Sunday, July 17, 2016

President Obama the Most Important Person in the Next 115 Days

So argue the NBC/WSJ pollsters in releasing their results today.

"President Obama's job-approval rating in the poll is at 51 percent - the third-straight NBC/WSJ poll in which his approval has been above 50 percent."

"Additionally, 49 percent believe the economy has improved during Obama's presidency and they give him credit for it, while another 15 percent acknowledge that it's improved but don't give the president credit."

"Just 36 percent of votes say the economy hasn't improved during Obama's presidency."

"President Obama remains the most important person over the next 115 days," says Hart, the Democratic pollster.

Of course. The rise of Trump is the answer to Obama on the Right. He is the candidate of embattled whiteness.

While you hear a lot about this being an electorate who wants change, Obama's consistently high approval rating in 2016 is the counterargument.

As for the poll itself, things are pretty stable. Despite the hyperventilating over the NYT poll that had them tied this week, today's NBC poll was right in line with where it was last month.

"On the eve of the GOP convention, Hillary Clinton maintains a five-point national lead over Donald Trump even after a period of negative news for the presumptive Democratic nominee, according to a new NBC News/Wall Street Journal poll."

"Clinton leads the presumptive GOP nominee 46 percent to 41 percent in a poll that finds both candidates facing sizable disadvantages and challenges ahead of the party conventions which kick off Monday for the GOP in Cleveland and a week later for Democrats in Philadelphia."

"Those numbers are unchanged from one month ago with the poll conducted after FBI Director James Comey strongly rebuked Clinton over her email practices even though he determined that she had not committed a crime. The poll was also conducted before Trump selected Indiana Gov. Mike Pence as his vice-presidential running mate."

"Clinton has the advantage among African Americans (84 to 7 percent), voters ages 18-29 (55 to 32 percent) and women (52 to 37 percent). The margin among Latinos will be released later on Sunday."

"Trump, meanwhile, is ahead among whites (50 to 37 percent) and men (46 to 39 percent), and the two candidates are tied among independents (36 percent each)."

That's actually not that great for Trump among white voters. He has consistently polled at just around 50 percent among whites. Traditionally the GOP candidate gets close to 6 to 10 white voters. Yet Trump struggles to get 50 of 10.

The 37 percent for Hillary is the same as Obama got-or about what most Democrats since LBJ get. What is different is Trump. He's losing about 10 percent of the white vote that normally is with the GOP.

Not hard to see where it's coming from. Hillary has a shot at being the first Democrat since 1952 to win the college educated white voter. She is crushing him with college educated white women.

"In a four-way ballot test - including Libertarian nominee Gary Johnson and the Green Party's Jill Stein - Clinton gets 41 percent, Trump 35 percent, Johnson 11 percent, and Stein 6 percent."

So in this case the third party candidates take more from Trump than Hillary.

"The more things change, the more they stay the same," says Democratic pollster Peter Hart, whose firm helped conduct this poll with Republican pollster Bill McInturff. "Things haven't changed an awful lot."

McInturff agrees. "Our overall ballot is very stable, and the contours of this election are becoming stable."

The more things change the more they stay the same. A stable race. This is the key to getting to excited over outlier polls. For instance, the big hypefest this week was the NYT poll which showed a tied race. It's possible that showed a short term dip for Hillary after the Emailgate thing. Comey sure went out of his way to damage her with what he said.

But then again, remember that in the NBC poll, there is a margin of error of 3.5 points. This means that Hillary at 46 percent may have as much as 49.5 percent or as little as 42.5 percent. Trump at 41 percent could be as high as 44.5 or as little as 37.5.

In theory this could then be a Hillary lead of 49.5 to 37.5 percent or a Trump lead of 44.5 to 42.5 percent.

In other words it could be as large as a 12 point Hillary lead or it could actually be a Trump 2 point lead. So when you bear that in mind, outliers don't seem so shocking. That tied poll was in the margin for error.

The only thing that should surprise us at all, then, would be a poll that showed Hillary up by more than 12 or Trump up by more than 2.

And, yes there have been a few poll with Trump up by more than 2-but these were Rasmusson so...

The CNN and ABC polls showed the same stable trend.

"Clinton+4, 7, 5 in the two way race for three national polls from ABC/Post, CNN, NBC/WSJ put this morning."

Don't get me wrong: that ABC/Post poll may be focused on by the media because while she leads by 4, she led in the same poll by 12 in June.

So that will be enough for those who want to claim this is more proof she's sinking like a stone. 
But again, recall what we just looked at in terms of MOE. If anything, the 12 point lead she had in June in ABC poll was at the upper edge of what you''d expect if this is in fact about a 5 point race. 
You should expect some polls that show her up by double digits and some that show a tie or even with him having a small lead. 
The CNN and NBC polls on the other hand are little changed since June-which further validate the idea that this is a stable race where Hillary has about a 5 point lead. Remember Obama won by 5 points and this was a near landslide in the electoral college. 
The media, of course, is anti Hillary all the way and will do anything to spin things as negatively for her as it can. Everything for Hillary is glass half empty. She isn't going to be indicted? That's just as bad for her as if she had. 
Obama gives a great speech for her? That's a negative because it diminishes her and reminds everyone how unpopular she is compared with him. 
Just don't get snookered. The media snookers people. It has convinced far too many that Hillary is dishonest though she's been 2016's the most honest candidate and Trump is the most dishonest candidate this year and perhaps ever. Maybe in the 1800s in the age of yellow journalism a candidate was more dishonest, I don't know. 
In 2002 it convinced Americans through its shoddy reporting and it's failure to ask tough questions from the Bush Administration that there was a tie between Saddam Hussein and 9/11 when there was none. 
The media not only fails to tell us accurately what's going on, they actively seek to mislead for their own biased and sloppy ends. 

UPDATE: Interesting point by Mark Murray:

"Rasmussen's entry into the 2016 race has changed the Pollster and RealClear averages as much as email story, IMO."

"You exclude Ras from the polling averages, and Clinton's lead is about 2pts higher."

Noah Rothman:

"If it's a 5-point race, that's not close or competitive. Obama won by nearly 5 points in 2012: 332 to 206 EVs."

Again, a 5 point lead means you should expect to see

1. Polls with Hillary leading by as much as 12

2. Polls with Trump leading by as much as 2. Excepting Rasmusson polls which kind of likes to publish polls that are good news to Trump in case you haven't guessed.


"I don't think ppl get this. This isn't Russia or N Korea. We won't see a huge blowout here. A 5 pts loss is a big deal."

And remember, Obama throughout most of the 2012 race didn't lead by as much as 5 points. So she is in truth ahead of where he was in terms of margin at this stage in 2012. He won by 5 eventually but it was often tied or he was up by just a point or two until September.


Kirsten S. Anderson observes:

"Going from RVs to LVs as a poll screen almost ALWAYS benefits GOP. Yet in ABC/WaPo, it takes Clinton from +4 to +7. …"


  1. One point about margin of error, if a poll says that Hillary is getting 45% with a margin of error of +/- 3%, it doesn't mean that 48% or 42% are equally likely as 45%, it means that you have a decreasing probability the further you get from 45%. So 45% is the MOST likely result to be true, 46% or 44% is still pretty likely, but 48% and 42% are at the very edge of the confidence interval. Normal confidence intervals are set at 95% probability, meaning 95% of the results will be between 48 and 42 for example.

    This is a very commonly misunderstood fact about "margins of error".

  2. Ok, thanks for clarification. I'm certainly not an expert!