Time to Kill the Red-Blue Maps

Now that the election is behind us, I have a little time to ponder one of my hopes for the future in our little corner of the world. I would really like to see the concept of the binary red-blue election map put out to pasture.

The whole concept went viral after the 2000 election as states were colored red or blue depending on who won each state. Since then, the media and our society has taken the idea of red states and blue states and run with it. We may have been on the road to polarization before then but I can’t help but think that such a stark visual has helped cement it in people’s minds.

Last night, as I watched the election coverage on CNN (yes, I saw the hologram), Bill Bennett commented on how after the 2004 election, a county-by-county view showed a sea of red across the nation with small islands of blue (those were the university towns according to him). This clearly showed that the vast majority of the country was conservative, in his mind. As a “blue” resident of a “red county” in a “blue state”, I wanted to throw something at the TV.

Interestingly, later in the same CNN coverage, there was analysis of how Obama was running in Republican counties as compared to Kerry’s 2004 performance. In almost all cases, Obama was outperforming Kerry. This discussion began to peel back the onion a little bit by showing that many “red” counties had a lot of “blue” voters.

This strikes to the heart of problem with the binary maps: they mask the true makeup of our country and potentially contribute to polarization. When the media picks up the red/blue analogy and runs with it, it helps to cement the concept. To be fair, there have been attempts at alternative visualizations. This year, I followed the polling on electoral-vote.com, as I did in 2004. That site makes an attempt to show the gradient a little. There were also a few “shades of purple” type maps after the 2004 election. I think this type of visualization should be the norm rather than the exception.

To that end, I offer up two views of my home state of Maryland after yesterday’s election. I used SharpMap and the CNN county-by-county results to produce them. The first is the standard red/blue binary view which makes it look like Democrats are being outflanked (the gray county is currently tied):

md_red_blue

The second uses purple gradients to attempt to show the mix of Republican vs. Democrat. I used standard RGB color values and applied the percentage vote for each candidate to weight R and B. G was left at zero. So, in a county where McCain got 58% to Obama’s 42%, the RGB value would be (255 * .58, 0, 255 * .42) or 148,0,47. I did not weight the counties according to population. Also, the CNN data did not include votes for third-party candidates so there is no “green” component here.

2008 Maryland Presidential Results in Gradient

There are a variety of ways to do this sort of thing that would be better than the binary red-blue maps we have come to know and loath.

Update: BTW, Princeton is already up with result of the 2008 election: http://www.princeton.edu/~rvdb/JAVA/election2008/ . They’ve been doing this kind of thing for years and have a great time-lapse view. I hope this becomes the prevalent way of looking at it.

Disclaimer: I’m a programmer. I threw these maps together for the purposes of discussion but I’m sure there are numerous better ways to do this. If they ever publish a “Maps That Suck” book, several of my products will probably figure prominently.

  • Chris McClain

    Great point, I think it is due to the focus that is put on the all or nothing (in most cases) electoral college system and the magic number of 270. I think we tend to lose sight of how close some of the races were in states like IN, VA, OH, etc. I think this points illustrates how the way you chose to represent your data spatially can greatly effect the the way it perceived and interpreted.

  • I think you miss the point. The simple red blue maps are aimed at an audience that does not understand basic cartography and statistics. Applying the purple map takes time to think about the meaning, where most people just want to see raw blobs of red and blue.

    Also, the people who create this content are probably not cartographers or even ex geography students. I did feel like John King on CNN was trying to break the simplicity and do some analysis.

    My 2cents worth.

  • I understand there’s something of a know-your-audience issue at play but I don’t think that absolves the media from a responsibility to educate their audience.

    I would agree that many of the authors are not cartographers but that doesn’t have mean they don’t understand presentation.

    I agree with your point about John King. I think he was doing exactly the right thing and was happy to see him doing it.

  • Your gradient approach is nice, and surely preferable for post-election contemplative analysis. But it doesn’t really communicate what most of us want to see at crunch time: On election night, I crave simplicity. Not binary, but I really just want called-red, called-blue, leaning-red, leaning-blue, and open. All the major online maps used this approach, usually offering additional details with interactive layers. I didn’t get to see anything on TV – do they really only show red and blue? They can manage holograms but not purple? Geez.

  • Right. I just threw that up as an example. I actually think electoral-vote.com did a good of conveying the strength of support and keeping it understandable. But yes, they were panning around showing red and blue counties. It can be very misleading, even if they don’t mean it to be. The binary view *may* be appropriate for the states but I think it really starts to send the wrong message at the county level or lower.

  • Would you have felt the same way if the CNN map had been reversed showing most of the country as blue?

    It’s curious how people react to things. Like Vista vs Ubuntu for example. ๐Ÿ™‚

  • Yes.

    Our system works best when people of all political stripes feel engaged and feel like their views are being represented.

    It amazed me that Howard Dean’s 50-state strategy seemed so radical at the time. By engergizing Democrats everywhere, they made states competitive that hadn’t been in years. I think that led to the huge turnout. If the Republicans had a corresponding strategy, we might end up with something resembling political discourse again.

    Despite the negativity heaped upon Hillary Clinton for staying in the primaries until the end, it had the effect making primaries meaningful in states that normally are forgotten after Super Tuesday. I think historians will one day look back at that primary and see that it probably helped the Democratic cause rather than hindered it.

    There will be a lot of history to be mined from this election beyond the obvious.

    So I would not find a map draped in blue any more acceptable. Creating a false impression of Democratic dominance would be just as bad as the false impression of Republican dominance has been.

    As for Vista vs. Ubuntu, give me politics any day! ๐Ÿ™‚

  • Terry

    I disagree. I think we should keep the red/blue maps as a reminder. The use of red and blue on maps of the country didn’t start or even contribute to the divisiveness in the US. Our country has been split in half by a few primary philosophical differences of opinion that have been polarized and exaggerated by the two-party system. It’s important for people to remember that, and to remember who did the splitting. I don’t think that attempts to gloss over the problem by playing with gradations of shading and/or color is such a great idea. There’s already too many people ignoring the problem.

  • markb

    I certainly like seeing a graduated representation given the appropriate scope and intent, but am I the only one that finds interpreting purple to be a little subject to ‘visual’ bias. Meaning it seems a little difficult to pick out the middle case (i.e. 50/50), particularly given widely varying displays (& perspectives). I think that a white or light grey would represent the balance case best, then feather it out from there. Hard to mistake pink from baby blue. This could also work much better for representing additional parties, if ever dominant enough (I know, wishful thinking). I think that Mark Newman’s cartograms and other peoples graduated color products might see some ease of interpretation gains.

    I don’t think that binary encoded thematic maps necessarily polarize people, but you can’t rule out reinforcement; it is an instant transfer of 1000+ words, many of which evoke an instant visceral response.

  • Isn’t this really a question of proportion? The standard red-blue maps work great if we count square miles as votes. But using a state outline like Wyoming (large areas with low population) with equal weight as a highly populated state with a small area (Mass) creates a misleading graphic… when after all it is population (number of electoral votes) that matters, right? I like the “County bubbles” approach used by the NYTimes here:

    http://elections.nytimes.com/2008/results/president/map.html

    You can have your red and blue and map it too.

  • Yes, I agree with you. My map was very quickly thrown together as just an example. I had someone point out the NY Times county bubbles to me on Twitter and I agree that it is effective. My ultimate point is that there are any number of ways to depict this concept that are better than what we usually see and that maybe our community should be taking the lead to improve it.