Sunday, February 21, 2016

Do three people know less than one (about Nevada's Latino Voters)

While reading this, spend a minute listening to one of the greatest Car Talk talk bits.

It will explain how bad political memes (and other dumb ideas) happen.

If you want to really get into that, here is the letter mentioned in the above one minute bit.

Since I am about to tread on sacred ground, let me first hedge by saying that on average when Ezra Klein, Nate Silver and Nate Cohn agree on something it is more likely to be accurate than 3 randomly chosen pundits. Sometimes by a wide margin.

But their notion about Bernie Sanders not winning the Latino vote in NV started with one of them not knowing what he was talking about and then another (cue the Car Talk bit above) and ....

It started when Nate Cohn wrote too much with too little time to research over 2 hours of tweets during NV election results day using simple back-of-a-napkin-sized math.

This nonsense was then blessed by the other two (along with an article of
IF you assume A, B, C, D, E, F...
THEN our theory is correct).

To go point by point on that article to detail in depth exactly why each of its assumptions are suspect is beyond the scope of this blog post (and my currently available time).

That article supposedly points to this NY Times analysis of Nate's, but the included URL just goes to a page showing the actual total delegate counts at the time. So I have included all the relevant tweets in this.

That other well known analysts in that tweet conversation like pollster/poll analyst Mark Blumenthal and Harry Enten and a few unknown data scientists (but one data scientist worked on the '12 Obama team) were skeptical, is never addressed by the trio (or even mentioned in the Vox article above). And since Ezra and the Nates carry such weight in the political pundocracy, you can count on this bad idea spreading like wildfire - starting with the Sunday talk shows and then really raging by Morning Joe early Monday.

I'm another unknown data scientist, but I've spent 3 decades analyzing/forecasting data and am currently working towards a master's at Harvard in my field to get some name-brand credentials. But credentials shouldn't matter. Common sense should suffice.

I'll end with one last ancient pearl of wisdom:

Occam's Razor

Among competing hypotheses, the one with the fewest assumptions should be selected.

A bunch of people (entrance pollsters) tried to randomly select voters at caucuses.
The lengthy arguments in the Vox article and Nate Cohn's Tweets

You decide which has the fewest assumptions.

And then if you still feel that Latino's can't possibly favor Bernie Sanders, please explain how the complex set of motivations all kinds of different people have when voting, can be boiled down to
  1. what race you are 
  2. that a Senator spent most of his life not living near people of your race (neither did the other Senator, but I digress)
And please do so with the smallest number of credible assumptions.

P.S. Yes I support Bernie Sanders, but I don't support bad math, bad statistics or bad data science.

Sunday, February 14, 2016

Rare South Carolina Primary Poll has Suspect Demographics

With so few current polls in South Carolina, looking in detail at these polls is even more important.

The demographic breakdown in the YouGov/CBS poll looks suspect.
The % of Democratic pollees that are black is 62.5%. In 2008 when there was a black candidate running it was only 55%. The racial composition of the state hasn't changed since 2008 so why would a race feature two white candidates have a surge in black voters? if anything, one would expect the black vote to be less than 2008.

Even if the demographics were like the extraordinary 2008 primary, the 73% of black voters for Clinton would suggest a 6 point reduction in Clinton's lead. A 56/43 vote split. More likely the demographics are not so favorable as 2008, meaning an even closer race.

Opposition Party Senate NEVER blocked an Election Year Supreme Court Nomination

The 80 year myth circulated by Senate Majority Leader McConnell, Senate Judiciary Chairman Grassley and every Presidential GOP candidate is easily dispensed with:

Contrary to the GOP Senate Majority Leader's claims that the Senate has not confirmed a Supreme Court Justice during an election year in over 80 years, that same person confirmed the currently serving Justice Kennedy in an election year (1988) and Justice Murphy was confirmed in 1940.

In the Supreme Court's history, 21 people have been nominated in an election year.

However, their obstruction based on fallacy is worse than that.

In only 4 years was someone appointed but not confirmed to the Supreme Court in an election year.

Those were all special cases.

In every case, the Senate was controlled by the same party as the President.

So an opposition party Senate has NEVER blocked a Supreme Court nominee in an election year. Not once in the 228 year history of the world's longest democratic constitution.

In all the 4 years above, the President was unpopular within his own party.

In 1968, the former Senate Majority Leader, President Johnson was at war with a large faction of his party. He was so unpopular as president that he decided he couldn't win his own party's nomination and withdrew. Sitting Supreme Court Justice Abe Fortas was nominated to take over for retiring Chief Justice Warren. Fortas had angered this faction of the president's party both because of his close association with he president and his own dubious acts. More details are here.

In 1852, the President was so unpopular he was not nominated for reelection by his own party. Like the case in 1844 the President had only become President by being VP when the President died and failed to win an election in his own right.

By 1844, the President had been thrown out of his own party and rejected all 7 of his nominations while the President continued the spat by renominating already rejected nominees multiple times. More details on this are here.

In 1824, President John Quincy Adams was chosen as part of the Corrupt Bargain by the House of Representatives despite the President losing to Andrew Jackson by 11% in the popular vote (no candidate had reached 50%). Both were members of the same party. The corrupt bargain poisoned his administration. A couple weeks after Andrew Jackson beat him in the 1828 election, John Quincy Adams nominated a Supreme Court justice, who was not confirmed and 22 days later Andrew Jackson nominated his own candidate.  More details are here

Thursday, February 11, 2016

Data on Why Bernie Sanders Will Win in Heavily Non-White States

If you believed many pundits and people from the Hillary Clinton campaign, she has 80% support among blacks (another poll with similar info) and therefore Bernie Sanders is doomed.

But these are old polls (August and November 2015).

And they are no surprise.

The first 11 months of 2015 the coverage on CBS/NBC/ABC national news:
Candidate Minutes
Clinton 113
Sanders 10

For years Hillary Clinton has had near 100% name recognition. 4 months after Sanders announced his campaign, 38% still knew too little about him to even give him a favorability rating.

We need CURRENT polls from Nevada and South Carolina.

Here is why there will be none in Nevada. There are similar reasons why you may get few to no polls for South Carolina.

(Feb 12 EDIT: One Nevada poll (from the GOP) just published. It shows Bernie TIED with Hillary)

There is other useful.

Look at the recent election exit polls of New Hampshire (and to a lesser extent Iowa).

In the New Hampshire Primary which Bernie Sanders won 60% to 38%, CBS exit polls showed
Bernie Hillary
White (93%) 61% 37%
Non-White (7%) 49% 50%

In New Hampshire 7% was non-white. Nevada has 35% non-whites and South Carolina registered Democrats are 44% non-white. If Nevada and South Carolina voters felt the same (based on their race) towards Hillary/Bernie,
then Bernie Sanders will get a
15 point win in Nevada and a 10 point win in South Carolina
Based on the data from this article about Nevada's Hispanic voter turnout, using that data which suggests a 27% non-white vote this would mean an even larger 18 point win for Bernie.

What about Iowa? NBC exit polls showed
Bernie Hillary
White (91%) 46% 49%
Non-White (9%) 34% 58%

If Nevada and South Carolina voters voted like Iowans (assuming anyone really votes solely based on their race), Hillary Clinton gets a

12 point win in Nevada and a 14 point win in South Carolina
(The above Nevada Hispanic turnout alternative reduces it to a 10 point win in Nevada)

Even this situation when Bernie was still being ignored by the media and the GOP competition, whit is not the 60 point blowout implied by the 80% black support claims.

So which result is more likely, voters acting like they did on Feb 1st or Feb 9th?

Campaigns get momentum, particularly in Bernie Sanders case where the media has ignored him (see above). The media is no longer ignoring Bernie Sanders. Particularly when he tied the all-time 24-hour fundraising record of 6 million dollars by taking 10 seconds to ask the people watching his NH victory speech to donate.

Caveats: Individual voters in South Carolina and Nevada won't have identical concerns as New Hampshire or Iowa. Both campaigns will do different things in these upcoming states than they did in New Hampshire. Particularly if Bernie Sanders wins in Nevada. The exit polls are samples, not the entire population of people who voted in this primary.

But the exit polls from voters is better than no information.

Less useful data: national polls. Pollees are asked about candidates even though they won't vote for months. They have much less time and reason to become informed now. Too many things will happen to change their minds.

Based on demographics and recent exit poll data,

I expect Bernie to have at least a 10 point win in Nevada.

This will only reinforce the likelihood of a victory in South Carolina.

For good non-data focused arguments supporting the overall claims in this post, read here and here.