Successive/subsequent/coherent/consecutive state trends

[eric82oslo]

I've measured each state's successive presidential vote trend relative to the national average. For a lot of states, their relative trend changed from 2008 till 2012. For instance states which tilted heavily Democratic in 2008, tilted somewhat back towards Republicans again in 2012. And the opposite for states tilting Republican in 2008. However, for a lot of other states, their relative trend continued for the past 2 or 3 elections, and for some states it has continued for an even longer period than that. This thread deals with the longest-running partisan trend of each state (+ D.C.) which still is valid.

I painted Democratic-trending states in blue and Republican-trending states in red. States which hardly have moved, were painted in green. I consider the important "walls of changes" to be 1%, 3%, 5%, 10% and 20% away from the national average. Less than 1% means no clear trend at all. Between 1% and 3% means a very slight trend, but which might be due to random changes/noise as well. Between 3% and 5% is more of a trend, but doesn't necessarily have to be a long term trend. Between 5% and 10% is clearly a trend, but still not a very strong one yet. Between 10% and 20% is a very strong trend. And everything about 20% is an extremely strong partisan-leaning trend.

So this is the consecutive trend of each state as it stands right now, after the 2012 election.

Hawaii – D +27.65%
Vermont – D +22.32%

California – D +14.89%
Alaska – D +10.96%
Virginia – D +10.49%

North Carolina – D +8.18%
Washington – D +6.99%
Maryland – D +6.78%
New Jersey – D +5.63%

Louisiana – D +4.83%
New York – D +4.73%
Mississippi – D +4.08%
Rhode Island – D +3.06%

Arizona – D +2.87%
Alabama – D +2.8%
Georgia – D +2.46%
South Carolina – D +1.92%
Ohio – D +1.8%
Florida – D +1.47%
Maine – D +1.38%
Oklahoma – D +1.16%

Minnesota – D +0.86%
Massachusetts – D +0.74%

Colorado – R +0.17%
Iowa – R +0.31%
Texas – R +0.6%
Oregon – R +0.85%

New Mexico – R +1.57%
Connecticut – R +1.63%
New Hampshire – R +2.11%
Nevada – R +2.4%
D.C. – R +2.53%
Delaware – R +2.96%

Idaho – R +3.07%
Kansas – R +3.35%
Pennsylvania – R +3.43%
Nebraska – R +3.44%
Wisconsin – R +3.55%
Michigan – R +3.56%
Illinois – R +4.86%

South Dakota – R +6.2%
Wyoming – R +7.35%
North Dakota – R +7.58%
Indiana – R +7.82%
Montana – R +8.08%

Utah – R +16.45%
Missouri – R +17.83%
Tennessee – R +19.88%

Kentucky – R +24.2%
West Virginia – R +38.08%
Arkansas – R +39.71%


Clearly, the 20 states which have trended the most strongly in the past few election cycles (going all the way back to 1992 for Arkansas, West Virginia, Kentucky & Missouri) are:

1. Arkansas - becoming 39.71% more Republican
2. West Virginia - becoming 38.08% more Republican
3. Hawaii - becoming 27.65% more Democratic
4. Kentucky - becoming 24.2% more Republican
5. Vermont - becoming 22.32% more Democratic
6. Tennessee - becoming 19.88% more Republican
7. Missouri - becoming 17.83% more Republican
8. Utah - becoming 16.45% more Republican
9. California - becoming 14.89% more Democratic
10. Alaska - beoming 10.96% more Democratic

11. Virginia - becoming 10.49% more Democratic
12. North Carolina - becoming 8.18% more Democratic
13. Montana - becoming 8.08% more Republican
14. Indiana - becoming 7.82% more Republican
15. North Dakota - becoming 7.58% more Republican
16. Wyoming - becoming 7.35% more Republican
17. Washington - becoming 6.99% more Democratic
18. Maryland - becoming 6.78% more Democratic
19. South Dakota - becoming 6.2% more Republican
20. New Jersey - becoming 5.63% more Democratic

New York would have been on the top 20 list above too, had it not been for their temporary GOP-trendline following the 9/11-attacks.

The states which have hardly trended at all, for the past two elections or so, are Colorado, Iowa, Texas, Massachusetts, Oregon, Minnesota, Oklahoma, Maine, Florida, New Mexico, Connecticut, Ohio, South Carolina, New Hampshire and Nevada.

Thoughts? Any surprises?

[eric82oslo]

That gives us this trendline map (with reverse colors used obviously).



Pink and light blue means literally no change/trend.

[opebo]

So it is Pennsylvania, Wisconsin, and Michigan which are the big concern, and Virgina and North Carolina which are the most positive stories.

[eric82oslo]

So it is Pennsylvania, Wisconsin, and Michigan which are the big concern, and Virgina and North Carolina which are the most positive stories.

Perhaps. And these stats should also worry those who still live in the illusion that Missouri is a swing or even battleground state. Few states have trended more heavily the past 20 years. And nothing seems to be stopping the trend, even for a short while. Which is curious, since Missouri used to be the bellweather battleground state next to Ohio. However, there's the simple fact that Missouri belongs to the Bible Belt, and that it's also somewhat whiter than the average state, and experiencing less of a diversifying demographic change than most of the country.

[eric82oslo]

Even more worrying for Republicans, these are the trendline changes we've seen in each state (again relative to the popular vote) since the presidential year 2000.



31 states + D.C. have moved towards Democrats, while only 19 states have moved in the other direction, towards Republicans. Although some fairly big states like Florida, New York, Pennsylvania, Massachusetts, Tennessee and Arizona are amongst those 18.

Almost no change (meaning less than 1%) during this 12 year period in New York and Georgia.

As is clear from both maps, the region which is most clearly tilting one way - in this case tilting towards the Republicans - is the "Upper South" or the Appalachian region. Meaning the states of West Virginia, Kentucky, Tennessee, Missouri and Arkansas. Sometimes the states of Kansas, Oklahoma, Louisiana, Alabama and (Western) Pennsylvania can be attached to the rapid changes of this region as well (even North-western Florida).

[Space7]

I see that I, a Canadian, and you, a Scandinavian, are the ones doing the data crunching for the American trends.

Obviously though, the first chart shows a Democratic trend for states like Louisiana, which is sort of misleading because if you disregard what seemed to be an upset in 2012, Louisiana has a Republican consecutive trend of over R+20.

You know what would be useful? A weighted map of the State trends, so you would value the 2012 trend at 50% of the general trend, 2008 at 25%, 2004 at 12.5%, 200 at 6.25%, etc., or something like that. Then make a map of it. That would kind of be useful for negating upsets.

It would also be nice to find a way to measure state elasticity, because that is an annoying factor when trying to figure out where a state is trending.

For example, low elasticity states will remain more stable despite whoever is winning the election, and as a result they will appear to move towards their favored candidate in an election where their favored party lost, and move away from their favored candidate if that party won the election.

High-elasticity states are the opposite, they will move towards their favored party if their party won, and they will move away if it lost.

Good work though, it's funny how the northern Rockies are blue in the first map and red in the second.

[eric82oslo]

You know what would be useful? A weighted map of the State trends, so you would value the 2012 trend at 50% of the general trend, 2008 at 25%, 2004 at 12.5%, 200 at 6.25%, etc., or something like that. Then make a map of it. That would kind of be useful for negating upsets.

Good idea actually, and so I did. Just tweaked your idea slightly, or rather the weighting of each election cycle. I decided to go back 20 years, starting with the 1992 election, since that was the year when several prominent political trends started, at least in the Appalachian region. Since then we've had 5 cycles. My weighting goes like this:

1992-1996: 1 weight
1996-2000: 2 weights (double the significance)
2000-2004: 3 weights (triple the significance)
2004-2008: 4 weights (quadruple the significance)
2008-2012: 5 weights

Finally adding them all together and then divide by 3 (since there are 15 weights added - meaning that the most recent cycle adds 33% to the total - and 5 cycles [15:5=3]).

So here is my Ultimate Trendline List of all 50 states + D.C. (1992-2012):


Hawaii: D +34.88%

Vermont: D +24.35%

Maryland: D +15.18%
California: D +14.54%

Virginia: D +11.91%
Alaska: D +11.64%
Colorado: D +9.42%
New York: D +8.72%
D.C.: D +8.3%

Nevada: D +7.84%
Delaware: D +7.07%
North Carolina: D +7.06%
New Mexico: D +6.86%
Washington: D +6.85%
Maine: D +5.31%

New Jersey: D +4.79%
Oregon: D +4.19%
Connecticut: D +4.01%
Rhode Island: D +2.96%
Ohio: D +2.62%
New Hampshire: D +2.38%

Wisconsin: D +1.88%
Illinois: D +1.74%
Nebraska: D +1.41%
South Carolina: D +1.32%
Iowa: D +1.21%
Florida: D +0.97%
Michigan: D +0.66%
Mississippi: D +0.48%


Indiana: R +1.31%
Georgia: R +1.64%
Minnesota: R +1.78%

Pennsylvania: R +2.72%
Massachusetts: R +4.79%
Texas: R +4.81%

North Dakota: R +5.19%
Montana: R +6.68%
Arizona: R +6.73%

Kansas: R +8%
Idaho: R +8.07%
South Dakota: R +10.13%
Alabama: R +11.66%

Missouri: R +17.52%

Louisiana: R +20.6%
Oklahoma: R +22.28%
Kentucky: R +23.66%
Wyoming: R +24.44%
Tennessee: R +25.25%
Utah: R +29.59%

Arkansas: R +39.32%
West Virginia: R +43.05%


Now let's create a trend map based on the groupings above. Meaning the following:

0-2%: Partisan trend of 30%
2-5%: Partisan trend of 40%
5-8%: Partisan trend of 50%
8-12%: Partisan trend of 60%
12-20%: Partisan trend of 70%
20-30%: Partisan trend of 80%
Above 30%: Partisan trend of 90%

And here is the Ultimate Trend Map (1992-2012):

This gives us a trend map of 329-209 EVs, favoring Democrats.

We see that more states are trending Democratic (29) than Republican (22). Yet those which are trending Republican, are trending much harder than the average trend for Democratic-leaning states.

How many of these states are likely to change partisan trends over the next election or two? I would say at least three; Texas, Arizona and Georgia. Probably Massachusetts too, although that is not certain. A few of those which right now are slightly trending Democratic might also turn around; Ohio, Nebraska, Illinois, Wisconsin, Michigan, Mississippi, maybe Iowa.

[Space7]

You know what would be useful? A weighted map of the State trends, so you would value the 2012 trend at 50% of the general trend, 2008 at 25%, 2004 at 12.5%, 200 at 6.25%, etc., or something like that. Then make a map of it. That would kind of be useful for negating upsets.

Good idea actually, and so I did. Just tweaked your idea slightly, or rather the weighting of each election cycle. I decided to go back 20 years, starting with the 1992 election, since that was the year when several prominent political trends started, at least in the Appalachian region. Since then we've had 5 cycles. My weighting goes like this:

1992-1996: 1 weight
1996-2000: 2 weights (double the significance)
2000-2004: 3 weights (triple the significance)
2004-2008: 4 weights (quadruple the significance)
2008-2012: 5 weights

Finally adding them all together and then divide by 3 (since there are 15 weights added - meaning that the most recent cycle adds 33% to the total - and 5 cycles [15:5=3]).

So here is my Ultimate Trendline List of all 50 states + D.C. (1992-2012):

Now let's create a trend map based on the groupings above. Meaning the following:

0-2%: Partisan trend of 30%
2-5%: Partisan trend of 40%
5-8%: Partisan trend of 50%
8-12%: Partisan trend of 60%
12-20%: Partisan trend of 70%
20-30%: Partisan trend of 80%
Above 30%: Partisan trend of 90%

And here is the Ultimate Trend Map (1992-2012):

This gives us a trend map of 329-209 EVs, favoring Democrats.

We see that more states are trending Democratic (29) than Republican (22). Yet those which are trending Republican, are trending much harder than the average trend for Democratic-leaning states.

How many of these states are likely to change partisan trends over the next election or two? I would say at least three; Texas, Arizona and Georgia. Probably Massachusetts too, although that is not certain. A few of those which right now are slightly trending Democratic might also turn around; Ohio, Nebraska, Illinois, Wisconsin, Michigan, Mississippi, maybe Iowa.

Excellent map!

Almost deserves a thread to itself.

This map would probably be cause for worry for the Republicans, even if some trends turn around. The only thing on the map which might concern the Democrats is Pennsylvania.

A bunch of weak Democratic trends in important swing states (Florida, Virginia, Ohio, North Carolina, Colorado, Nevada, Iowa, and North Carolina) is infinitely more useful than a couple extremely strong trends in safe states (e.g. Kentucky, Tennessee). You'd think more Republicans would support Proportional Representation, judging by the way things are going here.

[eric82oslo]

Excellent map!

Almost deserves a thread to itself.

Thanks for your kind words.



Based on the long term trends (1992-2012), this is about how partisan we should expect each state to be for the presidential election in 2016. Yet some states like Hawaii, Utah and Wyoming, as well as possibly some Appalachian states, might very well have maxed out in 2012 already. Plus the fact that Hillary will probably be in it, has the potential to mess up things a lot.

Anyways, trendlines expect the 2016 election to turn out approximately like this:


D.C.: D +81.4%

Hawaii: D +45.8%

Vermont: D +36.6%

New York: D +26.1%
Maryland: D +25.3%
Rhode Island: D +24.2%
California: D +22.2%

Massachusetts: D +18.3%
Delaware: D +16.2%
New Jersey: D +14.9%
Connecticut: D +14.3%
Illinois: D +13.4%
Maine: D +12.5%
Washington: D +12.4%

Oregon: D +9.1%
New Mexico: D +7.7%
Michigan: D +5.8%
Nevada: D +4.4%
Wisconsin: D +3.5%
Minnesota: D +3.5%
Colorado: D +3.4%
Virginia: D +2.4%
New Hampshire: D +2.2%
Iowa: D +2.2%
Pennsylvania: D +1.0%


Ohio: R +0.4%
Florida: R +2.8%
North Carolina: R +4.5%

Georgia: R +12.0%
South Carolina: R +14.1%
Arizona: R +14.3%
Indiana: R +14.3%
Alaska: R +15.5%
Mississippi: R +16.3%
Missouri: R +16.7%
Montana: R +18.8%

Texas: R +20.6%
South Dakota: R +23.9%
North Dakota: R +24.5%
Louisiana: R +25.2%
Nebraska: R +25.4%
Kansas: R +27.2%
Alabama: R +28.4%
Tennessee: R +29.3%

Kentucky: R +31.3%
Arkansas: R +35.4%
Idaho: R +37.4%
West Virginia: R +39.2%

Oklahoma: R +41.9%
Wyoming: R +49.6%

Utah: R +57.8%


If these predictions are right, the two most important battleground states - or more precisely bellwethers - in 2016 might very well be Ohio and Pennsylvania. Other important swing states will include Iowa, New Hampshire, Virginia, Florida, Colorado, Minnesota, Wisconsin, Nevada and North Carolina. The usual suspects in other words. Michigan might also be close, and round out the "Top 12". More long shots will include New Mexico, Oregon, Georgia, Washington, Maine, Illinois, South Carolina, Arizona, Indiana, Connecticut and New Jersey. The remaining 27 states + D.C. will probably be too far off to have any major impact on the outcome, including Alaska and Texas, unless there will be a tremendous Hillary landslide that is.

All in all, the map might very well look quite unchanged from 2012, if the trendlines continue uninterrupted. The projection expects 11 states to be within 5 points, 14 states to be within 10 points, 29 states within 20 points, 41 states within 30 points, 46 states within 40 points and 49 states within 50 points. The only outstanding state is then Utah, plus obviously D.C.

That gives us this map:

[Space7]

Yes, it seems states are settling into place. They are less volatile than they were 20 or so years ago. Not many states seem to cross over ground zero any more.

An interesting point, your list shows Minnesota and Wisconsin in the same place in 2016. Funny, because far more people are paying attention to Wisconsin than Minnesota.

We continue to see where the Democrat's advantage comes from: the massive swath of weak Democratic states, which the Republicans, on the other side, severely lack (they only have Ohio, Florida, and North Carolina).

Just a small mistake. In the map you have put Vermont as a super strong Republican state, rather than a Democratic one.

[barfbag]

Yes, it seems states are settling into place. They are less volatile than they were 20 or so years ago. Not many states seem to cross over ground zero any more.

An interesting point, your list shows Minnesota and Wisconsin in the same place in 2016. Funny, because far more people are paying attention to Wisconsin than Minnesota.

We continue to see where the Democrat's advantage comes from: the massive swath of weak Democratic states, which the Republicans, on the other side, severely lack (they only have Ohio, Florida, and North Carolina).

Just a small mistake. In the map you have put Vermont as a super strong Republican state, rather than a Democratic one.

There's something to be said for the lack of competitive states these days. With all things being equal, I could name about 22 battleground states, but with the way elections usually end up only a few come to mind; NH, PA, VA, FL, OH, WI, IA, CO, NV, and maybe MN. If all things are equal between candidates, fundraising, the nature of the times, and voter turnout, then we could look at NC, MO, GA, IN, MT, DE, ME, NJ, MI, WA, OR, NM.

[eric82oslo]

Just a small mistake. In the map you have put Vermont as a super strong Republican state, rather than a Democratic one.

Thanks for being my eyes when I'm acting blind!