I won't touch 1948 because I have no idea how accurate polling was overall. However, Carter was not polling over 45% at any point post-convention, nor was Reagan ever behind by 10-15 points (in fact Clinton leads where Carter was at that point by 10-15.)
As I understand it, Gallup stopped polling week(s) before the election and missed subsequent changes. Though, this doesn't touch on any potential issues with their methodology.
I think he can, but I certainly don't think he will. So far he seems to be a terrible general election candidate, and it's hard to see that changing in the next 90 days. It also seems impossible for him to make up the advantage Clinton will have in advertising and the ground game. Even if he does somehow fix those disadvantages, he'll have a tough time getting to 270; it's hard to see Pennsylvania, New Hampshire, Colorado, Virginia, Wisconsin, Michigan, Minnesota, or Oregon voting for him. It also seems that presidential elections rarely change dramatically after the conventions.
Agreed. Trump has given no one any reason to believe he'll turn it all around at this point. Ideally, he should have run a tight ship post-convention, when the reality is Trump's campaign is currently at one of the lowest points
(if not thee lowest) that they have been in the entire election cycle, all due to unforced errors from Trump and his big mouth. So can he technically bounce back? Sure. Is it likely? Not at all.
