Color Blindness

by tristero

Forgive a totally personal rant here, but dammit, I can’t be the only one who finds this rather annoying and, in some circumstances, potentially hazardous. This is a post about being color blind,. That’s not in the political/cultural sense of color as a synonym for race but color as in perception. So hardcore political junkies or those easily bored, just skip this post; you won’t miss anything. For the rest of you:

Like 8% of all males, I have color blindness. Color blindness is a misnomer. I see colors. I just don’t perceive color differences as well as most people. And I don’t have it as severely as my (maternal) grandfather, who literally couldn’t distinguish between red and green traffic lights (I see the green light as whitish and the red light as, well, red), so it’s usually just an amusing party chit-chat kind of a problem.

Except when it comes to reading technical drawings, charts and graphs, such as the ones in science journals. Here’s a fresh example. In order to illustrate a fascinating post on the differences between human and squid eyes, PZ reproduced this disastrous illustration from Nature Magazine. Well, at least it’s disastrous to me:

The caption reads:

Structural comparisons between squid rhodopsin and bovine rhodopsin. Squid rhodopsin is denoted by brown and magenta; bovine rhodopsin by cyan and blue.

Well I’m sure it’s really pretty, but I can’t *&#($^@#$ see the fershlugginah colors!

I deduced that the brown was the dark color; in fact, it kind of looks greenish to me, but nothing else remotely looks like brown, so brown it must be. I can’t see the magenta but I think it’s gotta be the grey color, except it seems to me there are two different grey colors, I’m think, and I have no idea whether they are both shades of magenta or one of them is supposed to be cyan. In isolation cyan, looks bluish to me, but in the drawing, there’s only one blue hue that I can find and since it doesn’t look like the cyan in Wikipedia, I assume it must in fact be blue. Maybe the cyan is what Wikipedia labels “dark cyan” (scroll down in the cyan entry). But that actually looks like grey to me which… looks just like what I think is the magenta.

As you can tell, this drawing hopelessly confused me. Dont get me wrong: I followed PZ’s lucid explanation of the squid/human differences perfectly, and grokked his later explanation of the drawing which seemed not to depend too much on the colors. But if the drawing had subtleties to offer that were not in PZ’s text, I completely missed them.

:You would think that using bold, distinctive colors in scientific visual material would be some kind of no-brainer. And you would be wrong. I have three of Edward Tufte’s beautiful books on the visual design of information and, while he does discuss color, he never mentions, as far as I can recall, choosing color schemes to minimize confusion related to color-blindness. Maybe he takes it into consideration in his new book. But frankly, I doubt it.

This is no aberrant incident, a single over-caffeinated graphic artist digging into the latest iteration of Photoshop. About 12 or 13 years ago, the color schemes at Scientific American got so confusing, I wrote a somewhat humorous letter to complain that I literally couldn’t make head or tail out of some of the charts. I received a very nice reply, with an apology and a promise to take color blindness into account in future graphics. And, miracle of miracles, for several years, the color coding dramatically improved before reverting back to an occasional head scratcher.

Now really, I’m simply a low level science fan; it’s not that big a deal if an amateur can’t tell magenta from cyan. But I know what sheer havoc is caused in my own field – music – before they changed the color coding for stereo wires and cables from red and green to red and white. Anyone who’s hooked up stereo components out of phase has some inkling of the problem – and sometimes, in my studio, I’ve had to make literally hundreds of connections. It was no fun at all trying to puzzle out which wire is red and which is green times 200 plus, especially in less than ideal light.

Given how many male scientists, doctors, engineers, etc, etc, there are (the disproportionate number of males in these professions is another post) and given that a significant fraction of those males must be color blind (8% is not a trivial percentage), I would guess that some very, very serious mistakes could be avoided simply by using less ambiguous color codings. I’m not trying to stifle creativity in the visual design of information. Go ahead, use magenta for all I care. But for heaven’s sakes, please contrast it not with brown and cyan but with eye-melting yellows, bright oranges, and the bluest blues this side of paradise.