Polls are supposed to predict election outcomes. But, as everyone knows, they can also affect the outcome. This is especially the case in multi-candidate primaries or caucuses. For one thing, some voters don't want to "waste" their votes on laggards. For another, some voters who are more-or-less undecided among two or more candidates (and there are many such in primaries, since candidates within a party often don't differ that much on issues) can see a candidate pick up momentum, and that can push them off the fence or strengthen an inclination to jump in a certain way. The final Des Moines Register poll is supposed to appear on line tonight, and in the paper tomorrow morning. The last DMR poll appeared a month ago, and had Huckabee at 29 and Romney at 24, followed by Giuliani at 13, Thompson at 9, and McCain and Paul at 7. What poll results tomorrow morning could affect the outcome Thursday night? Two results in particular could be interesting. If Romney has caught up to or passed Huckabee, that would presumably legitimize, so to speak, further defections from Huckabee waverers. And if McCain has moved into third ahead of Giuliani and Thompson, that could give him more momentum and enable him to secure third in Iowa--and perhaps even to pull closer to second (especially if a new New Hampshire poll out tonight shows, as is rumored, McCain leading in New Hampshire). Or perhaps the DMR poll won't matter at all.