Interesting - the only change between this poll and the previous one is in the likely voter model.
I wouldn't count Wilson out. She, like Anne Northrup, are always on the most endangered incumbent lists, but somehow manage to claw out wins.
I'm the guy who did this poll and wrote the text reproduced in the above post. The "likely voter model" did not change. It was exactly the same model as we used in the previous poll. We do not "weight to party", we rely on self-described party affiliation (NOT registration), and we saw that more people were identifying themselves as Democrats and fewer as Republicans. It's not a "model", or something we put in to the data -- it's what we observed while using the exact same methodology as before.
Not good news for the GOP I'm afraid, but there's still plenty of time for things to change.
True...
(Northup, not Northrup...)