Winning margin by county

[○∙◄☻¥tπ[╪AV┼cVê└]

[minionofmidas]

...in net votes.

[ATFFL]

This map is quite useless without some sort of key.

[nclib]

New England is also cut off.

[minionofmidas]

New England is also cut off.

Must be due to your browser settings, it's there for me.

Alaska isn't colored in, and the inclusion of Puerto Rico in the map is weird.

[nclib]

New England is also cut off.

Must be due to your browser settings, it's there for me.

Fixed that.

Alaska isn't colored in, and the inclusion of Puerto Rico in the map is weird.

Perhaps no county (actually election district) had a large enough margin to be colored in.

[minionofmidas]

New England is also cut off.

Must be due to your browser settings, it's there for me.

Fixed that.

Alaska isn't colored in, and the inclusion of Puerto Rico in the map is weird.

Perhaps no county (actually election district) had a large enough margin to be colored in.

Anchorage has a third of Alaska's population and went for Bush by 20 points. The map shows Boroughs and Census Areas, not state House Districts.

[nclib]

Lewis, do you (or anyone else) have a list of the largest victories in net votes for Gore and Bush?

[minionofmidas]

No; could do one. I do know the no.1 and 2 on the D side of the ledger -
Los Angeles County 838,575; Cook County 746,005 - but that's only because that much is apparent from the Election Statistics page in the 2000 results in the Atlas.

[○∙◄☻¥tπ[╪AV┼cVê└]

Anyways, this map is nice because it shows who won what geographically without making Bush look like a god for winning all these counties with no one in them. It still does make him look better, seeing as counties he won by a lot are often much bigger than counties Kerry won by a lot. Compare San Bernardino, CA to New York, NY.

[jerusalemcar5]

See the little blue county in the smack dab middle of New York?

THAT'S MY COUNTY!

[Kevinstat]

This map must be for the 2000 election.  Penobscot County (the red one you see in Maine) went for Bush in 2000 but Kerry in 2004.  It went for Clinton in both of his elections, and I doubt this map would be from before 1992.  I can tell that this map uses the modern convention of Red for Republicans and Blue for Democrats by looking at Massachusetts and metro New York.  The big light blue county in south-central West Virginia (Kanawha County, which includes Charleston) went for Gore in 2000 but Bush in 2004.

Looking at the West Virginia results in 2000 and then looking at other states, I think the intensity of the color may be based on the total vote (or perhaps the vote for the candidate carrying that county) rather than the net margin.  There were several counties in West Virginia in 2000 where the net raw vote margin (let alone % margin) for Bush or Gore was greater than Gore's net raw vote margin in Kanawha County.  I can't tell whether it is the total vote or the total vote for the winner that is being compared on this map, but it isn't the net margin either in raw votes or percentage of either the two-party or overall vote).  Until we know exactly what this is a map of, it's pretty useless, and even then it's pretty academic because the total vote in a county (or even the total vote for the winner, rather than the net margin) does not seem useful to anyone but some political stategists or campaign workers, and even they would not be very well served by this map in and of itself.

[Beet]

Anyways, this map is nice because it shows who won what geographically without making Bush look like a god for winning all these counties with no one in them. It still does make him look better, seeing as counties he won by a lot are often much bigger than counties Kerry won by a lot. Compare San Bernardino, CA to New York, NY.

True, but the most accurate map would not use political boundaries at all. This map still gives an undeserved reward for land area of county.

An impressionistically accurate map would ideally be built from data drawn from the smallest possible geographic divisions (precinct level) and the margin in each precinct would be assigned not to a shape but to a dot or centroid representing the geographic center of the precinct. And then the size of the margin would be represented by increasing the number of dots in the vicinity. This has to be done by a computer. It would look something like this. A 3-d image with mountains and valleys would do even better. here is one, but again it makes the error of using political boundaries.

[Kevinstat]

I think I've figured it out: The coloring is based on which candidate carried the county in 2000 (red for Bush, blue for Gore), and the darkness/intensity of the color is based on the number of votes cast for that candidate.  Notice that Hartford County, CT has a darker shading than Fairfield County, CT.  The total vote and total number of ballots cast (including invalid ballots) were both a smidgin higher in Fairfield County than in Hartford County, but Gore received 221,167 votes in Hartford County to only 193,769 in Fairfield County, where Bush was much more competitive (although Gore won Fairfield County by 9% of the vote).  Coupled with what I had already noticed in West Virginia and Maine, it seems like the most votes for the winner in 2000 is the only logical basis for the base coloring and shading of this map.

[Alcon]

Anyways, this map is nice because it shows who won what geographically without making Bush look like a god for winning all these counties with no one in them. It still does make him look better, seeing as counties he won by a lot are often much bigger than counties Kerry won by a lot. Compare San Bernardino, CA to New York, NY.

True, but the most accurate map would not use political boundaries at all. This map still gives an undeserved reward for land area of county.

An impressionistically accurate map would ideally be built from data drawn from the smallest possible geographic divisions (precinct level) and the margin in each precinct would be assigned not to a shape but to a dot or centroid representing the geographic center of the precinct. And then the size of the margin would be represented by increasing the number of dots in the vicinity. This has to be done by a computer. It would look something like this. A 3-d image with mountains and valleys would do even better. here is one, but again it makes the error of using political boundaries.

And the most representative of all (although not geographically - but warping precincts by votes would ruin that anyway) would just be a pie chart of the popular vote results.

[Beet]

Anyways, this map is nice because it shows who won what geographically without making Bush look like a god for winning all these counties with no one in them. It still does make him look better, seeing as counties he won by a lot are often much bigger than counties Kerry won by a lot. Compare San Bernardino, CA to New York, NY.

True, but the most accurate map would not use political boundaries at all. This map still gives an undeserved reward for land area of county.

An impressionistically accurate map would ideally be built from data drawn from the smallest possible geographic divisions (precinct level) and the margin in each precinct would be assigned not to a shape but to a dot or centroid representing the geographic center of the precinct. And then the size of the margin would be represented by increasing the number of dots in the vicinity. This has to be done by a computer. It would look something like this. A 3-d image with mountains and valleys would do even better. here is one, but again it makes the error of using political boundaries.

And the most representative of all (although not geographically - but warping precincts by votes would ruin that anyway) would just be a pie chart of the popular vote results.

What do you mean by "warping precincts by votes"?

[minionofmidas]

Actually, the most representative of all would not go by margins, but simply (simply? ahem) have one dot, in his or her party's colour, for every voter. So if one precint had 250 Dems, 212 Reps, and 7 Greens, it's got a grand total of 469 dots in three different colors.