Their methodology is making it a consistent outlier and may be embarrassing them, but they figure they'll learn if they should have tossed it in November.

Lots of people have questions about the USC/L.A. Times tracking poll; here are some answers
http://www.latimes.com/politics/la-na-pol-poll-faq-20161006-snap-story.html

to vote for each of the two major candidates or for some other candidate.

What the heck does that mean anyway? It is too subjective

Not unlike the medical profession asking on a level of 0 to 10, how much pain are you experiencing.

As when my Dad was asked that question, his reply was, I wouldn't be here if I wasn't in pain.

and am not convinced of the value of continuing with this model. As you say--doctors never explain what 10 pain might be, and the range of answers is very silly; but they do it because even sometimes very unreliable answers that have to be assessed against other observations are better than not asking the question. In this case, the pollsters are obtaining information resulting from different methodology, and it's apparently of value to some for that.

Until and if there's a horrible November surprise, I'm not among them either. I just have a problem with those few who call for keeping disagreeable results off DU. Yes, if they're from a poll designed specifically to fool people into believing what is not true. But that's not what USC/LA Times is doing.

But the best measure is accuracy predicting elections. And we don't have data for this yet for this poll.

Nobody should pay attention to this fraud poll.

this poll out. It is an insult to the legitimate pollsters, and the intelligence of the American people

from the start, and instead of correcting the flaws, they keep reporting the same bad data.

Garbage in Garbage out

Measured take. As always, cant use just one and toss it away because we don't like the results.

http://fivethirtyeight.com/features/election-update-leave-the-la-times-poll-alone/

A daily tracking poll is great in that it measures the same group of people to see what changes. The problem is that it doesn't re-sample so any skewness in the sample will remain constant.

embarrass themselves by using the same flawed modeling and sampling, instead of making the appropriate corrections.

Unless they provide him with evidence that they are doing a real poll with proper methodology.

Give it a week and even this poll will be Hillary +10. No need to fret over polls anymore, the race is over. The only question is how much the Dems can capitalize down ticket on this GOP calamity.