The PA primary exit polls showed that an unusually high percentage of Hillary supporters would not vote for Obama in November should he be the nominee.

Is this a big problem for Democrats?

The McCain campaign believes that this poses a significant problem for Obama if he becomes the nominee.  Below is an excerpt of the McCain campaign strategy memo, which I copied from another blog:

Exit polls reveal why this poses significant problems for Obama if he becomes the nominee. The most important problem: Clinton voters don't automatically become Obama voters after he becomes the nominee. In fact, Obama leaves large portions of Clinton's coalition on the table in November.

Obama only wins 72% of the Democratic vote in a general election match up among those surveyed last night. Clinton shows her broad coalitional strength and wins 81% in a general election match up against John McCain. A full quarter of the Democrats in Pennsylvania are not willing to cast their ballot for Obama against McCain (15% say they vote McCain and 10% say they stay home), however, Clinton loses only 17% of Democrats (10% for McCain and 7% would not vote). This gap of 8-points would be significant in a general election match up. President Bush lost Pennsylvania by 2-points in 2004, when 41% of the electorate were Democrats. That 8-point gap among Democrats is enough to swing the state the other way (8% of 41% is 2.8-points, turning Pennsylvania red). This dynamic is clearly visible in publicly released surveys; an average of April polls show McCain trailing Obama by an average of 3-points (3 surveys in April) and trailing Clinton by 8-points.

The cracks in Obama's Democratic coalition in Pennsylvania mirror what we saw in Ohio, and those cracks could have implications in November.

Hillary Clinton cleaned up with Union households - like she did in Ohio. In Pennsylvania, Clinton won 59% of Union members (Obama 41%). Obama won these voters by significant margins in Wisconsin (+9), but has lost his hold on their vote in both Ohio (Clinton 55% - 43%) and now Pennsylvania.

Clinton did better than Obama with lower income voters.
Our targeting and analysis of the 2008 political landscape puts voters who are on the lower economic brackets at the heart of either party's winning coalition. Hillary won at every income level below $150,000, and Obama only won with the wealthiest Pennsylvania voters. Obama's media foibles contributed to his inability to connect to voters who are suffering the real impact of this challenging economic environment.

This is also apparent in the number of voters who feel Clinton is more in touch with their views. Fifty-six percent of Pennsylvania Democrats say Clinton cares about people like them - again a significant switch from earlier contests. Wisconsin exit polls shows Obama had a 12-point advantage on that measure. By the time Ohio held their primary, Clinton had switched the dynamic and led by 12-points.

Clinton won Catholic voters. In Wisconsin, Clinton split the Catholic vote 50%-50% with Obama. Again, she changed the dynamic in Ohio and won Catholics by 27-points (63% - 36%). In Pennsylvania, she increased her margins and won by 38-points (69% - 31%). The strength of this coalition bolsters her argument that Obama would have had problems competing in Michigan and will not be able to carry key Midwestern states in November.

Clinton won Jewish voters. In Pennsylvania, the first state where both candidates competed for a significant block of Jewish voters, Clinton won by 15-points (57% - 43%). Again, the data suggests Jewish voters, a key Democratic coalition, pose a potential problem for Obama.

Clinton increased her margins in suburban and rural areas - without losing ground in urban areas. Clinton won Pennsylvania suburbs by a 12-point margin and won rural areas by 22-points. And Clinton lost in urban areas by 14-points. This is similar to her Ohio performance. But, it shows an increase in her performance in urban areas from earlier contests (in Wisconsin she lost urban areas by 21-points). Clinton has figured out how to increase her margins among suburban and rural voters and cut into Obama's base of urban voters.

Instead of the usual "what is the matter with Kansas" invectives against these would-be defectors, I think it would be more useful for us to direct our energy to find a solution.  To solve this problem, we need to know the demographics of the would-be defectors.    And then, we need to determine the reason(s) for their defection.

The PA exit polls did not provide any data that would help us determine the demographics of the potential "defectors." The demographic breakdowns of a head-to-head poll in a State showing a big defection of Hillary voters to McCain would help us determine the demographics of the potential "defectors."

Massachusetts is such a state.

The following is the link to the cross-tabs of a recent SUSA head-to-head Clinton-McCain  and Obama-McCain poll.

Link

Hillary  56%
McCain  41%

Obama  48%
McCain  46%

Recently, Anil Adyanthaya wrote an interesting article entitled:  Can McCain Win Massachusetts? 

He opined that the Independent voters, who are the majority in MA, would find McCain more appealing than Obama and that it's the Independent voters that would help McCain win Massachusetts in November.

However, a careful examination of the 4/11/08 SUSA poll cross tabs showed that while Mr. Adyanthaya may be correct in his opinion that McCain will be competitive in Massachusetts, he was wrong in his designation of the Independent voters as the cause of this potential outcome.  

Party affiliation: The SUSA poll Party Affiliation breakdown shows that neither the Independent nor the Republican Hillary supporters are inclined to defect to McCain.  The would-be defectors belong to a subset of Democrats.  

Ideology: The data showed few liberals would defect to McCain.  The would-be defectors are predominantly moderate and conservative Democrats.  A good percentage of them are a) church attendees, b) evangelicals, c) pro-life, and d) own guns.

Gender: If males were the only voters, both Clinton and Obama would lose to McCain by 9% and 17% respectively.  Clinton's decisive 15% win margin against McCain was completely due to the overwhelming support of her female supporters.  A goodly percentage of the would-be defectors are female.  The poll showed that a small percentage (3-5%) of male Hillary supports would also defect to McCain if Obama is the nominee.

Race: 

Hispanics -- 78% of Hispanics would vote for Hillary versus 22% for McCain.  This is completely reversed for Obama versus McCain.  73% of Hispanics would vote for McCain and only 23% would vote for Obama.  

Blacks -- The poll showed that a substantial percentage of black Hillary supporters would defect to vote for McCain.  Hmmm... this is very surprising and difficult to interpret.

Whites -- a smaller percentage of whites would defect.  These defectors are not likely to be racists as the same Democrats overwhelmingly voted for Deval Patrick.

Age: The would-be defectors are over 35 years of age.

Top Issues: 

Economy -- Among the economy is the top issue voters, Clinton beats McCain by 4% while  Obama loses to McCain by 5%.  Since 36% of all voters chose the economy as their top issue, we can conclude that a substantial percentage of the would-be defectors are worried about the economy.

Social Security - this is the only other issue that showed a huge flipping of Hillary voters to McCain.  While Hillary beats McCain 60-40, Obama loses to McCain 32-68 among these voters.  

These two issues are related to financial well being of the would-be defectors.  It suggests that the would-be defectors may be worried about their financial well-being.

Conclusion:

I think this study has provided a glimpse into the demographics of the would-be defectors.  They are moderate to conservative Democrats.  They are not too worried about the Supreme Court as many of them are pro-life, own guns and attend church.  The defectors are not restricted to whites.  Some of them are whites, some are black, but a huge majority of Hispanic Hillary supporters would defect to McCain if Obama should become the nominee.  This suggests that Obama has a Hispanic problem that hasn't been solved by Richardson's endorsement.  They are not racists as the same demographics voted overwhelmingly for Deval Patrick.  They are made up of both males and females but a greater proportion of them are female.  They are >35 years old.  Many of them are not very confident about Obama's ability in handling the economy and the Social Security.

More in depth polls and studies are needed to provide us with a more complete picture.  This is just a beginning to unravel this problem that we face.

This demographic profile of the would-be defectors is not specific to Massachusetts.  The Ohio head-to-head SUSA poll showed very similar profile.  Thus, it is applicable to other swing states such as Ohio and Pennsylvania.  I haven't done a study of the profile of the "Reagan Democrats." I would appreciate discussions of whether this demographic is similar to the profile of the Reagan Democrats.