>>2996038
Florida
https://www.realclearpolitics.com/epolls/2018/senate/fl/florida_senate_scott_vs_nelson-6246.html
Here we see the very same thing going on we saw in Arizona! We have 7 different polls done, but they DOGGEDLY hold on to the one from 6/17 that shows Nelson at +4, and it is the ONLY one with Nelson in the lead. This gives us a +1.7 RCP average, keeping it in the "Toss Up" category.
Let's examine that one, shall we?
http://media1.s-nbcnews.com/i/today/z_creative/Toss%20Up%20Questions_NBC%20News_Marist%20Poll_Florida%20Annotated%20Questionnaire_June%202018.pdf
In this poll, we see that they polled 344 registered Democrats, but only 326 registered Republicans. This was a HUGE tactic used in 2016: Oversampling Democrats. Still, there are only 670 of the 947 polled who are registered as Democrat or Republican. Are there REALLY 277 registered Independant people in this poll? That's quite a lot!
Either way, oversampling and a non-representative sample of registered voters points to there being fuckery afoot here as well.
Looking at the other polling, we show a CLEAR favorite in Rick Scott.
>SurveyUSA 9/7 - 9/9 634 LV 5.3 46 44 Scott +2
>Quinnipiac 8/30 - 9/3 785 LV 4.3 49 49 Tie
>Gravis 8/29 - 8/30 1225 LV 2.8 47 47 Tie
>Florida Atlantic University 8/16 - 8/20 800 RV 3.4 45 39 Scott +6
>Mason-Dixon 7/24 - 7/25 625 LV 4.0 47 44 Scott +3
>CBS News/YouGov 6/19 - 6/22 839 LV – 46 41 Scott +5
>NBC News/Marist 6/17 - 6/21 947 RV 3.9 45 49 Nelson +4
+2, Tie, Tie, +6, +3, and +5…
Let's look at the two "Tie" ones.
First, the Gravis one:
https://www.realclearpolitics.com/docs/Florida_(August_30_2018).pdf
This poll seems pretty much on the "good" side of polling… 12% of those polled are hispanic or have a member of their immediate family as being Latino, Hispanic, or Spanish Speaking background… 69% of those polled are white… There is no obvious over-sampling of Democrats… but again 33% are independent, which is INCREDIBLE to me that there are so many who claim that.
77% are VERY likely to vote, with only 9% being "Somewhat Likely"…
A big key in this one is to look at some of the other ideologically driven questions… 49% say Kavanaugh is qualified for the USSC. 43% support Brennan's security clearance being revoked (42% oppose)… 39% are more likely to vote for a candidate that supports Kavanaugh's appointment to the USSC (35% would be less likely), so that means Nelson is being given a clear indication: If you want to win this election, you had BETTER vote for Kavanaugh!
Looking at Question 23, the hatered for the MSM is obvious. 49% say they Disapprove of the MSM, with only 41% saying they approve… but look at that 36% that STRONGLY disapprove!
And 51% approve of the border wall!
Moving on to the OTHER "Tie" poll…
https://poll.qu.edu/florida/release-detail?ReleaseID=2566
First question I always ask myself? What is the breakdown of those polled… NOT THE PERCENTAGES, BUT THE RAW NUMBERS! They do not include that in this "press release" version, so you have to go to the "sample and methodology detail" link…
https://poll.qu.edu/images/polling/fl/fl09052018_demos_fjya13.pdf/
785 total people polled. 34% GOP, 31% Dem… so good job not oversampling DEMs, Quinnipiac!
Anyway, I can find nothing really wrong with either of these "tie" polls to show any sort of bias…
The bottom line is this… the +4 Nelson poll from 6/17 is bullshit. The RCP average of +1.7 for Scott is thus bullshit.
Without that +4 to Nelson, the RCP average shifts to +2.7 Scott. Still within the MOE of most polling, but definitely a huge swing towards calling this state for the GOP.
I'm going to call it for Scott now.
NEXT!