Half-Assed Rhode Island Delegate Projections (Now With Half-Assed Update!)

I'm a great admirer of the delegate projections that MattTX, redvolution, and especially poblano have posted in recent weeks. I don't nearly have the statistics (or HTML) chops that those guys do, but this evening it occurs to me that I might have what it takes to tackle a much simpler problem: the likely delegate splits in the Rhode Island Democratic primary on March 4.

If it matters, I'm not from Rhode Island, nor am I terribly knowledgeable about the place. I haven't even visited there in more than fifteen years. But if The Green Papers' Rhode Island page is accurate, things appear to stack up fairly simply.

A Rasmussen poll conducted last weekend has Clinton leading Obama 53%-38%. That and the Green Papers constitute the sum total of my extensive research (eat your heart out, poblano).

According to The Green Papers, Rhode Island's 21 pledged delegates are doled out as follows:

5 at large
3 PLEO (for our purposes, same thing as at large)
6 Congressional District 1
7 Congressional District 2

As a result of my steely-eyed "who needs research?" research strategy, I have no idea what demographic contrasts there are between CD 1 and CD 2. (It appears from the delegate allocation that CD 2 votes slightly more Democratic than CD 1, for whatever that's worth.)

I will presume that the current 53%-38% poll result is more or less accurate for both CDs (and for that matter for the state at large), because to do otherwise would be hard.

For delegate-pledging purposes, 53%-38% is better expressed as 58.2%-41.8%--the respective shares of the Clinton + Obama votes (throwing out undecideds and non-viable candidates).

If Rasmussen's findings are echoed in the election result, the statewide delegates will go 5-3 Clinton (3-2 at large + 2-1 PLEO). In fact, Obama has to win Rhode Island outright (50% + 1 vote) in order to do better than minus-2 net in the statewide delegates. He's also unlikely to lose any more than two there; he'd have to be under 30% to drop to minus-4.

The Congressional Districts are slightly more variable. Even behind 58.2%-41.8%, Obama splits CD 1's 6 pledged delegates evenly--though just barely. If he drops below 41 2/3% (of the Clinton + Obama vote), the delegates go 4-2 Clinton. Meanwhile, he has to get a ways over 58% of the vote himself to win any net delegates. So a dead heat in CD 1 is very likely, while Clinton +2 is not unthinkable.

In CD 2, Clinton stands to gain one net delegate unless she drops below 50% or gets above 64%.

So it seems very likely that Clinton will pick up three net delegates in Rhode Island--two in the statewide count and one in CD 2--for a 12-9 final tally. She could go as high as +5 (13-8) without a big change in projected vote totals.

On the other side, if Obama gets a significant swing from his visit Saturday or from his general nationwide momentum, picking up three delegates is within reach (he'd just have to get over 50% in CD 2 and in the state at large).

I should probably note that Kossack Twin Planets--after doing even less work at it than I have--several days ago predicted Clinton +3 in Rhode Island. That projection is obviously insignificant, because I want people to ignore it. Meanwhile, that old "leaked" Obama campaign spreadsheet had Clinton winning Rhode Island by five.

That's about the range at hand, I think.

----

UPDATE: A Fleming & Associates poll conducted Feb. 24-27 (last Sunday through Wednesday) shows Clinton ahead in Rhode Island 49%-40%. That's a substantial improvement for Obama from the earlier Rasmussen poll cited above.

But if this poll accurately reflects the state of the Rhode Island electorate, it still doesn't change any of the pledged-delegate-count results described above. At 49%-40% (i.e., 55.1%-44.9% after we throw out other candidates and undecideds), Clinton still gets +2 from the statewide delegates, a split in CD 1, and +1 from CD 2.

Obama's gain from the earlier Rasmussen poll suggests that it's slightly more likely now that he'll be able to pull out a win in the statewide count (which would mean +2 net delegates for him rather than minus-2) and/or the CD 2 count (turning a minus-1 into a +1). And the poll hints that it's now slightly less likely that Clinton could open up her CD 1 lead far enough to get +2 out of that district, leading to +5 for her in total.

But all of these are minor changes in probability at best. Clinton +3 still seems like a reasonably safe bet--the safest bet (about a state's specific delegate result), I think, of any of the contests on March 4.