« Ranking airlines: no easy task | Main | Unscientific American 2: a review of key concepts »


Rick Wicklin

You say "the rating was rounded up to the nearest 0.5, and presented using the iconic 5-star [motif]." I don't think so. If you look closely at the 2nd and 3rd images, you can see that the stars are filled to two decimal places! For example, the Journalists group (2.57) contains more than a half-star and the NGO group (3.09) has a little splash of pink in the 4th star. A strange choice, but the chart is at least self-consistent.

Tom Hopper

Two more serious problems are the design of the survey (highly biased sampling) and an analysis based on the mean of ranked data. Normally, we should only look at mode, maximum, minimum and the chi squared statistic for such data. If the scales were well constructed so that, to respondents, they were clearly symmetric about a central value, then we could argue that we could compute the median and percentiles.

I realize that this criticism is outside the core message of this blog, but we should be mindful not only of providing meaningful visualizations, but also of visualizing meaningful data. SciAm and Nature seem to have failed rather badly at both.


There is no way in hell that "people" trust scientists most on the topic of evolution. Is this poll seriously biased by the population that was polled?

Oh, it is.

dan l

Also seems to miss that the anti-science crowd have found themselves pet 'scientists' to fabricate facts. We see it with the global warming people and like with what the AFA does with Paul Cameron.

But, like you say, the poll group is a little biased and probably mostly hilljack free.


Rick: wonderful observation. I totally missed the cheeky, bizarre dissection of the stars. Even though my point is not relevant to this chart, I do want to emphasize the importance of the point that we should never print inconsistent things on the same chart.

Tom: What you mentioned is absolutely part of the core message of this blog. Look up my Trifecta checkup. What you are suggesting is that there is mismatch between the data that is collected and the underlying question that is being addressed.

The comments to this entry are closed.

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