I might argue the point on Zogby...
There are several much better polling firms...
Regarding a "safe" margin...
In 2000, if Gore had 1% more popular vote in every state (and Bush 1% less) the electoral college would have been. 296/242 for Gore.
A shift of 1.0% for Bush would have made it 327/211 for Bush
A shift of 3% for Gore would have made it also 327/211 - except this time for Gore
A shift of 3% towards Bush would have made it 366/172 for Bush
a gap of 3% or more it is almost inconcievable that the EV and popular vote would not agree
Didn't Zogby come closet in the past 2 presidential elections? I visited his site during the Dem primaries and he basically seemed to nail those too, with a few exceptions. Which firms would you consider better?
So going by what you say, a Gore PV win of 1.5% would have given him a 296-242 win in the EV? It would seem to me, just based on intuition, that it would be extremely unlikely for someone to win the PV by over 2% yet lose the EV. From 1-2%, I think with the right combination of states, it is possible but not probable. Under 1% it's quite probable.
I have a feeling that in both the EV and PV this election isn't going to be as close as 2000. The problem is I'm just not sure who's going to be ahead.