Edward Tufte knows some things about the visual display of quantitative information. And, as the author of a widely admired book on the subject, he wants to tell you about them. He thinks you're doing it wrong.
He's got a legitimate gripe. If you traffic in quantitative information, then you may have experienced the occasional twinge of unfocused discomfort that the data you are presenting is somehow unclear or misleading. This is often the result of failing to solve the fundamental problem in presenting quantitative data: how best to display 4-D data (3-D space + time) on a 2-D surface fixed in time.
Tufte has devoted his career to solving that problem. Several themes run through his work. One is resolution. He is an enthusiast of paper as a tool because of its resolution, which is many times higher than the best computer screen. But Tufte is no anti-computer Luddite. He thinks that paper and pixels are complementary tools. The problem is their misapplication.
Tufte thinks that typical computer-screen real estate and resolution artificially restrict both the rate at which a viewer can take in data and the viewer's individual methods of intake. He believes that it is better to present large amounts of data on a large piece of paper rather than successive screens on a computer. As Tufte puts it, "spatial adjacency is more effective than temporal stacking."
A related theme is to "design the data presentation to serve the analytical task." This is where Tufte thinks many presentations fail. Shuffling back and forth from screen to screen is less conducive to perception and analysis than having the data spread out on the same surface, even if it requires a wall-sized chart. To Tufte, a wall chart is not an unattractive option but rather an ideal method in situations that require it for the analytical task at hand.
The ability to coherently present a large amount of data, if it's required to prove your point, gives your presentation credibility, according to Tufte. And it diminishes viewer suspicions that the data is cherry-picked, which is a pervasive persuasion technique.
OK. You've accepted Tufte's theories. Now, how do you actually go about creating an effective presentation? Tufte thinks that most tools available to the typical user are inappropriate for the task. And one tool in particular, Tufte strongly believes, is especially ill-suited for creating effective presentations of data and information.
That tool is PowerPoint. Tufte has some interesting theories about the ubiquitous application, which he discusses in an essay, "The Cognitive Style of PowerPoint." The title suggests a neutral position.
Tufte's opinions are far from neutral. "Bullet outlines dilute thought," "PowerPoint has a low signal-to-noise ratio," "We should not abbreviate the truth to make the words fit," and other criticisms give you a better sense of his position on PowerPoint (a position Tufte also takes on every other presentation application of its type). If you agree, then you would enjoy Lincoln's Gettysburg Address done as a PowerPoint presentation, a hilarious parody created by a friend of Tufte. It's all the more so because his friend claims that the presentation was mostly done with the AutoContent wizard.
"PowerPoint's hierarchy is positively medieval," Tufte says, and forcing information into it creates a cognitive style that he calls "faux-analytical." He cites the example of Richard Feynman, who wrote a 600-page book covering most of basic physics using only two levels: chapters and headings within chapters. By simplifying thoughts and organizing them into a hierarchy, "PowerPoint is therapy for presenters," Tufte says. He thinks that it has little or no benefit for consumers of the presentation.
One reason why many PowerPoint presentations seem thin is, well, because they are. Tufte once made a study of 28 books on PowerPoint presentations. They contained a total of 217 data graphics, with an average of 12 numbers each. "This is astonishingly thin, nearly content-free," he says. Tufte points to data graphics published in a wide range of periodicals that far exceed this number. Other examples close to home are the sports and business sections of newspapers, which contain vast amounts of quantitative data, all of it spatially adjacent and easy to comprehend. Here, as Tufte would surely say, the data presentation is designed to serve the analytical task.
Making fun of badly done PowerPoint presentations can create the impression that no real harm is done, but the consequences can be serious. Tufte performed an analysis for the Columbia Accident Investigation Board of three PowerPoint presentations created during the last flight of the space shuttle Columbia. These were directed at NASA officials making decisions during the shuttle's flight. In its final report, the board said that it "...views the endemic use of PowerPoint briefing slides instead of technical papers as an illustration of the problematic methods of technical communication at NASA."
Edward Tufte has other strongly held opinions about presenting information. If a twist of fate should ever have you giving a presentation to him, do not - repeat, do not - use the word "significant" anywhere in your presentation. Trust us on this one.

Comments