Eh, those polls aren't laughable. Let's say that the next QPac poll showed Hillary up again -- that doesn't mean that they did something wrong. By the way polling works, you'll see outliers, you'll see polls that go against the aggregate.
To liberals: If there is a trend, don't ignore it for that one poll that is on your side, ala 2014. Ignoring trends and polling realities or betting solely on ground game is how reactions to 2014 happened.
To conservatives: Don't forget 2012 and ignoring polling in the aggregate. Latching on to one poll or one polling company, like you are on twitter right now, is repeating the exact same mistakes that you made in 2012.
Even mathematically there's a 1 in 20 chance that a poll will go outside the margin of error, and realistically it's higher than that. Given the number of polls we have for these things, polls like that are always going to happen.
And it seems the media's always going to hype those obvious outliers either way, but whatever, they can have their fun before it's all forgotten in a couple of days.
Yes, 1 in 20 polls should be an outlier if they're doing everything right. But an outlier means more than 2 standard deviations away (or outside the 95% confidence interval).
That doesn't mean an outlier should be 5 times away from the standard deviation, however. An outlier like that means you're doing something wrong.
For instance, let's say in Virginia, the true current polling should be Hillary 47, GOP candidate 42.
That means a true outlier could be something like Hillary 53, GOP 35.
Another one could be Hillary 43, GOP 45.
What it does not mean is that it should turn up Hillary 36, GOP 47. That is bonkers. That only comes from a mistake. And we're about to see why.
From their sampling information:
Colorado
PARTY IDENTIFICATION
Republican 29%
Democrat 26
Independent 36
Other/DK/NA 10
RACE
White 79%
Black 4
Hispanic 11
Other/DK/NA 6
Iowa
PARTY IDENTIFICATION
Republican 29%
Democrat 27
Independent 38
Other/DK/NA 6
RACE
White 91%
Black 2
Hispanic 2
Virginia
PARTY IDENTIFICATION
Republican 27%
Democrat 28
Independent 35
Other/DK/NA 9
Other/DK/NA 4
RACE
White 68%
Black 19
Hispanic 4
Other/DK/NA 9
I don't know the exact demographics or partisan breakdown for all of those states, but the first two look off to me.
Still worrying though.
CO was +5D in 2012. Quinnipiac is polling it as +3R.
So either it's a junk outlier poll or the party registration has suddenly drastically changed and any Democrat is fucked there in 2016. I'm going to guess it's the former.
It's fucking junk.
Here's the thing. In Co, it was +5D among ACTUAL voters. But this is not a poll of actual voters (of course) or even likely voters. Remember, the polling from the last election was close because the likely voter models were depressing Democrat ID because it assumed more would stay home than did.
But the above polling is neither. It's a REGISTERED VOTER poll. In most states, the RV is more favorable to Ds than the LVs.
So how can a state that clearly has more Democrats as registered voters in 2012 (something probably like +8-10D) all of a sudden become +3R? It's nonsense. It defies all logic. It would require party shifts we've never seen since the Civil Rights stuff. Sorry, but this isn't happening.
If the above poll was a Likely Voter poll (something kind of stupid this far out) then maybe it would be cause for concern. At least it would be a guess on what the actual electorate will look like. But it's not, it's a measure of registered voters and we are almost certain they're wrong.
Unless young people are registering as republicans and people are switching parties by significant margins in the last 3 years, you can wipe your ass with this polling.
Note: I am not trying to argue Hillary is ahead by a lot or that she may not be in trouble. But this poll is junk for the similar reasons I argued the Gallop polls were junk in 2012. If you recall, I didn't believe those polls were outliers but rather garbage because it oversampled the South by wide margins. Outliers happen 5% of the time if you do polling CORRECTLY. If you fuck it up, you're not going to get just outliers, you're going to get bad outputs.
On a side note, kudos to Lindsey Graham for a good video response to Trump. I found it funny.
