If there was some logic to why accept this poll other than “it supports my candidate”, I’d love to hear it. For example, if all of the other polls using D +12 samples and this one uses D +3. Not saying that’s the case, but then I’d be more inclined to believe this LA Times poll.
Their methodology. My logic to follow it is that polling was fairly accurate in 2012 using the same methodology -- while Gallup, which has been doing presidential polling since 1936, threw in the towel after 2012 because they were so far off. (True, the methodology used for USC Dornsife may just have been lucky in 2012. But most of the established pollsters were considerably off.)