Blog

Null Toleranz: Warum alle Achsen von Säulendiagrammen bei Null schneiden sollten

Ein ausführlicher Artikel bei The Mendoza Line thematisiert Achsen von Säulendiagrammen, die nicht bei Null schneiden, und welche verwirrende Wirkung diese auf den Betrachter dadurch haben. Diese Verwirrung kann sogar als  Lie Factor ("Lügen Faktor") quantifiziert werden. 

In the figure above, I condensed the original version from the Washington Post to a few bars and measured the length of the longest and shortest bar in millimeters. I did this to calculate the Lie Factor, which is a measure Tufte came up with to determine how inaccurate misleading graphics are. The Lie Factor is calculated by dividing the size of the effect shown in the graphic by the size of the effect in the data.
In this instance, the bars displaying the ages of Ronald Reagan and Marco Rubio are 402 mm and 54 mm, respectively. This means the increase shown in the graphic works out to 644.4%. However, this seriously overstates the effect in the actual data. Reagan was 77 when he was inaugurated into office and Rubio would be 45 on his hypothetical Inauguration Day. That’s an increase of only 71.1% from Rubio’s age to Reagan’s. The Lie Factor of the bar chart, then, is 9.1. As Tufte wrote, a graphic with a Lie Factor larger than 1.05 represents a “substantial distortion.”
To put it another way, the effect shown in the chart is commensurate with data that would have Rubio’s age as 45 and Reagan’s age as 335.

Sehr schön finde ich die im Artikel vorgeschlagene Alternative, die zusätzlich noch viel mehr Information beinhaltet!