Great Moments in Pedantry: Pie charts aren't so bad, after all

chartofcharts.jpg

Yes, that is a chart comparing the effectiveness of different kinds of charts. Apparently, it is Super Meta Science Day here at BoingBoing.

Pie charts get no respect, but Robert Kosara, Assistant Professor of Computer Science at UNC Charlotte, thinks the chart's bad reputation for being misleading isn't in line with the actual evidence. In particular, he cites a 1991 study that showed no significant differences between how well pie charts conveyed information, compared to other kinds of charts—unless the questions people were trying to use charted information to answer were particularly complex. In that case, pie charts were actually better than bar charts and tables.

In their discussion, they mention that adjacent segments in a pie chart are especially easy to combine, whereas the same is not true for a bar chart. And that, I think, is an important point: if we step away from the focus on accuracy in reading numbers, some of the more complex tasks are actually easier to do with the pie chart. The bar chart is great for reading and comparing individual numbers, but when it comes to adding up bars, it's a lot less effective.

Eager Eyes: In Defense of Pie Charts

Via Will Knight