Here’s an example of an outstandingly misleading data-graphic, appearing in this week’s LondonStudent freepaper.
It attempts to show the disparity of bar staff pay across London universities, but:
- The “empty” pint glass does not correspond to £0.00. To the casual observer it makes it look like CSSD work for free, until the figures are noticed. In fact, the text of the article mentions that, at two other London university bars, the staff do in fact work for free (or for beer – ouch) – but these are not shown on the graphic.
- The graphic is a 2D (i.e. print) representation of a 3D object (a pint glass, tilted slightly towards the viewer) but the scale appears to vary in 1D – the values form a straight line across the “glass”. Hence the graphic has a large “Lie Factor”, the concept discussed in detail in E. Tufte’s totemic book The Visual Display of Quantitative Information (p57 for those making notes!)
- LSE’s amount bizarrely isn’t represented at all in the graphic, but appears in a text box above it.
- The numbers are on their side, even though there is plenty of room to show them horizontally – making the real values harder to read, so the reader concentrates on the misleading graphic representation instead.
- The actual levels don’t bear any resemblance to the values – the ordering is correct but the relative value differences don’t correspond to the “beer levels”. For example, the drop between £5.90 and £5.95 is larger than the £5.95 to £6.25 drop.
- Why’s beer being used to represent pay anyway?