July 28, 2003

False Advertising

You've seen this map a bunch of times, yes? States colored bright red or blue by which candidate they cast their electoral votes for in 2000. It shows three islands of "blue" Democrats--the west coast, the upper midwest, and the northeast--swimming in a "red" Republican sea. This map has given rise to a huge amount of bad political commentary about the division of our nation, and the large cultural and sociological gap between the "red" and the "blue."

But suppose you colored states not by which candidate their electoral votes were cast for, but by who their citizens voted for. Suppose you mixed X% Democratic blue with Y% Republican red. What would it look like? It would look like this:

No islands. No sharp divisions. No yawning cultural and sociological gap--just slightly varying shades of purple, mixed blue and red. Only seven states in 2000 had a Republican presidential vote share more than sixty percent. Only five states in 2000 had a Republican presidential vote share less than forty percent.

The first map is false advertising--the combination of our quirky system of electing a president with the tendentious arguments political commentators interested in maximizing perceived differences.

The second map is reality.

E Pluribus Unum.


UPDATE: The highly intelligent and articulate Outside the Beltway gives the county-by-county map. Alas, it's just red-blue--Bush win-Gore win. It's not purple. I don't have the numbers (or the mapmaking software) to do the percentages-by-county graphic. Does anybody?


UPDATE: Sumana Harihareswara informs us that Leonard Richardson did it first, back last February. Yet more evidence that the most important skills I can teach my children are the skills of librarianship...

Posted by DeLong at July 28, 2003 10:51 PM | TrackBack

Comments

The Electoral College system needs to be abolished. It's downright silly to garner the most votes and still be able to lose the election.

Posted by: TIN Soldier on July 29, 2003 08:11 AM

Great way of making the point! This map deserves wide currency.

Posted by: Ted on July 29, 2003 08:27 AM

ok, that's not going to happen.

Also, I don't think that's the point of this entry. We are more alike than different.

It's also interesting to note the color difference between the Dakotas and south to Texas and Minnesota and South to Louisiana.

Posted by: NohoMissives on July 29, 2003 08:31 AM

Well, no. You've increased the precision of the measurement of each sample, but you've left the sample size unchanged (and an arbitrary and greatly varying sample size it is). If you make the sample size smaller (but still irregular) by using as your unit the political "county", then the greater variance would start to reappear.

The biggest divide is between urban and rural (and to a lesser extent, urban and suburban) and in those you will see redder reds and bluer blues. A better map would look at samples of equal-population, geographically contiguous areas. Say, of 100,000. There again, I think you would see red to blue varying with population density--but population density varies widely in the US and has a very uneven distribution.

Secondly, of course, the individual votes themselves are very low-resolution samples of the voters' political beliefs. Looking at local election votes, or all votes over a span of time, could show a greater or lesser disparity between voters' ideologies than your national election map above.

Posted by: Keith M Ellis on July 29, 2003 08:57 AM

A number of alternative representations are available here (http://www.geog.ucsb.edu/~sara/html/mapping/election/map.html) courtesy of Sara Irina Fabrikant & Cal Santa Barbara' Geography Dept.

Posted by: Davis X. Machina on July 29, 2003 09:11 AM

"The second map is reality." - No way, a map is a representation of reality. If it is "correct" it parsimoneously carries information about reality - no more, no less. The same thing holds for applied language, like texts.

Posted by: Mats on July 29, 2003 09:41 AM

By county:

http://images.usatoday.com/news/electmap.jpg

Posted by: George Zachar on July 29, 2003 09:50 AM

one thing I wonder about, however, is the variations in the white vote in the various states. As a Democrat, the one thing I wonder about is whether our whole national politics will become like the South writ large, where the Republican party becomes where white people *belong* and only "oddballs" and antedeluvian white people (especially men) vote Democrat.

Alarmist, but the Republican share of the white vote has been increasing in recent elections, and Bush's high approval numbers are not because he has been making inroads with minority voters, but because a very large proportion of the white population feels compelled, I daresay, to approve of him.

Posted by: roublen vesseau on July 29, 2003 09:56 AM

Brad: Thanks for the plug. I would be interested in seeing a "purple" version of the county-by-county as well.

Davis: The UCSB maps are interesting but, as they were all based on the election minus Florida, a bit odd.

Mats: I think your point is worth keeping in mind. Maps are useful ways to organize information, but they can be misleading. Indeed, one could color the entire US either red (Bush won the election) or blue (Gore got more votes); that wouldn't be very useful, even though either would be "correct."

Posted by: James Joyner on July 29, 2003 10:05 AM

Population density, presence of a college, and high percentage of govt employees (state capitals) are all strong predictors of Gore voters, according to the county map linked above.

Posted by: George Zachar on July 29, 2003 10:18 AM

>our whole national politics will become like the South writ large, where the Republican party becomes where white people *belong*

http://www.bear-left.com/original/2001/0520gop.html

Bush got 54% of the white vote. Gore got 90% of the black vote.

Racial voting patterns are clearly powerful. But not among whites.

Posted by: Bucky Dent on July 29, 2003 10:25 AM

You're missing some islands. Not just Hawai'i, but the island of hard-core Republican support in the Northern Interior West (UT, ID, WY). Even in the purple map, that's an island.

And when you look at the counties, you will see very dramatic island effects. The Mississippi corridor is really interesting. So is the distribution on the West coast; it's not the west coast states that go Dem but only the actual coastal counties in those states.

As for black and white, it's still the most interesting divide in America and will be until latinos start voting in significant numbers. Note that Mississippi is about 40% black and blakc people vote 90% Democrat there. That means white people in Mississippi are voting about 80% Republican. Now there's a divide to see.

Posted by: newt on July 29, 2003 10:35 AM

"Racial voting patterns are clearly powerful. But not among whites."

http://zebco.blogspot.com/2003_04_01_zebco_archive.html#93369826

If you work backwards from the voting totals, you get the interesting conclusion that "93% of Republican voters are white", to go wotih "54% of whites are Republicans." Depends on which end you come to it from, I guess.....

Posted by: Jason McCullough on July 29, 2003 10:41 AM

Bucky Dent wrote, "Bush got 54% of the white vote. Gore got 90% of the black vote.

"Racial voting patterns are clearly powerful. But not among whites."

Huh? The fact that a majority of whites voted for a complete fraud like Bush is a very powerful pattern.

Posted by: Stephen J Fromm on July 29, 2003 11:48 AM

Jason: If you work backwards from the voting totals, you get the interesting conclusion that "93% of Republican voters are white", to go wotih "54% of whites are Republicans." Depends on which end you come to it from, I guess.....

Heh. Well, considering that 75% of the population is white and that 90%-plus of the 13% of the population that's black votes Democrat, that's not too bad. Then, when you take in to consideration that whites vote in much higher proportion than their percentage of the population would indicate. . .

Posted by: James Joyner on July 29, 2003 12:05 PM

> The fact that a majority of whites voted for a complete fraud like Bush is a very powerful pattern.

This neatly parallels the racist cant about blacks' overwhelming support for Clinton.

How kind of you to proove that mindless racists exist on both sides of the aisle.

Posted by: Bucky Dent on July 29, 2003 12:18 PM

http://www.crummy.com/articles/purple/ : "How I did it: I used CNN's election results. Using the GIMP I made a 100-pixel-wide gradient from the "red" color on Salon's map to the "blue" color on Salon's map. That was Gore's gradient, the one shown on the legend. I flipped it around to get Bush's gradient (just for convenience, since the election results put the winner of the race in a state first on the list). I colored each state according to how many percentage points the winning candidate won in that state. For example, Bush got 58% of the vote in Montana, so I picked the 58th pixel from the left in Bush's gradient and used that color to color Montana.

For the nine states in which the winning candidate won only a plurality, I mapped the difference between the two candidates' percentages (usually a whopping 0%) onto the winner's gradient. In retrospect, I probably should have done it this way for every state to compensate for the effect of third parties. Not that it would have mattered much."

Posted by: Sumana on July 29, 2003 12:31 PM

Ogged (www.unfogged.com) links a pre-existing map that does pretty much what we all want. It doesn't do purple, but it shows 5 shades each of read and blue.

Posted by: James Joyner on July 29, 2003 01:33 PM

Bucky Dent wrote, "This neatly parallels the racist cant about blacks' overwhelming support for Clinton."

Really? How is it any more racist than "Racial voting patterns are clearly powerful. But not among whites."

My point is that 54% of whites voting for a total and utter fraud is every bit as compelling as 90% of blacks voting for a Democrat.

Posted by: Stephen J Fromm on July 29, 2003 01:51 PM

Bucky Dent wrote, "This neatly parallels the racist cant about blacks' overwhelming support for Clinton."

Really? How is it any more racist than "Racial voting patterns are clearly powerful. But not among whites."

My point is that 54% of whites voting for a total and utter fraud is every bit as compelling as 90% of blacks voting for a Democrat.

Posted by: Stephen J Fromm on July 29, 2003 01:53 PM

The server sees

Posted by: Chris Marcil on July 29, 2003 02:11 PM

The most important skills I can teach my children are the skills of librarianship...

Hey, don't forget, 6 months in the lab can save you three hours in the library.

Posted by: silicon_retina on July 29, 2003 02:26 PM

Bucky Dent wrote, "This neatly parallels the racist cant about blacks' overwhelming support for Clinton."

Another reason your position is nonsensical is the example of Louisiana, where David Duke won a majority of the white vote in 3 statewide elections. I'll go ahead and presume it was nowhere near 90%. Your logic would dictate that this voting pattern is not remarkable, simply because Duke (again, presumably) didn't receive a supermajority.

Now, Bush is no Duke, but...he's Bush.

Posted by: Stephen J Fromm on July 29, 2003 03:50 PM

I see no racism in blacks voting for the less racist party.

DSW

Posted by: Antoni Jaume on July 30, 2003 10:34 AM

"Heh. Well, considering that 75% of the population is white and that 90%-plus of the 13% of the population that's black votes Democrat, that's not too bad. Then, when you take in to consideration that whites vote in much higher proportion than their percentage of the population would indicate...."

Oh, of course, but to first order it's probably right. That the GOP actually has so few non-whites explains a lot in terms of, say, how Lott could get away with it for so many years.

Posted by: Jason McCullough on July 30, 2003 10:56 AM
Post a comment