Dear Ed,

I provide this link to a Kevin Drum posting from a few days ago because I think it's good common sense. But, more, it inspired an idea that -- were I even vaguely capable in mathematics and computer programming -- I'd be very inclined to put into action myself, just to see if it ultimately proved of value.

It seems to me that with some simple polling (or perhaps even just a close reading of the stated positions of all Democrats in the next Congress), combined with an application of Set Theory, you could pretty easily arrive at a priority list that would maximize the probability of successful passage of a vast range of Democratic policies while minimizing intra-Party friction ... _and_ minimizing the odds of the public turning against the Democrats in 2008.

I dunno, maybe this is nothing new. Maybe one or more mathematicians have done this in the past. But if so, I've never heard of it. I've never, for example, seen an illustration of it in, say, the textbooks I used to have on political game theory, which are nothing but mathematics.

So, if I had the brain for it, here's what I'd do, stated here on the off-chance that someone else -- someone who _does_ have head for math and programming -- might be inspired to act.

Start with one of your elaborate compilations of public opinion surveys, only this one being of what the public broadly claims to believe about various issues. Which issues are most salient to the general public? Which are the highest in priority? There's a big difference between those two things. For example, a substantial majority favor tough gun control laws. But they favor it, most of them, the way we favor world peace and an end to the illicit drug trade: "Gee, wouldn't it be a good thing if ..." On the other hand, folks who oppose gun controls REALLY oppose them. They aren't a majority or even very close to one, but because they can quickly become single-issue voters, they wield a lot of power. In short, they have high salience on this issue; taking it on can consume a _lot_ of political capital, especially if it's something you're likely to alienate a lot of your own erstwhile supporters over.

So anyhow, you list every issue you might imagine -- the Iraq War, national health care, gun control, raising the minimum wage, corporate reforms, abortion rights, and so forth -- and to each, you assign a composite image of what the public, nationally, thinks about them ... along with a composite "salience" calculation. So, for example, I'd bet that the Iraq war will stay right at the top where, most of us on the left, believe it to be and believe it deserves to be; that's because a substantial majority of the public are fed up with the direction, and they feel pretty strongly about it. Gun control would probably fall pretty far on the list, very nearly to the bottom.

That's one set.

To build the second set, you examine the stated national issues of concern to each Democratic Senator or Congressman. How does each of these elected leaders stand on the issues? How high is the salience of the issue to each of these people (perhaps one of the hardest -- or at least most subjective -- values to determine)? How about his or her constituents -- where do they, overall, stand on it? (Perhaps this is two numbers: overall public opinion _and_ the opinion of Democratic voters in his or her state or district). So, Casey of Pennsylvania is anti-abortion, and it might be of high personal salience to him, but probably isn't a key issue or his state's voters (indeed, many probably voted for him despite rather than because of his views on this). Tester, on the other hand, is very pro-gun rights, has high salience on it, and would get his butt kicked out of office if he crossed that. So ... mix all of this information together, and you come up with your second set.

Finally, the most subjective consideration of all: what issues can realistically be dealt with, with or without nearly unified Democratic support. A national healthcare plan ought to be in the cards, for example -- God, I'd give anything for it -- but, realistically, it might have no chance at all with Bush in the White House and Republicans still very powerful in the Senate (hence able to kill legislation with
filibusters and / or by merely attracting a stray Democrat or two to their position). Election reform and a better energy plan, however, might be issues that Bush would have a harder time blocking.

Anyhow, this is your third set.

Now crunch them altogether, and a rational priority list should emerge. Not perfect, not something to completely substitute for the experienced judgment of Democratic political leaders, but a lot better than the black box they use to figure out these things now (e.g., every member of Congress dropping things into the hopper, and all the committee leaders mostly exercise their own biases with little outside guidance).

I'm not suggesting that Henry Waxman or Charles Rangel or anyone else up there is going to allow a computer program to dictate what they do. But if all or most of these leaders are convinced that this would maximize ultimate effectiveness -- and, thus, their re-election prospects -- then they might be very inclined to give heavy value to such a priority list.

In truth, people like Reid and Pelosi already try to do this anyhow. I'm just inclined to think that a way to let a computer generate such a priority list might be recognized as a useful -- powerful -- tool.