Survey USA released a final poll for Indiana, it shows that Clinton has a 12 point lead over Obama (54%-42%), now, as you might imagine, is in direct conflict with my own Indiana projection so a little analysis is needed to see where these numbers are coming from, and whether this will lead to a significant movement in delegates.

First of all, I’m personally skeptical of SUSA’s number, especially considering how badly they messed up Missouri (this can be seen almost exclusively by missing how many Independents would turnout for the Democratic primary, they predicted that Independents would make up 11% of voters when they actually made up 22%), it’s not that they’re a bad pollster (on the contrary, they’re likely the best pollster in the field), but some of these numbers are suspect to me.

I’ll ignore the racial make up for the moment, as the sample of African Americans is way too small to be representative of anything (basically 68 people), and look at the regional and party affiliation break downs. SUSA predicts that Republicans and Independents will make up a total of 28% of the vote (14% each). I don’t believe that Republicans will make up that much of the voting sample (they will likely make up a sizeable part of the vote, either way), I think it will be more like Mississippi in that Republicans make up 11%-12% of the vote and Independents making up 16%-17% of the vote; although I must admit that changing those numbers don’t really move anything (really, it’s about a one point movement if SUSA is to be believed). The other thing is the regional cross-tabs, particularly the Northern Indiana number (SUSA projects that Clinton gets 53% of the vote here while Obama only gets 42% of the vote). My own projection for the three congressional districts which make up the region (a total of 16 delegates) would give Obama 9 delegates and Clinton 7 delegates, which would mean that I was also thinking that Obama would be able to both keep Clinton’s margin down in CD-02 and CD-03 and that Obama would be able to rack up his margin in CD-01. On top of that, I also projected that Obama would be able to keep Clinton below 62.4% of the vote in CD-04 and CD-05 (which gives 8 delegates) this should’ve given a delegate tie. However, according to SUSA, Clinton is expected to win the region with at least 62% of the vote, this means that in one (or possibly both) of the CDs, Clinton could force a 3-1 split in one or both districts. Finally, according to SUSA’s polling, Southern Indiana shows Obama only getting 34% of the vote (with Clinton getting 61%), if this holds up, it’s likely that my 3-3 split for CD-09 was overly optimistic, and will break 4-2 for Clinton.

In addition, Al Giordano of The Field makes a very good point in the comments section:

A lot of you folks are setting yourselves up for another cycle of over-expectations. Clinton will win Indiana by 10 points or more. Susa is right on the money on that. You can cherry pick things about the poll’s too-small-to-be-statistically-significant cross tabs (AA vote, Independent vote, i.e.) but, for example, Tara asks “how can IN be worse than PA?” Easy. Smaller AA population. And don’t forget about Operation Chaos and Fox News’ more-blatant-than-ever campaigning for HRC this week (how many times did they run that puff interview on the O’Reilly Factor? Five or six that I’ve seen, and Indiana is one of Fox’s highest viewership states.

So where does that leave us? Well, I stand by my initial projection, but if I’m wrong, the districts which will likely change are:

CD-01 (Obama 3, Clinton 3) Clinton +2 from before
CD-05 (Clinton 3, Obama 1) Clinton +2 from Before
CD-09 (Clinton 4, Obama 2) Clinton +2 from before
At-Large (Clinton 9, Obama 7) (Clinton +2 from before)

Leaving it at a grand total of Clinton 40, Obama 32 (Clinton +8 )

If my other projection was wrong then I’ll take my lumps (and my bruised ego ) and make sure I learn from what I did wrong.

UPDATE: PPP has released their final numbers from Indiana, they show that Clinton only has 51% while Obama has 46% of the vote. Naturally, I believe that these numbers are closer to where Indiana is right now (as my delegate projection has shown). This will be the third state where PPP and SUSA are in major conflict (the first was Texas where PPP came out the winner and the second was Pennsylvania where SUSA was right and PPP was wrong). I went with SUSA in Pennsylvania because I never bought that Obama was going to win the state. PPP is not a bad pollster though, and they’ve proven themselves to be solid in Wisconsin, Ohio, South Carolina (where they were among the only ones to accurately hit the demographic turnout), and Texas. I’m going to say that PPP is right here, but I’m also going to caution that SUSA’s numbers could very easily be correct.