The Worst Bar Graph in the World
Please take a minute out of whatever you're doing to make fun of this bar graph.
[Graph from Slate and HHS]This bar graph was posted by Timothy Noah of slate on Friday morning--it is horribly, fundamentally flawed, as any student of basic statistics can tell you. Yet Mr. Noah is so convinced of the power of the graph that he bases an entire two long, impassioned pages of text on it. He urges you to stop what you're doing and look at it.
For those who cannot tell just by looking at the graph what's wrong with it, I'll enumerate the atrocities: 1) the title is not clear--it doesn't explain whether this is a percentage (and if so, what is the sample size?), a per capita dollar amount, or a gross dollar amount being portrayed. 2) the y-axis is completely bare. There is no label (what is being measured?!) and there is, more importantly, no scale. 3) the x-axis is actually the most confusing. It is labeled "2008" while the title claims that it is measuring 1999-2008. Is this just 2008, or is this a composite?
It claims to be from the HHS, and it does appear on the healthreform.gov website. The site itself claims as a source the Kaiser Family Foundation. I could find this exact graph nowhere--in fact, the KFF site seems to be full of well-done, well-organized charts, graphs, and data. I even started looking for the data from which the graph was made, and I only think I found it (Additional Presentation Slides, page 3). This chart makes so much more sense, I can't quite tell if it's the same data.
Nonetheless, independent of your stance on healthcare, this chart should definitely give you pause.
Labels: rants, Stuff that Makes Me Angry
posted by Ben at
9:42 PM
4 Comments:
Ha ha ha! This looks like something I'd use for my presentation debate in my Principles of Health Promotion class...but only reference in passing and hope my teacher didn't pay too much attention to it. (By the way, my group was pro-mandating purchase of health insurance for everyone. And we used rather better figures and charts.)
I was a little disappointed when there wasn't any alt text, a la Dinosaur Comics or Dr. McNinja. "I'm not that great with visuals WHY DO YOU THINK I BECAME A COLUMNIST"
Good idea, Jim!
And yet... the graph isn't all that hard to interpret. It gives the relative average increase in wages (from 1998-2008) vs. the average increase in employees-sponsored premiums over that same period. The precise units aren't given because they aren't relevant to his point; the relevant point is the relative increase, making the units irrelevant. And if there's any residual confusion, you can always read the article.
I see your point, but that relativity is the problem. If it's percentages and not dollars, that could be 1% and 5%. The article does make some good points, but he's got to be more compelling with his visuals if he's going to demand that everyone who reads Slate stop what they're doing and check them out.
Post a Comment
Subscribe to Post Comments [Atom]
Links to this post:
Create a Link
<< Home