With the statewide election season upon us and with new electoral districts (I hear the Supreme Court will issue its redisricting law suit decision tomorrow) in place, it's time to reveal my numbers system. Tomorrow, I'll post the margins in the 24 Senate districts, but first let's get to know the system I use.
Methodology of My Rating System
If we want to discover how Democratic or Republican a particular city, town, or ward is—based on actual voting history and not based on the less reliable voter registration tables—the fist thing we need to do is establish a baseline.
In 2002, I discovered that if we used the five closest statewide races for the previous decade, the parties virtually split the votes cast in those five races statewide. (I always disregard third or fourth party candidates). The races were 1992 for President and U.S. Senate, 1996 for U.S. Senate, and 2000 for President and U.S. Senate. Statewide in those five races, Republicans received 1,214,890 votes while Democrats received 1,212,805 votes. That amounts to 50.04 percent for Republicans; 49.96 percent for Democrats, close enough to 50-50 so that we can then add up totals in any town, city, or ward and get a number to judge just how much any advantage one of the parties would enjoy in that location, all other factors being equal.
The exercise worked so well that I decided to do it again for the last ten years. Data from one close race is hardly enough, so I sought out five of them and discovered that we had a large share of blowouts the past then years. So the best I could come up with was four statewide races (for U.S. Senate in 2002; for President and Governor in 2004; and for Governor in 2010). For the fifth, I did something I would prefer not to do, but it works fairly well…I used the two Congressional races of 2010.
If you add the numbers for Democratic candidates in those five races, you get 1,329,197; for Republicans, it’s 1,319,928. That breaks down to 50.17 for Democrats to 49.83 for Republicans, not quite as close to 50-50 as we were in 2002, but it’s good enough for this exercise.
Here are the numbers:
2002 Senate—John Sununu 227,224, Jeanne Shaheen 207,478.
2004 President—John Kerry 340,511, George W. Bush 331,237
2004 Governor—John Lynch 341,299, Craig Benson 325,981
2010 Governor—John Lynch 240,346, John Stephen 205,616
2010 1st CD—Frank Guinta 121,655, Carol Shea Porter 95,503
2010 2nd CD—Charlie Bass 108,610, Kuster 105,616
Using these same races, we can get a total for any town, ward, or city, and then we can add these numbers to get a number for any race we desire (it’s limited only by the amount of time you want to spend running the numbers; or creating a computer program if you're into that) from all State Rep districts (including floats) to the 24 State Senate districts to Executive Council districts…to anything.
Not only that, we can compare the number from 2002 to 2012, and if we really want to create more work, we can look for example at how an unchanged Senate district would stack up to how the ones approved for this year look. Remember I did that last week to show how Republican Senate mappers had made District 17 less Republican (Jack Barnes didn’t seem to need the help; now, he’s decided not to run, the decision could come back to haunt Republicans) in order to make District 23 more Republican (to help Russ Prescott). Be careful what you wish for.
I’ve run the numbers for all 24 Senate districts, and it’s simply amazing (albeit not really surprising) on how Democrats have been stacked into five specific districts (5 around Lebanon-Hanover; 10 around Keene; 15 around Concord; 21 around Portsmouth-Durham; and to a somewhat lesser extent the new district 4 around Dover-Somersworth). It's a real eye opener.
This will become more evident in chart form, and I’ll post the chart separately.
For the purpose of consistency, Republican totals will always be listed first, Democrats second. The higher above 50, the more Republican an area is; the lower the number, the more Democratic the area. I arrive at an advantage by subtracting the two numbers.
Let’s run just one example here, and I’ll use my own Manchester Ward 8 which just happens to be the most Republican ward in the city (note that it’s the only ward which gave more votes to Stephen than to Lynch for Governor in 2010).
Ward 8 Totals For The Five Races
2002—U.S. Senate—Sununu 1804, Shaheen 1321
2004—President—Bush 2613, Kerry 1983
2004—Governor—Benson 2598, Lynch 1959
2010—Governor—Stephen 1603, Lynch 1576
2012 1st CD—Guinta 1768, Porter 1298
Five Combined Races—Republican 10,386, Democrat 8137
That’s 56.07 percent for Republicans, 43.93 percent for Democrats or a GOP advantage of 12.14 percent. (The 2002 number for ward 8 was 54.15-45.85 for a GOP advantage of 8.30. Thus, Ward 8—in these five races at least—has become quite a bit more Republican in the past ten years. Republicans should be winning more down ballot seats here. That's why it was especially sad for Republicans that they could only come up with two candidates to run for three state rep seats two years ago; it might not matter to get Republicans to run in highly Democratic ward 3, but it's a missed opportunity to fail to fill the slate in Ward 8). This knowledge will obvioulsy prove useful throughout the state.
As a quick example, Rochester is just about 50-50, so races should be close there in a "normal" year. The fact that Democrats won most seats there in 2006 and Republicans in 2010 shows that those were not "normal" years. I suspect 2012 will be fairly "normal"; in other words, neutral at the top of the ticket.
A great thing about this system is that actual raw numbers are used. Thus, when it comes to creating a Senate district tally (for District 18 for example),Ward 8 with a fairly high turnout would contribute much more than low turnout ward 5. (The Ward 5 totals are 4649 for Republican candidates in the five races; 5380 for the Democratic candidates thus a ranking of 46.36 for the inner city ward).
Since so many numbers are involved, my assumption is that I’ve made a few mistakes, but I trust they are minor. You can run the numbers for anything. Let’s make Litchfield a test case. Dig out the red books for the five elections. Fill the numbers into the chart, and see what you come up with. If it’s not 59.38 (a very strong Republican number indeed), either you or I have erred. Ten years ago, Litchfield was 54.14, so it’s gotten much more Republican…again based on top of the ticket close races.
Go for it. Let me know what you get and I’ll verify any numbers sent in.