Perhaps that earlier Selzer poll and Quinnipiac were outliers, because I can't imagine things have improved this quickly for Bruce Braley.
Things didn't improve this much for Braley, even though I think he might be gaining in this race very slowly.
The big reason for the difference in the 2 polls is the likely voter. This is all because of Iowa having early voting.
A pollster can't kick a low interest left leaning voter out of the likely voting pool if they say the already voted. She even admits this in the write up for the recent poll.
I would bet this makes up 2-3% of Braley's gains between the 2 polls.