Jerome Armstrong's fuzzy logic

Jerome Armstrong, father of the netroots, has a post up on MyDD arguing that "Obama has a huge electability problem in [Ohio]." Armstrong's evidence?

[Obama] took a total of 5 counties [in last night's primary], and lost in 82 counties. ... You don't win a general election in Ohio if you can only win in 5 counties.

Arrrgh. This is such obviously, transparently flawed reasoning. As I've noted previously, "winning a primary in a given state [or county] does not mean you'll win the general election there, and likewise, losing a primary doesn't imply that you'll lose the general."

In a Democratic primary, between two Democratic candidates, some Democrat has to lose every county. (And every state, for that matter. But cf., Hillary's bogus swing-state argument.) The fact that somebody loses in the primary doesn't in any way suggest that the losing candidate couldn't win in November, running against a Republican. Those are two entirely separate calculations.

The fact is, in order to win in November, either Obama or Clinton will need to win the votes of a large majority of the Democrats who supported his/her opponent in the primaries. But it's impossible, by definition, for either of them to win a majority of their opponent's votes when they're running against each other. Again: someone's gotta lose. So a loss, by itself, tells us nothing about the November landscape.

Put another way: a vote for Clinton, without more, does not imply "I would never vote for Barack Obama." And a vote for Obama, without more, does not imply "I would never vote for Hillary Clinton."

Armstrong's argument could be just as easily, and with equal validity (i.e., none), be turned against Hillary Clinton:

She lost all of Ohio's major urban counties. You don't win a general election in Ohio if you can't win the urban counties. Clinton has a huge electability problem in Ohio.

But that's crap! Obama's victories in those urban counties in no way suggest that Hillary would lose to John McCain there. Likewise, Clinton's victories in various "swing" counties (as well as tons of solidly Republican counties) tells us nothing about her, or Obama's, general-election prospects there. Nothing. Nilch. Nada.

Simply put, last night's Ohio results provide no valid cause for concern about either candidate's chances in November, because primary elections simply are not a remotely reliable or relevant gauge of general-election prospects.

Well... maybe not quite no valid cause for concern. This, I'll admit, is a little disconcerting:

[B]oth exit polling and copious anecdotes from my colleagues on the trail suggest that some Ohioans were less ready than white voters elsewhere to elect, as Obama says, a black guy with a funny name. ... [T]he exit polls had about 20% of Democrats (and these are Democrats, remember) saying that race was a factor in their decision; of those, three quarters went for Clinton.

That data point is, I must admit, potentially a reason to worry about Obama's general-election chances. (If Ohio's Democrats cite race as a reason for voting against the black guy, what will happen when you add Republicans and independents into the mix?) TNR's John B. Judis elaborates:

The exit polls ask voters whether the "race of the candidates" was "important" in deciding their vote. If one looks at the percentage of Clinton (and earlier Edwards) voters who said it was "important," that is a fair estimate of the overall percentage of primary voters who were not inclined to vote for Obama because he was black. ...

In some February 5 states, the overall percentage of white (or Latino) primary voters who voted for white candidates partly because of race was pretty high. It was 9.5 percent, for instance, in New Jersey. In the general election, that percentage is likely to double; and some of these additional voters will be white working class or Latino voters that a Democratic candidate needs to win. In Wisconsin, the number was very low--only 6 percent. But in Ohio, a crucial swing state, it was 11.4 percent. That's a real danger sign for Obama in a state where elections can be decided by one or two percentage points.

If Armstrong wants to build a case that Obama has an electability problem, those are the numbers he should be citing, and then maybe he'd have a respectable argument. But the mere fact, without more, that Obama lost a bunch of Ohio counties to Clinton, is an absolutely ridiculous basis for such an analysis.

P.S. It should be noted that people who cite "race" as an important factor in their vote, and then vote for the white candidate, are not necessarily racists. They may simply be partisan Democrats who are worried about electability, and believe that other people, who are racists, would derail Obama in November. Such a stance enables racism, but it is not itself based on racism, as such. More importantly, this rationale would not carry over to a general election.

Comments

[yea]

good job brendan. interesting stuff and on-point. stuff like this makes this blog rock.

[yea]

i wonder if this was part of the reason/strategy that clinton did so well in ohio.

http://dailykos.com/storyonly/2008/3/5/131156/5021/187/469677

[Gardner]

Mitt Romney won 100% of the county's in Utah. That means that McCain has not chance?

[PenguinSix]

Yea, but it's not about the numbers....so the pundits say.

http://blog.washingtonpost.com/the-trail/2008/03/05/beyond_math_to_a_tougher_messa.html?hpid=topnews

[Brendan Loy]

From that article:

inevitability is no more an argument for him than it was for her last fall

Um... yes it is?

Last fall, Hillary had received zero votes and had earned zero delegates. Her "inevitability" was based purely on media hype and assumptions about what would happen.

Obama's "inevitability" is based on actual results... numbers... math... facts.

I'm not saying Obama should just sit back and run out the clock, but it's ludicrous to claim that he's "no more" inevitable now than she was then.

[Andrew]

If Ohio's Democrats cite race as a reason for voting against the black guy, what will happen when you add Republicans into the mix?

Nothing, because 90% or more of those Republicans who would cite "race" as their reason for not voting for Obama would not vote for HRC or any other prominent Dem candidate anyway. Additionally, Republicans are statistically far less likely to "defect" to the Democratic candidate than vice versa. Thus, your real concern should be "independents" who will or won't be voting for Obama based on his race, because in order for him to have a chance to win come November, Obama must be competitive and/or win significantly among independents, who previously have highly regarded McCain.

[Brendan Loy]

Good point, Andrew... one I've made before, but lost sight of in my rhetoric here. I changed the quoted sentence to read "Republicans and independents."

[Andrew]

I'm not saying Obama should just sit back and run out the clock, but it's ludicrous to claim that he's "no more" inevitable now than she was then.

I think your position is also unsupportable. You can't make claims about inevitability based on where the horses are coming into the third turn when they are still running neck-and-neck. HRC is probably going to gain a lot of media-driven momentum out of these recent victories, and whereas her earlier strategy hamstrung her post-Super Tuesday (February edition) efforts because she allocated all her resources to advertising in order to get a knockout punch, she now has time to build a ground game in the coming states and tweak her message to be more competitive in places where Obama would otherwise be heavily favored because of demographics.

[Brendan Loy]

I'm not saying Obama is inevitable now. I'm saying he has a much more plausible, substantive, fact-based claim at "inevitablity" than Hillary ever did. I don't assert that this claim is correct, just that it's facially plausible, and based on something real. (There is, after all, no mathematically plausible way Hillary can even come close to winning the pledged-delegate tally.) Whereas Hillary's "inevitability" was always just a creation of spin, polls and media punditry.

[Andrew]

Are you really discounting that "spin, polls and media punditry" is every bit as powerful as actual results?

[Brendan Loy]

It depends on the circumstances. Spin and punditry can be powerful and important, but recent elections have shown that spin/polls/punditry before a single vote has been cast is NOT a terribly good predictor of presidential nomination battles (e.g., Dean '04, Giuliani/Thompson/Romney '08, Hillary '08).

[Steve]

(If Ohio's Democrats cite race as a reason for voting against the black guy, what will happen when you add Republicans and independents into the mix?)
Sorry, Brendan, I enjoy your blog but I take exception to that statement. I am a white guy and a registered Republican and have NEVER considered voting for a Democrat for President...until now.
It isn't because I agree with Mr. Obama's policies (I don't), but at this point he has (as you have rightly observed) taken the high road and he represents a potential departure from all the rancor of the last 16 years.
Make no mistake, I'm not totally convinced and November is a long ways away. But if Obama is the Democratic nominee I will give the man serious consideration...if the nominee is Hillary than I will vote (and probably campaign for) Mr. McCain.
It has nothing to do with race, or gender. It has to do with leadership.

[JD]

Nilch?

[Brendan Loy]

Steve, there's really nothing to take exception to. I'm sorry if you interpreted my remarks as impugning independents' and Republicans' racial attitudes generally, but that certainly is not my intent, as I think my history of comments on this topic makes clear. (I have, for instance, repeatedly talked about the possible "Obamacan" -- Republicans for Obama -- phenomenon... I was talking about this before it was cool. :) My point is not that all, or even most, Republicans or independents are racist. I have firmly argued against that viewpoint on many occasions. My point is just that, in the aggregate, there are probably more anti-black racists among independent and Republican general-election voters than there are among Democratic primary voters. But I myself am an independent, so I'm hardly suggesting that all independents and Republicans are racists. :)

[Brendan Loy]

Nilch?

LOL. Oops. :)

[Paul]

Armstrong is not using fuzzy logic.

http://www.science.uva.nl/~seop/entries/logic-fuzzy/

Rather, Armstrong is guilty of making an unsound inference or, as you stated, flawed reasoning.

Love the blog!