It's time for someone to post a rundown of polling methodologies, their strengths and weaknesses.

Some are good at tracking trends over time.  Others are better at producing cross-tabs.  Some are inexplicably wrong all the time while others have a good track record.  RCP just averages them all as far as I can tell.

Sample size?  Response rate?  Survey duration?  Robo-calling versus human?  Different RDD methods?  Cell phone coverage?  Historical accuracy?  Likely voter/caucuser methodologies?  Demographically adjusted versus unadjusted?  

Don't make me surf over to MyDD or a site like that.  Surely the Stumpers or Planksters can boil it down for us so we know which outliers to marvel over and which to ignore.

thanx, Love,

STGLA

Who knows. Maybe the poll is right. Mabe Hillary hit a spot of good media in Iowa during that period. It'll probably fall right back to deadlocked though.

These polls are pointless. The Iowa Caucus can't be polled. Period.

I agree with Richelieu's point.  In Israel polls are meaningless, because nobody takes them seriously.  Everybody makes up answers.  In any given election, the polls before the vote, and exit polls, can all point to victory for party X, but in reality, party Y wins more seats.

Some might see that as evidence of vote-tampering, but Israelis of my acquaintance don't, generally.

The polls do make sense if we don't over-read them. We know with reasonable certainty that Obama, Clinton and Edwards are the first tier of candidates and that the others are pretty much out of it at this moment.

We can also say with a fair degree of certainty that as of now, the only thing that makes sense is to regard Obama, Clinton and Edwards as in a three-way tie in Iowa. This is pretty clear if you look over the past several weeks. Mark Blumentthal "Where things stand in Iowa" (pollster.com) gives a very good explanation of all the polls, their limitations and how to look at them.

If a true trend were to start to develop for any one candidate (which it has not) over the coming days, then it should show up in three (or at least two) of the polls taken and it needs to last more than a couple of days.  Absent such a trend, the best thing to do is to regard the race in the Iowa Caucus right now as a three-way tie.

This does, actually, make sense.  

Hillary by 15 or whatvever: let's keep our fingers crossed that she wins in Iowa.