« Bumps charts and NYT | Main | The dots don't connect »



Why is that a problem? The sampling error should very similar regardless of the size of the country, shouldn't it?


Yes, the tick mark at the bottom is confusing. Especially hard to grok in that small image!

I'm not so sure about your point about scale. I guess you could adjust the thickness of the bars according to population and not lose too much readability. But is the point of the chart to compare national attitudes or the number of people behind each opinion?

John S.

Which special partner, Canada or Mexico? I think those two countries would have very different charts. It would indeed be interesting to put the U.S. in its geographical context and compare with other countries of the Americas.

Antonio Cavedoni

Wrand: you could also rotate the map by 90 degrees, have the 50% axis in the vertical middle and have the countries as bars hanging from that 50% mark. That way you could display both the yes/no for each country, and the bar length would signal the population.


Which dataset is the source? It is irritating not to mention a proper survey name.

Is the data taken from one survey or did they ask the same question in a European survey and an American survey and an Japanese survey?
If it is one survey, the composition of part-taking countries is odd: the US and Japan in a European survey?? If we have data from different surveys, the it is less comparable: several isssues besides translation and response rates, e.g. the context and position of the question in the questionnaire conquer the credibility of the data.


You can compare/contrast this with the chart on the original paper (http://www.sciencemag.org/cgi/reprint/313/5788/765.pdf)

I agree with the suggestion that the tick mark at the bottom was a particularly bad idea. I would have also drawn a subtle faint line all the way down tracking the 50% point. I disagree with the comment that the population size is of much relevance to this chart, but agree that the peer selection seems odd (the paper explains this somewhat and has some additional data source info online).

Two questions. One, the NYT version uses dotted lines between clusters of five countries. Presumably this is to facilitate reading. What do others think of that?

Second, the chart on the original paper uses "red" to indicate false and "blue" to indicate "true". While I find myself firmly in the camp of those who believe in evolution, I do wonder if there is some subtle association of red with "bad" or "wrong" that convey an editorial point of view with the choice of colors...


Totally agree on the uselessness of clusters of 5, as well as having a faint 50% line running down the chart.

As for big vs small countries, I'm just making a point about practical statistics

And John - don't forget the U.S.'s special partner in Europe!

derek c

Kaiser, could you stop being mysterious and just say which country you're referring to?

John S.

Ah, that special partner. Ironic that it is not on this list, given that it is the very country that gave the world the idea that humans are descended from earlier species of animals.


Look, if y'all are referring to Britain, it's the sixth one down on the list, between Japan and Norway!


Derek - it's embarrassing to admit I was referring to the Brits and indeed it is the 6th country in the list. Mea culpa! And I thought I scanned it several times to make sure...

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