USC Dornsife / LA Times Presidential Election "Daybreak" Poll

http://cesrusc.org/election/ ^ | 8/24/2016 | LA Times

[FirstFlaBn]

We’re good. Nothing at all wrong with your challenge. It gave me an opportunity to do a better job of explaining my point. The site is improved with your participation.

I worked really hard in my elementary statistics class and learned a lot. I’ve even used some of it in the real world! Could you answer one question for me?

Am I correct in thinking MOE applies to each result (and not to the difference between them)? To simplify, if a random sample from a large population of red and blue balls came out 52% red and 48% blue with MOE of 3, wouldn’t that be a statistical tie?

One bonus question - isn’t the 52-48% result reported that way for convenience and that all values within the MOE are equally likely? IOW, in the red/blue example, aren’t 49% and 55% just as likely as the reported midpoint of 52% for red?

If I’m correct on both, then all can see the futilty of pundits proclaiming “Oh, he moved up 2 points this week!”

Please straighten me out if I’m in the ditch.

[mac_truck]

If the enthusiasm meets registration, polling on 2012 numbers is meaningless.

In that case the 2012 numbers become a floor which is why having Trump -2 at this point is really no big deal.

I like this poll because its different and (to my eye at least) it doesn't seem like the LA Times is interfering with the methodology.

[AngelesCrestHighway]

It’s all bullcrap!

[Albion Wilde]

Ping

[enumerated]

They have trump back up by .7 as of the latest. Well within the moe, as you said, but for some reason, it still comforts me to see a shift in the right direction.

If for no other reason, I like a little ammunition to counter the perception being pushed across the board that Clinton is ahead and Trump may as well throw in the towel.