« Energy- and search-efficient | Main | Light entertainment: cougars alert »


Alex Kerin

Two other flags for me:

Their suggestions for the 'middle' ground of chart adornment resulted in one ugly chart and one unreadable chart.

Second, for the prompts they asked things like "Did you remember the chart called Monstrous Costs?" Because of the labeling, the subject would be much more likely to remember that the chart looked like a monster if they had seen that version.


"The sample size consisted of 20 students. Flipping open any elementary statistics textbook, you will find the standard advice to ignore experiments with fewer than 30 observations."

Umm, I guess you haven't flipped open that same elementary statistics book? The number of participants is based entirely on _power_, not an arbitrary number. Many famous studies have have had a dozen or less participants.

You're right in many of your other critisisms, but the main value of this paper is its willingness to ask the question in the first place. Much of Tufte has been assumed as gospel, and this is the first time someone has checked whether his guidelines should always be followed. Note that the paper clearly states the value of chartjunk depends on the authors goal, and that ample additional studies are needed.


Hi Kaiser, I'm very happy to read a critique of this paper by you. The thesis of the paper is very appealing to me but I couldn't believe it would be embraced unanimously by the charting community. (see also http://www.excelcharts.com/blog/no-tuftes-charts-are-not-plain-and-simple/)
about your red flag 7 -

Is a good chart defined as one that leads readers to retain displayed values? Not in my book.

this is key.
In some contexts this is the whole purpose of a display. In others, it is completely secondary.
this depends on, say, the intended audience, the subject, and the degree of engagement of the reader.

so I work at OECD, and in charts that appear in specialized publications which are read by researchers, I advise to remain completely Tufte-compliant with a very clean and sober design. People who find these charts were asking for it. They will spend as much time and energy as they want to understand them.

but for charts that appear on leaflets, on the side of our web sites or on posters at our HQs, since they need to conquer the reader's attention, I'm happy if they are a little quirky.

Andrew Gelman

Interesting that the Information Aesthetics people fell for this study. Perhaps because it told them what they wanted to hear?


Surely there's a bigger picture here. Tuftian design principles are for decision-makers who are already interested in the underlying data and who have to analyse it day in/day out. If we ask the random person in the street whether they remember a chart of random numbers they arent interested in or a stylised picture of a chorus line girl, most people (weeks later) will remember the chorus girl. They then may remember that the "graph" traced the line of her leg and sales went up, then down. Our brains are hard-wired to remember images. However for decision makers this isnt an option... unless we hire a graphic artist every time our diamond sales are updated. Sales are flat, the chorus line girl is doing the splits. Sales are up, she's doing a high kick. Sales are up even more, she's doing a high kick which is a few dgrees higher than last month. If you have one piece of information to impart to an ambivalent audience then junk charts have value. Otherwise, stick to the facts and only the facts. Just my thoughts!



If you had bothered to read the paper you would know that the authors clearly state that whether chartjunk is useful or not depends on the context. It really depends on what you are trying to do, and the authors state this in their paper. That doesn't make their study invalid.


The comments to this entry are closed.


Link to Principal Analytics Prep

See our curriculum, instructors. Apply.
Kaiser Fung. Business analytics and data visualization expert. Author and Speaker.
Visit my website. Follow my Twitter. See my articles at Daily Beast, 538, HBR.

See my Youtube and Flickr.

Book Blog

Link to junkcharts

Graphics design by Amanda Lee

The Read

Keep in Touch

follow me on Twitter