I don’t think it’s much of a stretch to believe that we are more likely to remember information that is presented visually than as text. Indeed, the Pictorial (or Picture) Superiority Effect has been long studied showing how much better we are at remembering and recognizing information presented visually. Unfortunately, because some people want to assign specific quantities to this concept, bad statistics are passed around as truth.
In some of the seminal research that laid the theoretic groundwork for the Picture Superiority Effect, Allan Paivio (1971, 1986, 1991) developed the “dual-coding theory.” The theory states that when we are exposed to different stimuli, we form mental representations and, depending on the stimuli type, we have the possibility of encoding them twice, once as a verbal code and again as an image code. For instance, if I said the word “chair,” you can encode it twice: a verbal code for the word and an image code because you can easily picture what the word is about. When we have two codes we increase the likelihood for memory because we have two memory traces. Compare this with something abstract like “optimization.” This creates only one code—the verbal one—and it is harder to come up with an image for it, and is thus less likely to be recalled.
The scientific research around the Picture Superiority Effect and dual-coding theory attempts to measure the differences in recognition and recall between different types of stimuli, be them text, images, videos, and oral presentations. In one study, for example, participants reviewed more than 2,000 pictures at a rate of 10 seconds each and, on average, they were able to accurately recall more than 90% of the images several days later.
But there are lots of other statistics about how we process, recognize, and recall information that are not attached to actual citations in the scientific literature. Some bastardized version is grabbed and published—and because it sounds sexy and true—it gets passed around and around. Here are three numbers that I regularly see that do not appear to be rooted in actual scientific research.
The 60,000 Fallacy
I wrote about this one a couple of weeks ago, so I won’t rehash the entire argument here. The basic claim that “visuals are processed 60,000 times faster than text” is rarely cited and when done so, the root citation goes to a document that includes no references to an actual scientific study.
As an aside, I don’t think infographics should be used as the source material for research findings and results (I’m looking at you, Harvard Business Review). Doing so is just lazy. Many infographics are not presenting original research, but instead are summarizing or presenting research in different (and sometimes better) ways. If you are in the habit of citing infographics, go instead to the original source material and cite it directly.
This trifecta is thrown around pretty regularly (see, for example, this infographic found on Hubspot). The basic story is that people remember 10% of what they hear, 20% of what they read, and 80% of what they see and do. Do I believe the basic idea behind these numbers? Sure I do. Do I think these numbers are accurate? I have no idea. But the study on which these numbers are based appears to be a bastardization of some early work on audio-visual research (this 2006 post by Will Thalheimer is a great summary of the history).
The primary source for this claim most often used today comes from a four-page publication by Hewlett-Packard entitled “The Power of Visual Communication.” The authors (who are not named) point to a paper (apparently unpublished) by Paul Martin Lester (professor of Communications at California State University, Fullerton) with the title, “Syntactic Theory of Visual Communication.” In it, Lester states that, “Educational psychologist Jerome Bruner of New York University cites studies that show persons only remember ten percent of what they hear, 30 percent of what they read, but about 80 percent what they see and do.” There are no citations or references provided for the statement.
I emailed with Professor Bruner, now 99, a few weeks ago and met with him while in New York in early October; he does not recall saying or writing anything to this effect. I was also unable to find any such claim in Bruner’s voluminous writing, but I’m no psychologist, so perhaps it’s in some paper I have not read.
Dr. Carmen Simon, cognitive scientist, confirms that the “10%-20%-80%” is false. “We know it’s false in several ways,” she told me. “Over the decades, the numbers have appeared in many formats, not just the 10-20-80. For instance, another version is cited as ‘people remember 10% of what they read, 20% of what they hear, 30% of what they see, 90% of what they do…’ Very few statistics from studies form such a conveniently progressive string. And how is reading different than seeing anyway?”
Another numbers trifecta, similar to the one above. This is perhaps my favorite because it comes from John Medina’s terrific book, Brain Rules, and is one that I have used in the past myself. The problem is that the numbers come from a single document that includes no references or source documents.
In Chapter 7 of Brain Rules, Medina writes, “Text and oral presentations are not just less efficient than pictures for retaining certain types of information; they are way less efficient. If information is presented orally, people remember about 10 percent, tested 72 hours after exposure. That figure goes up to 65 percent if you add a picture.” The online companion to the book includes a number of sources, but none refer to these specific numbers. I emailed Medina some time ago and he provided three additional possible sources:
- McBride, D.M. & Dosher A. B. (2002). “A comparison of conscious and automatic memory processes for picture and word stimuli: a process dissociation analysis,” Consciousness and Cognition 11(3): 423-460.
- Renninger, KA, Hidi, S and Krapp, A (eds). 1992. The Role of Interest in Learning and Development. Lawrence Erlbaum Associates (London).
- Occupational Safety & Health Administration (OSHA). 1996. “Presenting Effective Presentations with Visual Aids,” U.S. Department of Labor (May).
The first paper is certainly in the ballpark and tests the Picture Superiority Effect but the authors do not make explicit tests of oral stimuli. Similarly, the Renninger et al book does not test oral stimuli recollection; certain chapters are concerned with learning from text, but they do not explicitly test the Picture Superiority Effect or recollection of oral stimuli. Admittedly, I did not read the entire book word for word, but Professor Renninger confirmed to me via email that the book does not include these numbers.
It appears that Medina’s numbers come from the last source, a May 1996 memo from the Occupational Safety and Health Administration (OSHA) entitled, “Presenting Effective Presentations with Visual Aids” (good luck parsing the title). In it, the authors (who are again go unnamed) include the graph shown here and the subsequent text.
The document includes no references or links to any other paper, report, or website. Although I love the Medina book, I hesitate to use these numbers because there is no way to review the supporting scientific evidence.
There are countless blogs, infographics, books, slidedecks, and other materials that argue the importance of visual stimuli over simple text or tables. I don’t disagree with this basic premise: My personal experience is consistent with this idea and there is a sufficient scientific literature that supports the Picture Superiority Effect showing it to be the case. But it’s nearly impossible to find the scientific evidence to support such specific numbers like those shown above. Don’t get me wrong: I’m not arguing that everything you see in a peer-reviewed journal is correct (that’s clearly not true!). The next time you are writing a paper or blog post, making an infographic, or posting a slidedeck touting how much better visuals are than text, find a number that you can actually stand behind.