The only way out is by “holding the course”, keeping our current force levels & training the Iraqi security forces until they are ready to take over for us, make it clear to them we will leave when we are through, then when we are through, leave.

Your regular army generals you keep quoting & propping up are experts at running over people with tanks and smashing enemy armies, not making friends and favorably influencing people. They are the wrong tool for the job.

I haven’t trusted the results of any poll on any subject in several years. You shouldn’t either.

We need to focus on continuing to train the Iraqis to fight their own war. Unfortunately that is going to take more time, more money, and more casualties. But in my opinion there isn’t another good alternative.

If you want to deny that we need more troops, right now, to secure Baghdad and al Anbar province, I can’t persuade you of anything.

No one, including the 3 generals, wants more troops to run over people with tanks and smash enemy armies. Counterinsurgency methods are quite the opposite, but they are labor-intensive.

(I don’t want to make a personal attack here, and I guess I should charitably assume that you have missed my numerous posts about counterinsurgency techniques and my criticism of “smash ‘em up” generals like Ray Odierno, but … your comment really doesn’t reflect at all what I’ve advocated.)

You don’t believe polls? The consensus of polls done in the week before the Nov. 2004 election were within a point or two of the actual results.

Capt Ed and Powerline have been joined by Atrios in denouncing unfavorable polls, invariably because of perceived sampling bias. (”Not enough Reps/Dems in this sample .. *snort*”) It’s a bogus argument. The pollsters get a sample that is demographically representative (age, sex, geography, etc.) and then ask “self select your party.” Those percentages vary. I know you’re a scientific guy. Check this topic out. Honest, it’s no-brainer.

No, I don’t believe polls. I distrust all polls as a matter of course. I have a passing acquaintance with statistics, and most polls I look at have visible mistakes in either the sampling, the questions asked, or the analysis applied afterwards. Even pollsters who are trying to be honest often get it wrong. When I consider that many pollsters aren’t trying to be honest, I can’t see any reason to believe a word they say.

Consider, for example, a poll that made the news a few weeks ago, about American attitudes toward evolutionary theory as compared to other countries. I took the time to drill down through links at Panda’s Thumb and elsewhere to get at the actual poll questions and answers. What I found was that the poll actually exposed Americans’ ignorance about evolution. The poll questions included two that were directly contradictory — no one should have answered yes to both. At least two percent of the respondents did exactly that. Yet the poll analysis focused on the respondents’ acceptance of evolutionary theory, not their understanding of evolutionary theory. These are two different topics.

I really stopped trusting polls several years ago, when I realized that the standard “President’s approval rating” number that is reported so often in the news is reported in a basically slanted and misleading way. It simply says X% of people disapprove of the president’s performance. It never says why they disapprove. If you listen to mainstream media, you’d get the idea that all of the 60% or so of people who disapprove of Bush’s performance do so because they think the Iraq war was a mistake, and Bush has been too aggressive about fighting terrorism. How many of those 60% disapprove of Bush not because they think he’s been too aggressive against islamist terrorism, but because they think he hasn’t been aggressive enough? You might find that out by looking at raw polling data, but you’ll never find it out by reading mainstream media articles on polls.

The pollsters get a sample that is demographically representative (age, sex, geography, etc.) and then ask “self select your party.” Those percentages vary.

Indeed they do. However, over time the pollsters (if they’re honest) should be able to get a baseline average for party affiliations nationwide and within any specified region and/or demographic. If a poll claims to be nationwide but produces a party-affiliation ratio that differs by five or ten points from the established national averages, then my BS-meter goes off.