State polls lag behind national ones?

[geek tragedy]

The national polls showed Bush's bounce while the state ones showed Kerry still leading. Now the national polls are showing the race tightening while the state polls are showing bad results.

My theory? Bush's convention bounce is just now registering in the state polls, and in a week or two the state polling will bounce back to sanity, especially in states like WI, NJ, OH.

Just a hunch--can't back it up with statistical analysis.

Thoughts?

[cthrumatrix]

1. it means they cannot be 100% trusted...get people registered and

tell truth about awol to swing voters

[bunkerbuster1]

2. Several thoughts about by-state polls/results reported

1. There's way more media attention, way earlier, on these state-by-state races than there was back in the Olden Days (i.e., before Nov. 2000). All of us amatuer pundits have little historic background on the volatility of these polls.

2. Rassmussen has an interesting comment here at

http://www.rasmussenreports.com/MembersOnly/ScottPage.htm

and I quote (don't necessarily agree 100%, but worth reading):

Many polling firms are releasing "Battleground State" polling data. This makes sense given the closeness of the race. However, all Battleground State data should be viewed with a bit of caution. Here's a few things to watch out for...

1) Sample Size--Just under one-third of the population lives in the 16-States commonly viewed as Battlegrounds. This means that a survey of 1,000 Likely Voters will have roughly 300 respondents in the Battleground States. Smaller surveys have even fewer Battleground State respondents. Needless to say, such small samples lead to a pretty healthy margin of error. This is not to say that the data is bad, but simply a warning to avoid reading too much into it. At Rasmussen Reports, our three-day samples include just over 900 Battleground State interviews, but we still prefer to review the 7-day data for a more stable sample (over 2,100 respondents).

2) Shifting Battlegrounds--While everyone pretty much agreed on the Battleground States when the year began, there is some evidence that the Battlegrounds are shifting. Some states (i.e.--New Jersey) were thought to be safely in one camp or the other when the year began, but now may be a bit more competitive.

3) All changes are not equal--At the moment, Kerry appears to be outperforming Gore in Washington and Oregon. While that's good for Kerry, it doesn't secure any additional Electoral Votes for the Democrats. On the other hand, Kerry is doing better than Gore in Ohio, but lagging in Wisconsin, Minnesota, and Pennsylvania. Those changes could be VERY significant.

4) In the end, the best indicator to watch is still the national Tracking Poll. If either candidate can open up a sustainable lead in the popular vote (beyond a point or two), that candidate will win the Electoral College.