15 replies on “Happy Tax Day! (Or, Who is Teabagging?)”

  3. That chart sucks balls for trying to figure out the relative tax rates. Why even bother including the tax information? It’s only useful when looking at the overall box size.

  4. Indeed, there is data showing viewers are terrible at estimating relative sizes of shaded areas on charts. This is why pie charts are so bad, even though everyone like them. Edward Tufte has lots of alternatives to this kind of stuff. Probably the first one might be to try just printing the numbers in a simple table; it’s only about 15 numbers, right?

    At least it’s not in 3D. Thank god.

  5. @2 – Hear, hear.

    The socialist in me would really like to see the tax rate go up as an exponential curve where the middle of the curve is set at a multiple of the poverty line, say 200x. The poor would pay next to nothing, and the assholes at the top would be effectively prevented from taking over the world.

  7. I tend to really like Tufte’s ideas about the visual display of quantitative data. This is one area where I tend to disagree.

    While area is a poor way to depict small changes in magnitude, for differences on this scale, it works better than numbers in a table for most people’s comprehension or a line chart.

    38,200,000 vs 78,100 lacks the impact of the huge square next to the small one, in my opinion. Partly I think this goes back to how counting and numerical comparisons are handled in our brains. Up to three, a specialized area of the brain handles the task. Above that, all math operations are through the speech processing centers.

    And I’m intentionally trying to diminish the differences in marginal tax rate. The point is, even with the (ever so slightly) progressive marginal income tax rate in the US, the net income differences are still dramatic. Much of the whining going on today involves the black bar for the CEOs moving a little bit higher into the green. In any honest way, it’s a meaningless change….

    I don’t like the use of whitespace in this figure. With more time, I could do a better job. Too busy right now.

  9. My point is that all you’re communicating is that the difference in magnitude is “big.” You do actually communicate the same fact with a table. The problem is that what you’ve communicated isn’t all that accurate; the viewers are going to each give you wildly varying answers if you were to ask them how much larger the CEO’s boxes are. A hundred times, fifty times, two hundred times. The average of their answers will probably be way off and the variance is going to be larger than other, better graphs.

    It’s because some viewers will tend to say: that box is twice as tall as the next box, while others will say: that box is four times the area of the one next to it. Some will sort of dither between height and area. Add a third dimension and it becomes a disaster

    You could have careful written instructions telling the viewer to estimate the area they see and ignore the height, but a graph that relies on a lot of written instructions to work is pretty sub-optimal, especially since you can communicate 15 data points in the same number of words as the instructions for reading the graph.

    The argument about visual vs math/language processing would favor a simple bar chart over a table, but it doesn’t give any reason why you have to use area instead of distance.

    I’d have hope for a scatter plot too, perhaps but I’d have to see it. But area? Iffy. Volume? Hopeless.

  10. elechos: I thought about making a line chart and mapping this as distance. The difference is SO big, it becomes silly. Switching to a logarithmic scale helps, but makes the figure less intuitive for the non-mathematically literate.

    With area, I get to run things through a square root. I like using symmetry (squares) to make the area to scale transition for the viewer easier. I thought about gridding the areas (so you can see how many teacher’s net income fits in the CEO box) but that adds a lot of visual cruft.

  11. Since we’re their boss now for their hard work I suggest them taking in half of what we pay teachers until they give us our money back. You don’t like it? No one’s forcing you to be here today.

  13. The grid is actually not a bad idea. A field of identical symbols, like dots or dollar signs or money bags or little old ladies with their pockets turned out can be effective too. In essence you are communicating area with a certain number of icons, but it visually cues you to not look at the width or height but to focus on how many icons there are, which is essentially a measure of the area.

  14. weren’t we just making fun of senators for their use of silly charts to demonstrate how big our debt will be? something about stacks of $100 bills to the moon?

  15. Actually, “big” is sufficient to make your case here. The correct response to this factoid is “holy crap, that’s a lot of green”. Precision is your enemy here; “78.3 times as much” simply does not have the impact of a big visual space.

Comments are closed.