Over at Pollster.com, they do an excellent job of keeping track of pretty much every national poll thats out there - and analysing 'em.
One thing that they'll warn against time and time again is not to read too much in any one single poll. Margins of error and the like mean that any one poll can swerve up and down quite markedly without statistical significance, and differences between pollsters mean that on top of that, there can be conspicuous differences between the results of different pollsters - and then you have the occasional outlier.
So what they strongly suggest is to look at the average of the last X polls out. They track exactly that, charting out a very cautious trend estimate.
In this image, the lines represent the trend estimate; whereas every individual dot represents one poll result. (And just look at how large the variation from one individual poll to another is..)
(Note that the graph was last updated three days ago.)
As you can see, none of the three major candidates have made much headway in the last three months or so. Hillary is slipping somewhat; Obama and Edwards have stagnated.
Notice though that the trend estimate was deliberately made very, very cautious - there really need to be a lot of new polls out showing a marked change for the trend estimate to drop or surge markedly.
If you look at the dots (you can click the image for a larger version), you can that see that a more sensitive estimate would, for example, have shown Edwards really jumping around March-April, when the bad news about his wife's cancer came out; and that he has appears by now to have fallen further than the cautious trend estimate shows yet.
Then, of course, there is the curious matter of Al Gore actually steadily rising in the polls - despite not making any signs of running.
This is the analysis that went with it:
link
Quote:<snip>
The Democratic frontrunners haven't suffered the sharp declines that Giuliani and McCain have, but none are showing strong positive gains either. Clinton has fallen off just a few points recently, Obama seems stalled and Edwards has a very slight decline in support. So who do Democrats like more and more? Al Gore, the non-candidate.
The Gore increase is no where close to that of Fred Thompson, but of the four possible nominees pictured above, he is the only one with steady gains throughout 2007. Given the low level of encouragement Gore has given to a possible candidacy (FAR less than Thompson) it is remarkable that he's moved up at all. And while the Thompson candidacy looks increasingly likely, a Gore campaign seems a remote possibility to me at least. Nonetheless, a non-trivial number of Democrats are looking longingly at him while passing up the easy opportunity to support Clinton or Obama or Edwards. Clearly they are looking for someone else who can take this opportunity to exploit the moment. Other evidence (here and here) makes Gore seem an unlikely white knight. In terms of partisan feelings, polarization and support in a general election, Gore looks a lot like Clinton-- well known and well liked among Democrats but not very popular among independents and actively despised among Republicans. But that isn't the point here. He is "someone else" at the moment within the Democratic party.
The front-runners have won substantial support within their parties and one may yet go on to win. But the current flatness or decline in their support trends argues strongly that none have sealed the deal with their primary voters. The widespread public disaffection with current leaders and conditions can be seen as a difficult environment to run in. But it is also the great opportunity to be seized by an able politician, one who can convince supporters that they have a vision of how to lead the country out of these bad times and into a new "morning in America". Based on the evidence here, I don't think any of the top six candidates has managed that yet. And that leaves them all vulnerable to someone who can. Fred Thompson is evidence that it is not yet too late for such candidates to emerge.