That's a charming graph, as far as it goes, but the lines don't actually MEAN very much, since there's no correspondence between the y-axis on the left and the y-axis on the right. If there was, it would imply that Americans should expect to live to 100 for the money they're paying. If anything, the way it's scaled suggests that Mexico's non-universal system is the most efficient in the world, since it has a steeper slope even than Japan's and yet people hardly ever have to go see a doctor!
My main point (which I admittedly didn't make very well) was that there's no inherent correspondence between the scale being used on the left and the one being used on the right, which appears to be $1000 vs. 1 year. Like I said, as scaled, the implication is that the US should have an average life expectancy of 100 for the money we're spending, which is absurd. If you stretched out the right scale so that it covered as much y as the left scale - say, so that the 75 year mark was opposite the $1000 mark - it would make Mexico look more impoverished than super-efficient, while still retaining a rather steep slope to show the US's inefficiency, and really highlight countries like Japan, Spain, South Korea and the Czech Republic that seem to be getting a lot of bang for their buck. And leveling out the lines of the many other countries to make the correspondence between amount spend and level of care quite clear. If you stretched that scale, it seems that most of the lines would be flat or nearly so; the particular correspondence of
( ... )
Right - the compression of the scale is slightly arbitrary, though the centering at the averages makes sense. Your argument seems mostly against the centering of the average lifespan and average health care costs. I'm not sure why - that makes a ton of logical sense.
I think however that the scale compression point is valid, though I do feel that no matter the scale you would still get the same impression of the data.
If it was much more compressed it would over exaggerate the disparity between the extremes and if it was stretched more it would exaggerate the differences in the opposite way (compressing things that aren't that close).
I having just played around with it in photoshop a little it is actually one of the better ways to scale it while still keeping the data clear.
Comments 6
Reply
Reply
Reply
I think however that the scale compression point is valid, though I do feel that no matter the scale you would still get the same impression of the data.
If it was much more compressed it would over exaggerate the disparity between the extremes and if it was stretched more it would exaggerate the differences in the opposite way (compressing things that aren't that close).
I having just played around with it in photoshop a little it is actually one of the better ways to scale it while still keeping the data clear.
Reply
Reply
Reply
Leave a comment