2016 elections - because it's never too early

[Caliga]

Me to!

[garbon]

Fault the student not the fictional black teacher.

[Valmy]

What? She taught typing not English.

[DGuller]

Fault the student not the fictional black teacher.

I never realized she was fictional.  I thought she was a famous typist.

[garbon]

Fault the student not the fictional black teacher.

I never realized she was fictional.  I thought she was a famous typist.

Because there are so many of those?

[KRonn]

Lots of activity in my town tonight. Trump is landing at the local airport and being hosted at a party/dinner by a local businessman, Ernie Boch who owns several car dealerships in town. Protesters were here early and there was supposed to be a good crowd. I haven't heard/seen what the event was like as I was out talking with neighbors, watching the helicopters hovering over the area waiting for Trump. We didn't see anything much but it was different, the noise of several news helos hovering over the area attracted a lot of attention.

[viper37]

Sometime ago, people laughed at how the candidates would try to outdo themselves, who would come up with the crazyest idea.

today we got Republican candidates proposing:
- tracking immigrants, like we track livestock
- building an anti-immigration wall between Canada and the USA
- And Bush, I think, was the one who said Asian immigrants were the ones with anchor babies.

Your politics are fun

[Admiral Yi]

I've read about Asian anchor babies.

[Phillip V]

I've read about Asian anchor babies.

CHINESE ANCHOR BABIES

"Taking advantage of the system and giving nothing back in return, local unemployed, single mother Mei Xiang reportedly gave birth to two more children out of wedlock this week and continued to pathetically leech off the government."

http://www.theonion.com/article/single-unemployed-mother-leeching-government-51176

[alfred russel]

Sometime ago, people laughed at how the candidates would try to outdo themselves, who would come up with the crazyest idea.

today we got Republican candidates proposing:
- tracking immigrants, like we track livestock
- building an anti-immigration wall between Canada and the USA
- And Bush, I think, was the one who said Asian immigrants were the ones with anchor babies.

Your politics are fun

Scott Walker really had an opening to be a strong conservative with a record of effective union bashing but still a reasonable guy. The whole being against abortion even if the life of the mother is at risk and US Canadian wall has moved him into the realm of just another goofball.

[Admiral Yi]

No, it was somewhere else Phil.

[Berkut]

Sometime ago, people laughed at how the candidates would try to outdo themselves, who would come up with the crazyest idea.

today we got Republican candidates proposing:
- tracking immigrants, like we track livestock
- building an anti-immigration wall between Canada and the USA
- And Bush, I think, was the one who said Asian immigrants were the ones with anchor babies.

Your politics are fun

Scott Walker really had an opening to be a strong conservative with a record of effective union bashing but still a reasonable guy. The whole being against abortion even if the life of the mother is at risk and US Canadian wall has moved him into the realm of just another goofball.

No, that was a symptom of him being a goofball, not the cause.

[jimmy olsen]

Dems better hope the polls are screwy.

http://www.dailykos.com/story/2015/08/30/1416196/-The-2016-presidential-polling-is-deeply-weird-but-does-that-matter

Sun Aug 30, 2015 at 09:29 AM PDT.

The 2016 presidential polling is deeply weird (but does that matter?)


by Steve SingiserFollow
for Daily Kos Elections.

Perhaps predictably, most of the attention being lavished upon the nascent 2016 presidential election is centering on the increasingly entertaining primary contests. This is particularly true as it relates to polling. Nationally and at the state level, pollsters are eager to measure if the Trump-mentum on the GOP side is durable, or if Bernie Sanders is really making a run at Hillary Clinton's frontrunner status on the Democratic side.

Meanwhile, the less-scrutinized polls that look forward to November 2016 are still fascinating. To put it simply: The national numbers, when matched with recent state polling, make pretty damned close to zero sense.

Can Hillary Clinton really be leading national polling by margins akin to Barack Obama's 2008 and 2012 presidential victories, and be losing in states like Michigan and Pennsylvania? Can Virginia be a slightly safer bet for the Democrats than . . . Minnesota?

Simply put, recent polls don't conform to what we think we know about the electoral landscape.

This leaves us with four plausible explanations:
The state polls are wrong
The national polls are wrong
They're both wrong
They're both right

Keep reading as we explore all of these potential answers to a puzzle which, on the surface, doesn't seem to make much sense.

To understand the state versus national polling divergence, let's take Hillary Clinton's average performance in the polls against two foes: Donald Trump and Jeb Bush. These are the two Republicans that, in the last few months, every pollster has been willing to test.

If you take the most recent RCP averages for the matchups between Clinton and the two Republican aspirants in question, this is what you get:


Clinton +4.0 vs. Bush
Clinton +8.8 vs. Trump
As stated in the introduction, those numbers mirror the last two presidential elections, when Barack Obama bested his Republican rivals by margins of 4 to 7 points.
But the state numbers that have come out during the same time frame (post-July 20) bear very little resemblance to the state numbers we saw in either 2008 or 2012.

Take, for example, the state of Michigan. Michigan was solidly in the Obama column in both 2008 and 2012, with his margin of victory twice exceeding his national margins. But last week's Mitchell Research poll in the state shows the leading Democrat trailing both Trump and Bush, albeit by single-point margins.

Is it really possible that Clinton is running anywhere from 5 to 10 points behind her national lead in a state like Michigan? The likelihood of that scenario, it must be said, seems remote.

It seems especially unlikely when comparing those Michigan numbers with the numbers in neighboring Wisconsin. In a poll that was in the field at roughly the same time, Marquette found Clinton leading Bush (47 percent to 42 percent) and easily leading Trump (51 percent to 35 percent). Those numbers square with the numbers from 2008 and 2012, to be sure. But they don't square with the Michigan numbers, since Obama was a couple of points stronger in Michigan than he was in Wisconsin during both of his victorious presidential bids.

It is tempting (and easy) to dismiss this as an apples-to-oranges issue, since we are dealing with pretty disparate polling outfits. Except we are seeing a similar phenomenon with pollsters that have offered both national and state polling.

Take, for example, the past two weeks of Quinnipiac polling. There are few ways to reconcile the national poll that dropped this week, with the state polling that raised so many eyebrows last week.

The national Q poll this past week had Clinton leading Bush by 2 points, and Trump by 4 points. While not the most optimistic data for the Democrats we've seen to date, the results aren't that far from where the other national pollsters have had the race.

But if one looks at their state polls last week, it becomes damned hard to reconcile the two together. How can Jeb Bush be trailing Clinton by 2 nationally, while leading in populous Florida by 11?

Sure, Florida is Bush's home state, and one expects a little home field advantage in presidential elections. But that doesn't explain the Q poll also having Bush leading in Pennsylvania by 3 points. The last time a Republican has run ahead of his national numbers in Pennsylvania, for those keeping score at home, was in 1948.

Even in disparate state polls by a common pollster, oddities abound. In early August, PPP had Clinton leading the Republicans by very modest margins in Minnesota (2 to 5 points). The following week, a PPP survey in the firm's native North Carolina had Clinton within 3 points of both Republicans. It helps to remember that while Democrats have been competitive in the Tar Heel State, they've only carried the state twice in the past half century. Yet there was Clinton within striking distance, dead even with Bush and down just 3 to Trump. Given the presidential trends over the last several cycles, it is hard not to ask: If Clinton is within striking distance in North Carolina, how can Minnesota, which has gone Democratic in every election since 1972, be in peril?

So, yeah, the numbers are a bit strange. What gives?

Possibility No. 1: The Polls Are . . . Right!

In 2012, a number of poll watchers noted a similar dynamic. The national polling showed a tantalizingly close race, one that had optimistic Republicans dreaming of a President Romney. But the state polling, most of it centered on the so-called battlegrounds, did not look quite as pleasant for the red team. Republicans trusted in the national polling—and wound up disappointed on Election Day.

What happened, as several folks (including Markos) predicted, was that Romney was keeping the national numbers close by running up the score in the lightly polled deep red states. Which, in the scope of a presidential contest, means precisely jack shit. If Obama is sweeping virtually all of the swing states, who gives a damn whether he is losing Tennessee and Arkansas by 12 points, or 22?

It is not entirely implausible that, given the relative paucity of statewide polling to date, a similar dynamic is taking place here. How can Hillary Clinton be ahead nationally by margins similar to what Obama won with in 2008 and 2012, but struggling in some places where Obama did well? For one thing, it is not unthinkable that Obama did so unnaturally bad in certain corners of the South that a Democratic nominee with the surname Clinton could do considerably better.

Again, this is cold comfort to Clinton. If she loses Kentucky by 9 points instead of 21, that doesn't mean anything as it relates to her prospects for election. If that is the reason for the state versus national poll oddities, it is entirely possible that Clinton could be close to where Obama was in 2012, but in a considerably more perilous position vis-à-vis the electoral college.

Possibility No. 2: The Polls Are . . . Wrong! (All of 'Em)

For several years, political junkies have debated whether or not the political polling business is on the losing end of several trends, which could be leading to less accurate results. The most recent criticism of polling came in June, courtesy of Rutgers professor Cliff Zukin, a former president of the American Association for Political Opinion Research (AAPOR). In his piece, Zukin underscored that even those on the inside of the industry recognize the problematic new world for pollsters:

Election polling is in near crisis, and we pollsters know. Two trends are driving the increasing unreliability of election and other polling in the United States: the growth of cellphones and the decline in people willing to answer surveys. Coupled, they have made high-quality research much more expensive to do, so there is less of it. This has opened the door for less scientifically based, less well-tested techniques. To top it off, a perennial election polling problem, how to identify "likely voters," has become even thornier.

In a piece authored just before the 2014 elections, Nate Silver noted that between 1998 and 2012 the accuracy of political polling had been fairly static. If anything, it might've improved slightly in most cases. However, even as his data noted no catastrophic collapse in polling accuracy, Silver warned that this didn't mean there wasn't some peril:

How can a poll come close to the outcome when so few people respond to it? One way is through extremely heavy demographic weighting. Some of these polls are more like polling-flavored statistical models than true surveys of public opinion. But when the assumptions in the model are wrong, the results can turn bad in a hurry. (To take one example, the automated polling firm Rasmussen Reports got fairly good results from 2004 through 2008, but has been extremely inaccurate since.)

If Silver is right, the nightmare scenario for poll followers is that any election could be the one where the assumptions are wrong, and the results become a shitshow in no time flat. Conservatives would argue that we saw something to that effect last year, though it is worth noting that our own Daily Kos Election Outlook, a poll-based model created by Drew Linzer, was right on the fairway in forecasting the Democrats' crappy election night. So, the polls might've undersold the Republican performance, but it wasn't like nobody saw it coming.
All that said, when it comes to presidential elections, being off by a few points could mean a heck of a lot more. And the threat of the "assumptions being wrong" is quite real. Could this be the cycle in which that happens? It is certainly an—ahem—unusual cycle to date. If ever there was a series of variables that could make an election challenging to forecast, this would be it.

Possibility No. 3: The State Polls Are Wrong

This is the possibility that the Democrats, of course, hope is the correct one, given that their candidates seem to do considerably better in national polling than in most of the recent state polls. The possible exception is Vice President Joe Biden, if you put stock in Quinnipiac.

Certainly, this possibility is largely rooted in the identity of two of the pollsters giving the best numbers for the GOP. I took Quinnipiac to task just last month for swing state numbers (in that case in Iowa, Colorado, and Virginia) that seemed a little wonky. Others, in political conversation and in the Twitterverse, were quick to point out that the racial and partisan makeup of the Ohio, Florida, and Pennsylvania polling seemed to reflect more of a 2014 midterm turnout than a typical presidential one. Likewise, with some notable exceptions, Mitchell Research has been more likely than not to forecast rosy outcomes for the GOP in Michigan.

Of course, the danger there is that this theory is based on the faith that Quinnipiac's assumptions are wrong, but that the national outfits (which include firms like CNN/ORC and ABC's polling unit) are correctly forecasting the composition of the electorate. The graveyards, of course, are full of election prognosticators convinced that the polls were off the fairway.

Possibility No. 4: The National Polls Are Wrong

Republicans, conversely, would love to endorse this view. And it is a theory that has some plausibility, given an intriguing quirk in the polling this year. After all, how could Hillary Clinton have a national polling lead (and, as our own David Nir pointed out last week, a pretty stable one), given that her favorability has been sinking palpably all year?

Of course, experienced election observers already know the answer to that question. It has happened in both of the GOP's recent midterm wave elections. In both 2010 and 2014, the Republican Party's favorability ratings were absolutely for shit. In the final analysis, however, those voters (and there were plenty of them) that disliked both parties went largely in one direction.

Another somewhat reasonable argument for assuming the national polls are an issue is the relative lack of consistency in the national numbers. Observers looking at the recent Marist or CNN numbers, for example, would feel pretty good about the chances for the Democrats to continue their hold on the White House. But the Quinnipiac and Fox News polling seem much more equivocal about Democratic prospects.

Of course, there is always the very real prospect that the truth is found in...

Possibility No. 5: We still have no freaking clue!

There are two important details that fortify this final possibility. For one thing, it is over 14 months out from Election Day, and this thing is going to twist and turn a couple of hundred more times. For another, the "sample size" of polls is still incredibly small. A random day in late October 2016 will in all probability see more polls than we've seen over the past three months.

If, however, the national polls still make little sense when married to the state polling a year from now, this piece may be worth revisiting. In that scenario, it would be far more fair to suggest that something has gone quite a ways off the tracks.

[Admiral Yi]

Timmy, the medium of the internet is simply not well suited for the dissemination of information in chunks that large.

[jimmy olsen]

Timmy, the medium of the internet is simply not well suited for the dissemination of information in chunks that large.

That's 2,000 words, max. That's nothing. 

I read entire novels via kindle PC.