealth an individual in a particular percentile has.
- When you get to the far right of the chart (99.995th percentile), the figure is represented as being 933 feet tall. Of course, all you see is the figure’s shoe.
- What message is this chart conveying about those who’ve accumulated an enormous amount of wealth?
The next discussion is about several charts and graphs related to HIV/AIDS infection among African Americans. The first link below goes to a webpage with bright graphics and facts and figures clearly and effectively placed on the page–good design overall. The second link goes to the full report and opens on page 11, which is the “Technical Notes” section that discusses limitations to data collection.
- HIV/AIDS among African Americans
- Don’t miss the “Technical Notes” section (p. 11) of the 2018 Surveillance report
- I’ll give you a hint: Read the 2nd paragraph that starts out “Please use caution when interpreting data on diagnoses of HIV infection. HIV surveillance reports may not be representative of all persons with HIV because not all infected persons have been (1) tested or (2) tested at a time when the infection could be detected and diagnosed.” (p. 11). The link should jump right to page 11.
- This discussion would be better if we were face to face, but I’m going to present some further critical analysis of this report.
- First, I’m not suggestion that anyone is manipulating data or flat out lying about disease transmission; however, we can’t ignore the word “surveillance” in the title without considering health departments’ desire to keep track of infections.
- Second, the CDC derives these statistics from states reporting infections, so the CDC relies on the accuracy of over 50 different entities (don’t forget Washington, D.C. and US Territories) reporting the results from all the different health departments with a state of territory. It’s unlikely there is anything like uniform reporting.
- Even a cursory look at the news will show you that African Americans are disproportionately affected by COVID-19 and a host of other ailments. Much of the health disparities may be linked to socioeconomic status, and, if African Americans are disproportionately affected by poverty, and, if we can assume health care is lacking in poorer communities, health care disparities become obvious.
- One theory (that hasn’t been verified because I’m not an epidemiologist) I have based on a critical thinking thought experiment is that it’s possible African Americans are under more surveillance for infections than other groups.
- Health clinics in economically depressed areas might be more likely to report all cases of infection to the CDC because they’re mandated to in order to comply with community health rules. On the other hand, people in more affluent areas with private doctors screening them might be less likely to report all cases.
- The parallel here is, I hope, obvious during this tumultuous time and protests against police brutality. One thing you might have heard is that predominantly African American communities are “policed” more; they have more police surveillance and, thus, more attention leads to more arrests.
- Looking at the CDC numbers in regard to the amount of surveillance is important, so we aren’t pathologizing a group because of deep-seated prejudice conditioned by systemic racism.
But don’t just take my word for it–do some research. Again, read the fine print. On that same CDC Surveillance report’s “Technical Notes” they state, “testing patterns are influenced by many factors, including the extent to which testing is routinely offered to specific groups and the availability of, and access to, medical care and testing services” (p. 11; emphasis added). In order to think critically, you have to not just read the headlines and captions; instead, you have to understand how the data were compiled.
“How to Lie With Statistics”
Huff’s “How to Lie with Statistics” is on Canvas, so I expect that you read it already 9depending what class you’re in). Below are some old images that might still be useful for this discussion on distortion. When do visuals lie? What should you do to make sure you’re accurately representing the facts in a visual?
- Averages: mean, median, mode
- Pictorial representations
- Decimals
What’s going on in the chart below? Does it represent twice as much?
Maybe we have time for this old PowerPoint.
One More Visual
Although we would all agree that purposely manipulating data and visuals would be unethical, we might not readily recognize that even inadvertent manipulating data or presenting unclear information is unethical. A technical communicator–and this goes for anyone communicating technical information regardless of title–must convey information accurately and in a way the audience will understand.
Take a look at the visual below. It’s a graph that was included in an old business writing textbook as a good example of an area graph. What do you think? What does the graph represent?
I’m not 100% sure, but my assumption is that this is supposed to represent income distribution equity based on population…but I’m not betting the ranch on that assumption. Although there are other problems, below are four major ones:
- There is no key or details on what those numbers represent. Are they actual figures? Are they out of 1,000? 10,000? 100,000?
- The gray population values have six numbers, but the X-axis only has five areas: Africa, S.E. Asia, Middle East, Far East, South America.
- The blue ‘GNP Distribution’ has five values marked.
- There are hash marks on the x-axis that don’t correspond to any discernible region, making the reader ask, “are they missing geographic areas, or am I supposed to know what ‘Africa and a half’ is”?
- The order of geographic regions follows no discernible order. The order isn’t alphabetical or based on an east-to-west or west-to-east orientation.
Most bizarre, huh? This “image” comes from the following Business Writing Textbook:
Murphy, Herta A., Herbert W. Hildebrant, and Jane P. Thomas. Effective Business Communication. 7th ed. Boston: McGraw-Hill, 1997: 575.