Statistical literacy is more than being able to crunch numbers. Statistical literacy is the ability to think critically, objectively and logically to assess claims and data that are presented to us. Consider three claims that have been made by corporations or individuals in (relatively) recent years. Apple claims its M1 powered Macbook Pro gets "up to 20 hours of battery life". POM Wonderful, a pomegranate juice company, claimed consuming their product could aid in fighting heart disease. This past fall, and still today, some people claim the Presidential election was stolen due to voter fraud. Assuming all of these claims are false (spoiler alert: one of them is not) the repercussions vary from being relatively harmless to severely damaging. The more damaging the repercussions could be, the more important it is that people have the skills needed to assess these claims and take appropriate action when needed. Let's start by taking a look at the claims Apple made about its M1 processor during the Keynote presentation prior to the launch of the product.
Apple's claims fall in the relatively harmless category. These claims don't have any real effect on the daily lives of most people. If the computer doesn't really get 20 hours of battery then you'll need to plug it in a bit more often. That's the major repercussion. Even so, numeric claims like these that are a measure can be verified or proven false with relative ease. We would want to get a well represented set of use cases (video editing, web browsing, streaming video, etc) and then let the battery run out while completing each task. Since the product launched a few months ago, this data is now available. Some users are able to get the claimed 20 hours of battery life while others who do more cpu intensive tasks (like video editing) are not. This is why companies like Apple include the phrase "up to". They are not saying that every user will get 20 hours, just that its possible to get 20 hours under the right conditions. This example is a more base level of statistical literacy. Do we understand what the company is claiming? Do we know what kind of data we need to assess the claim? Can we get that data and perform the analysis? For measurable claims like this, we can almost always prove or disprove the claims. Others are a bit more difficult to prove and require a deeper level of statistical literacy.
From 2003 to 2010, POM Wonderful advertised that daily consumption of its product could treat, prevent, or reduce the risk of heart disease. That is a pretty powerful claim that can have a major impact on someone's life. The repercussions could be much more costly than needing to carry your laptop charger with you. What evidence should be required to validate that claim? Once a metric is chosen, say cholesterol and blood sugar - two of The Simple 7 from the American Heart Association - we would expect to see a randomized controlled experiment that measures these traits and supports the claim. More technically, we would expect a double-blind, randomized, placebo-controlled clinical trial. Does the necessary data exist? If it does, where can the consumer find it? Often this can take some digging. Sometimes it's a footnote in the ad itself, others require a bit more research either on the company website or another online resource. This is where our critical, objective and logical thinking skills come into play. We need an understanding of the different types of studies and trials that are used to support claims such as these as well as the ability to assess them.
The company supported their claim by citing their $35 million worth of research and clinical studies, published in peer reviewed journals, that supported the health benefits of pomegranates and pomegranate juice. These clinical studies are largely observational. Unfortunately that means they are much less rigorous and do not provide reliable scientific evidence. During their trial against the FTC, they were unable to provide a single clinical experiment that supported their claims (see page 4 of FTC's opinion).
Let's step things up further with a look at a claim that can be severely damaging to the integrity of a nation.
Even now, months after President Biden's inauguration, there are claims that the 2020 Presidential Election was stolen from the former President through rampant and widespread voter fraud. This is the type of claim that can have severely damaging consequences. Those consequences came to fruition in the form of the January 6th insurrection in Washington DC. While the topic itself can be polarizing due to the current political climate in our country, it does give an excellent example of how statistical literacy can help solve even our biggest problems. A case like this can be assessed by calmly and logically asking questions -- questions that we can actually answer. How many cases of voter fraud have been reported? Are they isolated incidents? What was the nature of the cases? Were people caught voting multiple times or filling out ballots for deceased family members? If so, how many times? How can we classify or categorize the reported cases? Are there enough verified cases of fraud that it seems conceivable that the election was swayed by such votes?
To paraphrase Charlie Eppes from the TV show Numb3rs, "Statistics is more than formulas and equations. It's logic. It's rationality. It's using our minds to solve the biggest mysteries we know". Everyone needs to be statistically literate on some level. We need the skills necessary to analyze and interpret information on a daily basis. It helps us decide if we need to bring our laptop charger with us to class. It helps us decide what we can consume to promote better health. And it helps us ensure the integrity of our democracy.