Tuesday, April 17, 2012

Tinting the Pregnant Teens

Don't look for the key on this map, and think of what impression you get of the data.

US map showing pregnancy rates with dark green states as high rates and dark blue as low, with shades in between, meeting at neutral gray
Personally, I get the impression that the gray states have the lowest rate of teen pregnancy, which is quite wrong. (Here's the key.)

Key shows rates per 1000 girls
Maybe it's partly because the tiny New England states that have the lowest rates (shown in dark blue) are so.... well... tiny and unnoticeable?

But I think it's more likely a problem with the color coding, which is done as if the map were showing a five-point Likert scale, where the mid-point is neutral, when in actuality one end is bad and the other is good.

To me, the best way to represent that kind of data is one continuous gradation of the same color. My tendency is to have the color that represents the highest rate -- the one that has the most of whatever the map is talking about -- be the darkest, and the tint that represents the lowest rate be the lightest, since it has the least of it:

Same map using dark green to very light green

The New England states are still a bit hard to spot, but isn't that more immediately readable, even without a key?

1 comment:

Pearl said...

I can't actually say that I would've interpreted it as you did without the key, but I can see how a person would!