Blue Mass Group

Reality-based commentary on politics.

BMG / RMG: MA Senate prediction composites 1/19/2010

By

BlueMassGroup
Brown Coakley Kennedy Total
BMG Straw Poll 5 31 0 36
% 13.89% 86.11% 0.00%
BMG Predictions
Brown Coakley Kennedy Turnout
sleeples 48 50 2 1100000
JoeTS 49 47 3
Carey Theil 45 53 2
Peter Porcupine 49 47 3
Wookie 46 52 2
BrooklineTom 46 61 3
Hoyapaul 47.5 51 1.5 1200000
cater68 51 45 4
Steve Stein 43 51 5 1200000
sabutai 47 51 2 785000
Ray M 45 52 3
HLPeary 46 52 2
PatrickLong 44 51 5 660000
ruppert 46 53 2
jconway 47.5 49.5 3 600000
petr 32 59 850000
eddiecoyle 44 54 2
JohnD 49 48.5 2.5
PeteSimon 47 49 3
FrankSkeffington
1900000
Argyle 50 49 1
Manny Happy Returns 53 41 3
Witch7 54 44 2
BigD 46 49 5

BMG and RMG composites below … along with the beginnings of a predictive model for how these numbers could potentially be used.

And the straw poll and predictions from RedMassGroup:

RedMassGroup
Brown Coakley Kennedy Total
RMG Straw Poll 18 1 0 19
% 94.74% 5.26% 0.00%

RMG Predictions
Brown Coakley Kennedy Turnout
sleepleswithamoustache 48 50 2 1100000
Ken Pittman 52 47 1
Rob “EaBo Clipper” Eno 50.5 48.5 1 870000
yankeepundit 51 48 1 1050000
nomad943 54 44 2
garyrlake 51 47 2 880000
MerrimackMan 49 48 3 1000000
Vote3rdpartynow 48 48 4
ElectricStrawberry 48.1 48.1 3.8 3000000
Mike “DD4RP” Rossettie 1400000
Republican Ram Rod Radio 51 48 1 680000
Peter Porcupine 49 47 4

The composite values:

Composite Predictions
Brown Coakley Kennedy Turnout
BMG Average: 46.74% 50.39% 2.77% 1036875
RMG Average: 50.15% 47.60% 2.25% 1247500
BMG/RMG AVG 48.44% 49.00% 2.51% 1142187.5
Model Adjusted**: 47.42% 49.57% 2.95%

So there we have it: BlueMass people overall think it will be a 50-47 Coakley victory and RedMass people average a 50-48 Brown victory, and RedMass predicts a higher turnout (1.2 mil to 1 mil).

The average between the prediction margins, which is a really quick and dirty way to aggregate the percentages, gives an absolutely razor thin 49-48.5 Coakley victory. Apparently, the collective Conventional Wisdom is that this will be an extremely close race.

**The “Model Adjusted” field is my attempt at creating a prediction model out of this data. I initially put a description of where it came from here, but it was too lengthy so I will add it as a separate diary entry and link later.

Please share widely!
fb-share-icon
Tweet
0

Comments

  2. and not at all eligible to be included…

    <

    p>I think Coakley gets 54 and Kennedy gets 2, leaving 44 for Brown.  I think high turnout helps her, and substantially.  Sort of a “silent majority”.

    <

    p>Of course, this could just be wishful thinking… 54% have showed up to vote for a GOP governor in the past.  Well, not exactly.  I looked it up.  GOP candidates for governor:

    <

    p>1986 525,364 votes 30%
    1990 1,175,817 votes 50%
    1994 1,533,390 votes 71%
    1998 967,160 votes 51%
    2002 1,091,988 votes 50%
    2006 784,342 votes 35%

    <

    p>So, what have we learned?  1994 was a massive wave, and the incumbent was popular.  That doesn’t match this at all, although we may find 2010 to be a wave.

    <

    p>1990, 1998, 2002 all had the GOP candidate right at 50%, at around 1,000,000 votes.  1986 and 2006 had a terrible GOP result.

    <

    p>It’s pretty clear that Brown will get less than 71% and more than 35%.  I don’t think he’ll pull in 1,000,000 votes, which seems to be the natural GOP ceiling.  My guess is that he’ll bring in right around 700k.  The question is: can Coakley drive out the Democrats?

  3. And you even included some cool charts.

    <

    p>It does appear that turnout will be quite a bit higher than expected, and perhaps even a decent amount higher than my 1,200,000 prediction. We may even hit 1,500,000, which certainly would be a shock for a January special election, even with all of the national attention on the race.

Archives