That poll was statistically worthless because of the "special sauce" the LAT polling staff added.
Why did they add the special sauce?
Because they found out they had a bad sample to begin with.
In my statistics class, I was told that if you started with a bad sample, you were going to get bad results, no matter how much you fiddled with them.
So that poll is completely worthless.
I don't think the LAT added any "special sauce" to a SurveyUSA poll. (The LAT poll had McClintock at 12%.)