Wednesday, November 10, 2004
Fallacious Meme 3: Dems have a Comparative Disadvantage in "Away Games"
In this post explaining why he thinks Bush doesn't have a mandate (I point I find debatable, not wrong, just debatable), Josh Marshall injects the following paragraph.
Ok let's only examine those current office holders who were elected for the first time from 2000 forward (a date when I think we can agree the red/blue divide was defined.) We then have the following statistics:
2004: 8 Red, 1 Blue, 0 R|B (0 Reps given a blue state), 1D|R (1 Dem given a red state)
Colorado(R) = Ken Salazar(D)
Florida(R) = Mel Martinez(R)
Georgia(R) = Johnny Isakson(R)
Illinois(B) = Barack Obama(D)
Louisiana(R) = David Vitter(R)
North Carolina(R) = Richard Burr(R)
Oklahoma(R) = Tom Coburn(R)
South Carolina(R) = Jim DeMint(R)
South Dakota(R) = Jim Thune(R)
2002 Election: 8 Red, 2 Blue, 1 D|R, 1 R|B
Arkansas(R) = Mark Pryor(D)
Georgia(R) = Saxby Chambliss(R)
Minnesota(B) = Norm Coleman(R)
Missouri(R) = Jim Talent(R)
New Hampshire(R)=John Sununu(R) (Considered Red at the time. Blue assignment considered below)
New Jersey(B)=Frank Lautenberg(D)
North Carolina(R)=Elizabeth Dole(R)
South Carolina(R)=Lindsey Graham(R)
Tennessee(R) = Lamar Alexander II (R) (Teve Torbes!!!)
Texas(R)=John Cornyn(R)
2000 Election: 4 Red, 6 Blue, 2 D|R, 0R|B
Deleware(B)=Thomas Carper(D)
Florida(R)=Bill Nelson(D)
Michigan(B) = Debbie Stabenow(D)
Minnesota(B) = Mark Dayton(D)
Nebraska(R) = Ben Nelson(D)
Nevada(R) = John Ensign(R)
New Jersey(B) = Jon Corzine(D)
New York(B) = Hillary Clinton(D)
Virginia(R)= George Allen(R)
Washington(B)=Maria Cantwell(D)
Totals for all new senators elected from 2000-2004.
20 Red, 9 Blue, 17 Rep 12 Dem, 1 R|B, 4D|R
The relative probabilities of winning an "away game" are as follows: Probabilities: p(R|B) = 1/12 = 8.3%, p(D|R)=4/20 = 20%.
If I choose NH in 2002 to be blue instead, then the probabilites become: p(R|B) = 2/11 = 18.2%, p(D|R)=4/19 = 21.1% and Dems still have a greater chance of winning "away games" than Reps.
As a further indication that these are not outlier statistics, consider the implication of these probabilities as applied to the current red/blue divide:
62 * (.8) + 38 * (1/12) = 53 Rep Senators.
62 * (.8) + 38 * (2/11) = 56 Rep Senators.
So judging from the current Senate breakdown (55 Rep), the two estimations appear to bracket what should be the "true" probability, giving us further confidence that Dems do hold some advantage in away games.
Bottom line: The exact opposite of Marshall's1 thesis is reality. Democrats have a (slight) comparative advantage running in states that lean in the opposite direction in the Presidential election.
Note members of the House are not considered in this analysis as (most) House seats are not statewide offices and thus not easily considered within the context of the red/blue state divide. Also note the 2000 Carnahan election is not considered (wouldn't help the original thesis anyways) as that seat was determined in 2002. Also the 2000 Miller election is not considered as that seat was determined in 2004 (again this wouldn't help the original thesis).
Election Results gathered from wikipedia:
http://en.wikipedia.org/wiki/U.S._Senate_election%2C_2004
http://en.wikipedia.org/wiki/U.S._Senate_election%2C_2002
http://en.wikipedia.org/wiki/U.S._Senate_election%2C_2000
Footnote
1. Instead of "Marshall's", the post originally said "your." However, as the post is not a letter to Marshall, the use of the second person seemed a bit stupid.
Comments(0) |
But what I mean by that clunky phrasing is that Republican senators can still often run and win in blue states despite the unpopularity of the national Republican party in those states. But Democrats have a far harder time doing the same thing -- as Daschle, Bowles, Tenenbaum, Castor and Knowles found out to their dismay.This prompted me to email Josh the following:
I don't follow. There were 31 red states and 19 blue states. If it were harder for a Dem to be elected in a red state than for a Rep to be elected in a blue state, then shouldn't there be more than 62 Rep senators? As it stands there are 55 Rep senators, so the exact opposite appears to be true.Josh then replied with:
your reasoning doesn't make sense to me for several reasons. for starters, the tenure of most senators predates what we now call blue and red states.This prompted the following email from me (some editing for the post - mostly coloring which I couldn't do in the email).
Ok let's only examine those current office holders who were elected for the first time from 2000 forward (a date when I think we can agree the red/blue divide was defined.) We then have the following statistics:
2004: 8 Red, 1 Blue, 0 R|B (0 Reps given a blue state), 1D|R (1 Dem given a red state)
Colorado(R) = Ken Salazar(D)
Florida(R) = Mel Martinez(R)
Georgia(R) = Johnny Isakson(R)
Illinois(B) = Barack Obama(D)
Louisiana(R) = David Vitter(R)
North Carolina(R) = Richard Burr(R)
Oklahoma(R) = Tom Coburn(R)
South Carolina(R) = Jim DeMint(R)
South Dakota(R) = Jim Thune(R)
2002 Election: 8 Red, 2 Blue, 1 D|R, 1 R|B
Arkansas(R) = Mark Pryor(D)
Georgia(R) = Saxby Chambliss(R)
Minnesota(B) = Norm Coleman(R)
Missouri(R) = Jim Talent(R)
New Hampshire(R)=John Sununu(R) (Considered Red at the time. Blue assignment considered below)
New Jersey(B)=Frank Lautenberg(D)
North Carolina(R)=Elizabeth Dole(R)
South Carolina(R)=Lindsey Graham(R)
Tennessee(R) = Lamar Alexander II (R) (Teve Torbes!!!)
Texas(R)=John Cornyn(R)
2000 Election: 4 Red, 6 Blue, 2 D|R, 0R|B
Deleware(B)=Thomas Carper(D)
Florida(R)=Bill Nelson(D)
Michigan(B) = Debbie Stabenow(D)
Minnesota(B) = Mark Dayton(D)
Nebraska(R) = Ben Nelson(D)
Nevada(R) = John Ensign(R)
New Jersey(B) = Jon Corzine(D)
New York(B) = Hillary Clinton(D)
Virginia(R)= George Allen(R)
Washington(B)=Maria Cantwell(D)
Totals for all new senators elected from 2000-2004.
20 Red, 9 Blue, 17 Rep 12 Dem, 1 R|B, 4D|R
The relative probabilities of winning an "away game" are as follows: Probabilities: p(R|B) = 1/12 = 8.3%, p(D|R)=4/20 = 20%.
If I choose NH in 2002 to be blue instead, then the probabilites become: p(R|B) = 2/11 = 18.2%, p(D|R)=4/19 = 21.1% and Dems still have a greater chance of winning "away games" than Reps.
As a further indication that these are not outlier statistics, consider the implication of these probabilities as applied to the current red/blue divide:
62 * (.8) + 38 * (1/12) = 53 Rep Senators.
62 * (.8) + 38 * (2/11) = 56 Rep Senators.
So judging from the current Senate breakdown (55 Rep), the two estimations appear to bracket what should be the "true" probability, giving us further confidence that Dems do hold some advantage in away games.
Bottom line: The exact opposite of Marshall's1 thesis is reality. Democrats have a (slight) comparative advantage running in states that lean in the opposite direction in the Presidential election.
Note members of the House are not considered in this analysis as (most) House seats are not statewide offices and thus not easily considered within the context of the red/blue state divide. Also note the 2000 Carnahan election is not considered (wouldn't help the original thesis anyways) as that seat was determined in 2002. Also the 2000 Miller election is not considered as that seat was determined in 2004 (again this wouldn't help the original thesis).
Election Results gathered from wikipedia:
http://en.wikipedia.org/wiki/U.S._Senate_election%2C_2004
http://en.wikipedia.org/wiki/U.S._Senate_election%2C_2002
http://en.wikipedia.org/wiki/U.S._Senate_election%2C_2000
Footnote
1. Instead of "Marshall's", the post originally said "your." However, as the post is not a letter to Marshall, the use of the second person seemed a bit stupid.
Comments(0) |
