Begining with those states I consider to be Leaning Obama (Wisconsin, New Mexico, Nebraska Congressional District 2 and Pennsylvania) I'll now explore likely Party ID and key demographic changes in the electorate since the 2004 Presidential election. Also included is the 2006 exit poll data. Taken together there is a good comparison of the high and low watermarks for both parties with regard to ID and demographic breakdowns (2004 being the GOP sweep year and 2006 being the Democrats' Congressional sweep). Lastly, I've checked this information against the latest voter registration estimates in the battleground states.
The purpose of this model is to serve as a likely prediction of what the composition of the electorate on November 4th is actually likely to be. By so doing, we create a sanity check against polls which deviate wildly from these fairly conservative estimations of party ID and demographic change. Such an undertaking is of course fraught with all the dangers of multiple changed variable predictions so let me first state clearly the assumptions for my estimations:
Assumptions
- Higher African American turnout
- Higher Hispanic turnout
- Higher youth turnout
- GOP turnout at 2004 levels (although a GOP turnout estimate from a credible source would be appreciated)
- Increased Democratic Party ID from 2004
- Decreased Republican Party ID from 2004
Party identification and polling demographics in WI, NM, NE-2 and PA
Wisconsin (10 electoral votes)
The 2004 network exit poll for Wisconsin revealed the following demographic breakdown of the WI electorate:
Female: 53%
Male: 47%
White: 90%/89%
African American: 5% (estimated 2008 increase: 6%)
Hispanic: 2%
Other: 3%
Ages:
Under 30: 20% (estimated 2008 increase: 21%)
31-64: 66% (estimated 2008 decrease: 65%)
Over 65: 14%
Party ID 2004:
Democratic: 35%
Republican: 38%
Independent: 27%
Party ID 2006:
Democratic: 39%
Republican: 34%
Independent: 27%
Party ID 2008 estimate:
Democratic: 39%
Republican: 34%
Independent: 27%
New Mexico (5 EVs)
The 2004 network exit poll for New Mexico revealed the following demographic breakdown of the NM electorate:
Female: 55%
Male: 45%
White: 57%/54%
African American: 2%
Hispanic: 32% (estimated 2008 increase: 35%)
Other: 9%
Ages:
Under 30: 17% (estimated 2008 increase: 19%)
31-64: 66% (estimated 2008 decrease: 64%)
Over 65: 17%
Party ID 2004:
Democratic: 40%
Republican: 33%
Independent: 27%
Party ID 2006:
Democratic: 41%
Republican: 32%
Independent: 27%
Party ID 2008 estimate:
Democratic: 42%
Republican: 32%
Independent: 26%
Nebraska Congressional District 2 (Omaha) (1 EV)
No exit poll or party ID information exists for NE-2 from 2004 and there is a lack of reliable polling at present to engage in a decent analysis. As soon as credible polling emerges (e.g. Survey USA) I shall both post anew as well as update this post. In the meantime, I strongly urge anyone with NE-2 data to send it to me please!
Pennsylvania (21)
The 2004 network exit poll for Pennsylvania revealed the following demographic breakdown of the PA electorate:
Female: 53%
Male: 47%
White: 82%
African American: 13%
Hispanic: 3%
Other: 2%
Ages:
Under 30: 21%
31-64: 65%
Over 65: 14%
Note: Such was the success of the Kerry campaign in Pennsylvania then as of this time, no clear opportunities for demographic change in PA appear to exist.
Party ID 2004:
Democratic: 41%
Republican: 39%
Independent: 20%
Party ID 2006:
Democratic: 43%
Republican: 38%
Independent: 19%
Party ID 2008 estimate:
Democratic: 42%
Republican: 38%
Independent: 20%
Conclusion: Between the scale of Obama's ground game in these states, the positive shifts in party ID even under conservative projections and the conservative demographic increases for youth and minority turnout, I estimate these states to all break for Obama by larger margins on election day then polls currently project.
Next up, the toss-up states.
No comments:
Post a Comment