Bradley effect/Shy Tory effect??
A sizeable amount of people tend to report to pollsters the more popular answer when polled which might of made LBJ looked better in the polls than in reality despite polling upwards of 70% at one point.
Good question. He could have done better in terms of the electoral votes; as you pointed out, AZ was very close. 486-52 is actually weak compared to other landslides; Nixon in 1972 and FDR in 1936 both breaked 60%, but won 520-13 and 523-8. Reagan outperformed it twice, although he was more than ten points short in 1980 (489 EVs and 50.7% vs. LbJ's 486 EVs and 61.1%). Even John Anderson took more votes from him than from Carter.
GA and SC were winable for LBJ and maybe LA. But he was not on the ballot in the latter, for whatever reason (just "unpledged Democratic voters").
As far as the PV is concerned, he could have done better with RFK as his running mate. But both didn't like each other and LBJ wanted to prove that he is able to win an election without a Kennedy on the ticket. However, I think that HHH was a smart choice for the ticket balance.
I always wondered why he didn't break 60% in California (he got 59.1%).
Maybe this would have been possible with RFK:
President Lyndon B. Johnson (D-TX)/Attorney General Robert F. Kennedy (D-MA): 511 EV. (62.7%)
Senator Barry Goldwater (R-AZ)/Congressman William Miller (R-NY): 27 EV. (36.7%)
SC was going to go to Goldwater whether of not the ticket was different, as long as LBJ signed the Civil Rights Act. Goldwater won that state by 59% so I doubt it was winnable for LBJ. It would also be hopeless for the rest of the Deep South states.
However if LBJ did pour more resources into Arizona, he would of been able to beat Goldwater's home state advantage.