I sent the following suggestions to Richard Morin, one of the authors of the WP poll. The question numbering is that of the poll data document linked in Mike's article.

(a) Questions 13 and 14 are not rationally paired. As I wrote in my letter [to the editor], "There is no contradiction between wanting the troops to come home now and wanting troop levels to be adequate." The linking of 13 and 14 may therefore underestimate (or, conceivably, overestimate) the real level of support for immediate withdrawal. Questions 7 and 14 would make a more rational pairing, since they are both about withdrawal, whereas question 13 is about troop levels.
(b) Question 7 is not a neutral measure of support for withdrawal vs. staying in, because it argumentatively links US military presence in Iraq with civil order in Iraq. It assumes, and encourages respondents to assume, that
(i) civil order can be achieved with US military presence
(ii) civil order can only be achieved with US military presence
(iii) US military presence is conducive to civil order
The part about civil order should be removed from the phraseology, as it tends to bias respondents toward saying "stay," at least those who care about Iraqi civil order.
(c) A question about the relationship between civil order and US military presence should be added, modeled on propositions (b)(i) - (b)(iii) above. Then question 10, which is flawed in a similar way to question 7, would no longer be needed.