As always, lots of data goodness from me. Last three Presidential Elections, Senate Elections, Governor elections, redistribution statistics and demographics.
But first of all, a shout-out to shamlet from RRH, who sent me the .drf file this is based off.
**cross-posted from RRH**
DRA has a lot of major issues for NJ-- a bunch of Trenton, Elizabeth, Newark and Lakewood is completely off on the election numbers. So in the areas that are affected by this-- the 8th, 10th, 4th and 12th districts my election numbers will be about as problematic as the DRA numbers are.
Luckily, none of the districts involved are even remotely swingy, so it doesn't matter too much.
DRA has a lot of major issues for NJ-- a bunch of Trenton, Elizabeth, Newark and Lakewood is completely off on the election numbers. So in the areas that are affected by this-- the 8th, 10th, 4th and 12th districts my election numbers will be about as problematic as the DRA numbers are. [EDIT: I built a fix for this]
Luckily, none of the districts involved are even remotely swingy, so it doesn't matter too much.
Below are the estimated "real" (adjusted for DRA problems) Obama-McCain numbers. I'll leave the wrong numbers posted at the bottom of the diary, as well as an explanation of my algorithm for estimating the adjusted numbers.
District |
Obama |
McCain |
1 |
65.4 |
34.6 |
2 |
53.9 |
46.1 |
3 |
51.9 |
48.1 |
4 |
45.7 |
54.3 |
5 |
48.9 |
51.1 |
6 |
59.1 |
40.9 |
7 |
47.8 |
52.2 |
8 |
74.4 |
25.6 |
9 |
64.2 |
35.8 |
10 |
85 |
15 |
11 |
47.4 |
52.6 |
12 |
66.3 |
33.7 |
District |
WhiteVAP |
BlackVAP |
HispanicVAP |
AsianVAP |
1 |
69 % |
15 % |
10 % |
4 % |
2 |
71 % |
12 % |
12 % |
3 % |
3 |
80 % |
10 % |
6 % |
3 % |
4 |
80 % |
6 % |
9 % |
4 % |
5 |
76 % |
4 % |
10 % |
9 % |
6 |
55 % |
9 % |
18 % |
16 % |
7 |
77 % |
4 % |
10 % |
8 % |
8 |
30 % |
9 % |
50 % |
9 % |
9 |
46 % |
9 % |
31 % |
12 % |
10 |
24 % |
50 % |
17 % |
7 % |
11 |
78 % |
4 % |
8 % |
9 % |
12 |
54 % |
17 % |
14 % |
14 % |
District |
|
|
|
|
|
1 |
Andrews 90% |
Runyan 10% |
|
|
|
2 |
LoBiondo 94% |
Runyan 6% |
|
|
|
3 |
Runyan 76% |
Smith 23% |
LoBiondo 1% |
Andrews 1% |
|
4 |
Smith 66% |
Holt 19% |
Pallone 14% |
Runyan 2% |
|
5 |
Garrett 79% |
Rothman 21% |
|
|
|
6 |
Pallone 64% |
Lance 20% |
Sires 11% |
Holt 6% |
|
7 |
Lance 61% |
Freylinghuysen 27% |
Garrett 5% |
Holt 3% |
Payne 3% |
8 |
Sires 68% |
Rothman 16% |
Payne 11% |
Pascrell 5% |
|
9 |
Rothman 52% |
Pascrell 44% |
Garrett 4% |
|
|
10 |
Payne 72% |
Sires 15% |
Pascrell 10% |
Lance 3% |
|
11 |
Freylinghuysen 65% |
Pascrell 32% |
Garrett 3% |
|
|
12 |
Holt 68% |
Pallone 14% |
Smith 10% |
Lance 7% |
|
District |
Kerry |
Bush |
PVI |
Gore |
Bush |
2004PVI |
1 |
60.5% |
39.5% |
D+12 |
65.4% |
34.6% |
D+13 |
2 |
49.1% |
50.9% |
EVEN |
55.7% |
44.3% |
D+3 |
3 |
47.9% |
52.1% |
R+1 |
54.6% |
45.4% |
D+2 |
4 |
42.5% |
57.5% |
R+7 |
49.9% |
50.1% |
R+3 |
5 |
45.7% |
54.3% |
R+4 |
50.6% |
49.4% |
R+1 |
6 |
55.3% |
44.7% |
D+6 |
61.4% |
38.6% |
D+9 |
7 |
43% |
57% |
R+6 |
44.8% |
55.2% |
R+6 |
8 |
68.9% |
31.1% |
D+20 |
73.2% |
26.8% |
D+22 |
9 |
60.6% |
39.4% |
D+11 |
64.9% |
35.1% |
D+13 |
10 |
81.4% |
18.6% |
D+32 |
83.6% |
16.4% |
D+33 |
11 |
42.5% |
57.5% |
R+6 |
45.3% |
54.7% |
R+6 |
12 |
61.4% |
38.6% |
D+13 |
65.1% |
34.9% |
D+14 |
|
2001 Governor |
|
2005 Governor |
|
2009 Governor |
|
District |
McGreevey |
Schundler |
Corzine |
Forrester |
Corzine |
Christie |
1 |
64.7% |
35.3% |
61.3% |
38.7% |
55.9% |
44.1% |
2 |
55.6% |
44.4% |
52.7% |
47.3% |
46.1% |
53.9% |
3 |
55% |
45% |
49% |
51% |
40.7% |
59.3% |
4 |
49.6% |
50.4% |
43.7% |
56.3% |
32.8% |
67.2% |
5 |
49.7% |
50.3% |
49.7% |
50.3% |
43.2% |
56.8% |
6 |
61.6% |
38.4% |
56.8% |
43.2% |
45.7% |
54.3% |
7 |
43.5% |
56.5% |
42% |
58% |
34.8% |
65.2% |
8 |
70.6% |
29.4% |
76% |
24% |
71.7% |
28.3% |
9 |
63.6% |
36.4% |
65% |
35% |
59.7% |
40.3% |
10 |
82.4% |
17.6% |
84.3% |
15.7% |
81.4% |
18.6% |
11 |
44.5% |
55.5% |
44.5% |
55.5% |
37.6% |
62.4% |
12 |
65.3% |
34.7% |
61.4% |
38.6% |
55.1% |
44.9% |
|
2002 Senator |
|
2006 Senator |
|
2008 Senator |
|
District |
Lautenberg |
Forrester |
Menendez |
Kean Jr. |
Lautenberg |
Zimmer |
1 |
63.1% |
36.9% |
60.9% |
39.1% |
64.9% |
35.1% |
2 |
52.2% |
47.8% |
49.1% |
50.9% |
54.9% |
45.1% |
3 |
50.4% |
49.6% |
48.2% |
51.8% |
52% |
48% |
4 |
45.4% |
54.6% |
43.5% |
56.5% |
45.1% |
54.9% |
5 |
48.7% |
51.3% |
47.8% |
52.2% |
50.1% |
49.9% |
6 |
58.1% |
41.9% |
57.5% |
42.5% |
58.6% |
41.4% |
7 |
41% |
59% |
43% |
57% |
43.9% |
56.1% |
8 |
73.7% |
26.3% |
75.2% |
24.8% |
75.6% |
24.4% |
9 |
63.2% |
36.8% |
62.7% |
37.3% |
66.1% |
33.9% |
10 |
82.8% |
17.2% |
83.6% |
16.4% |
85% |
15% |
11 |
42.5% |
57.5% |
43.4% |
56.6% |
46.1% |
53.9% |
12 |
61.7% |
38.3% |
62.9% |
37.1% |
64.1% |
35.9% |
__
As mentioned above, here are the faulty Obama/McCain numbers from DRA, and the D/R numbers from DRA that are just as wrong as the Obama/McCain numbers, as well as an explanation of my algorithm for generating the "real" Obama/McCain results (which are just an estimate, but should come fairly close to the real numbers).
District |
Obama |
McCain |
"D" |
"R" |
1 |
65 |
35 |
62.6 |
37.4 |
2 |
53.6 |
46.4 |
51.9 |
48.1 |
3 |
51.6 |
48.4 |
49.2 |
50.8 |
4 |
46.7 |
53.3 |
45 |
55 |
5 |
48.5 |
51.5 |
47.1 |
52.9 |
6 |
59 |
41 |
58.4 |
41.6 |
7 |
47.1 |
52.9 |
42.7 |
57.3 |
8 |
72.8 |
27.2 |
72.8 |
27.2 |
9 |
64 |
36 |
64.2 |
35.8 |
10 |
84.7 |
15.3 |
81.3 |
18.7 |
11 |
46.9 |
53.1 |
45 |
55 |
12 |
65.5 |
34.5 |
60.8 |
39.2 |
Here's my algorithm: First of all I add up the population of all the precincts in a county where the DRA-reported turnout was less than 10% of VAP.
In counties where this adds up to less than 10,000 people I apply a simple fix (for any of the following, do the same for McCain as for Obama):
Let's call the sum of Obama votes in County x Obama_DRA. I calculate Obama_DRA and look at the discrepancy between Obama_DRA and Obama's real votes. I calculate how many percent Obama got more in every county than DRA says he did: Obama/Obama_DRA.
Then I multiply the votes DRA reports in any given precinct with this variable: Obama_DRA*(Obama/Obama_DRA).
In counties where precincts with more than 10,000 people are missing I multiply with the average of Obama/Obama_DRA in all counties where there were fewer problems. (This is about 1.09 for Obama and 1.07 for McCain).
Then I calculate how many votes are still missing from the real election results, and how much population the precincts with problems have. Say, in Ocean County 8,000 McCain votes and 5,000 Obama votes are missing and the precincts with problems have a population of 50,000. Then I will assign each precinct .1 Obama votes per person and .16 McCain votes per person.
This should come fairly close to the real results.