This article, from an academic and research librarians blog, reviews the debunking of the diagram I refer to as "learning heads" and often use when introducing a course at the beginning of a semester. I use the diagram more as an argument, than as a proof, so the fact that the claims have been debunked doesn't totally rule out its use, provided that distinction is made.
However, it is a great example of an abuse of "statistics" and the need for sources. In this way it is a good teaching tool. As the author says, it is appealing because of its "truthiness." I got it from what I took to be a trustworthy source -- Educating for Change -- where it appears with a seemingly legit citation: "NLT Institute for Applied Behavioral Science."