Discussed at these sites, among others:
Chart junk-ies
When bar charts go bad
World's Most Expensive Places to Have Sex.
Catch the flaw(s).
Click here to see large.
Go under the fold to see small:


I am the Online Community Manager at PLoS-ONE (Public Library of Science). My job is to try to motivate you to comment on the papers there. My scientific specialty is chronobiology (circadian rhythms and photoperiodism), with additional interests in comparative physiology, animal behavior and evolution. You can contact me at: Coturnix@gmail.com
Buy the 2007 Science Blogging Anthology:
Buy the 2006 Science Blogging Anthology:
Stay abreast of your favorite bloggers' latest and greatest via e-mail, via a daily digest.
Read the archives of my old blogs:
Add Scienceblogs to your Technorati Favorites!
Dictionary of Circadian Physiology
Basic Terms and Concepts in Math and Science
« Now that's old! | Main | Innumeracy and related academic turf-defenses.... »
Category: Reproductive Health • Science Reporting • Sex
Posted on: July 26, 2008 6:58 PM, by Coturnix
Discussed at these sites, among others:
Chart junk-ies
When bar charts go bad
World's Most Expensive Places to Have Sex.
Catch the flaw(s).
Click here to see large.
Go under the fold to see small:


YES! Send me a free issue of Seed.
If I like what I see, I'll receive 5 more issues (6 in all) for just $14.95. That's 50% off the cover price! If I'm not completely satisfied, I'll simply write "cancel" on the invoice and owe nothing. The free issue is mine to keep.
(Non-U.S. subscribers, click here.)
Comments
Well, the first thing I see is that their vertical scale does not start at zero, but at $1, which exaggerates the difference between the data items.
Posted by: Kevin C. | July 26, 2008 7:52 PM
Well, the first thing I see is that their vertical scale does not start at zero, but at $1, which exaggerates the difference between the data items. Beyond that, there are clarity/visual design problems, such as the unnecessary world map background.
Posted by: Kevin C. | July 26, 2008 7:54 PM
Sorry about the double post; the browser had some sort of glitch while I was writing.
Posted by: Kevin C. | July 26, 2008 7:57 PM
There is vertical scaling, for which an explicit scale stands to the left, as well as horizontal scaling, since the columns get wider, with no explicit scale, as well as implied depth scaling, as the implied volumes get larger, again with no explicit scale.
It is flat out wrong to use a three dimensional object to scale a single dimension. Dots at the data points would have been unambiguous.
This may be due to some artsy-fartsy helper wanting to make the graph more decorative. I would not assume anything deliberately sinister like the 'food pyramid' -- where a list of categories is turned into a three dimensional figure with no dimension scaled or labeled.
Posted by: Ken Shabby | July 26, 2008 8:42 PM
Not much is wrong with this chart except:
1: it does not say what it is a chart of. I know whatever it is costs 18 dollars in irleand. It can't be a condom because that would not be possible.
2: the left to right gradation in color is meaningless and/or distracting OR meaningful and unlabeled.
Otherwise it is a perfectly good graph. Of something, but we don't know what.
Posted by: Greg Laden | July 26, 2008 10:20 PM
Oh Gods, the colours... Seattle and Chicago just disappear into the murk.
I don't mind them using condoms as bars, but they really do need to make them the same width ('coz that's what's important). They also need to dump the dollar signs on the y-axis, and give it a label. And start it at zero (as Kevin C.) has pointed out. And not have a labelled tick at every dollar. Use less, and have a slightly larger text size.
The grid lines and map are just clutter: I might not mind the map if the graph was aimed at the general public (e.g. in a newspaper), but the condom colours need to be strong enough to dominate it.
The title is barely informative: I understand the gist of what the graph is about, but more details need to be given somewhere (again, if its in a newspaper, it might be OK in the main text). Mind you, it might be better to have a more accurate title: "Prices of 12 packs of Durex elite condoms around the world" perhaps.
Personally, I would flip it round so that the price is on the x-axis: it makes the names of the places easier to read. It also means that the y-axis can be extended in size, so that the text is larger. The plot should look OK at the size Bora has it, but I think the labels would be difficult to see even with a better quality graphic.
I might also have the US cities in different colours: it depends a bit on what the graph is trying to show, and how it would look.
Finally, why just England? It isn't even an administrative unit! It makes one wonder why Scotland and Wales have been excluded.
Aside from that, it's fine.
Posted by: Bob O'H | July 27, 2008 4:04 AM
I think the horizontal axis ought to compare like-to-like: countries-to-countries or cities-to-cities. (I could see comparing European countries to U.S. states, but not countries-to-cities.) I agree with the complaints about the products not being properly labeled. I had to follow a few links to figure out what the prices represented.
Posted by: John | July 27, 2008 3:26 PM
Prices of 12-packs of Durex elite condoms say nothing about the prices and availability of condoms in general in various places, so the premise of the study is flawed to begin with and the title is thus misleading. Just because this particular brand may be expensive or hard to find, does not mean that locals do not have access to cheap, quality local brands. As for graphics, I agree with most of the above (and the threads I linked to).
Posted by: Coturnix | July 27, 2008 3:44 PM
I'm sticking with my number 1 answer. Because, it does not say what this is a graph of. Like, it does not say that this is a graph of elite condom prices!
(Or am I supposed to recognize these condoms)
Posted by: Greg Laden | July 30, 2008 12:32 PM