The Clinton Victory Thread

[Columbus]

http://fivethirtyeight.com/features/ber ... delegates/

Phil Wasserman explains why Bernie is even less electable than his standing in the Democratic party nomination race makes him look.
Tom

[Willie71]

http://fivethirtyeight.com/features/bernie-sanders-is-even-further-behind-in-votes-than-he-is-in-delegates/

Phil Wasserman explains why Bernie is even less electable than his standing in the Democratic party nomination race makes him look.
Tom

These "experts" fail to account for the fact that Sanders wins in caucuses, which by their nature have much different numbers. Whether good or bad, caucuses don't represent a proportionality accurately. Fivethirtyeight is for the most part a respected entity, but they are trying to understand something that is elusive to measure statistically. I am quite familiar with this principle as a therapist. We have a lot of statistical evidence regarding potential treatment outcomes, but much is simply in the unmeasurable "relationship" that cannot be measured statistically.

Taking Wisconsin as an example, if all the polls over the past two months were averaged, Sanders should have lost by 10+ points. It was only in the past couple weeks the lead evaporated, and Sanders beat the latest polls by 10+ points. In New York, Clinton was 40 points ahead, now down to 10 points. Almost two weeks to go. Sanders is attracting crowds of 10,000+, while Clinton is having a hard time filling a school gym. How does that translate to votes? None of us could say with certainty.

Wasserman is trying to use math to explain a psychological/sociological process, and missing the mark.

[Columbus]

These "experts" fail to account for the fact that Sanders wins in caucuses, which by their nature have much different numbers. Whether good or bad, caucuses don't represent a proportionality accurately.

Exactly. Caucuses are much less representative of the final outcome in the general election.

Their turnout tends to be tiny, and hardcore supporters of a particular candidate, and much better informed than the typical voter in the real election. So while Sanders is doing quite well in caucuses, November is not one. And the superdelegates of the DNC know that.

So while most people agree that as more and more people become familiar with Sanders policies those policies acquire support, that doesn't mean he'll have enough in November to accomplish anything.
Tom

[Thommo]

Taking Wisconsin as an example, if all the polls over the past two months were averaged, Sanders should have lost by 10+ points.

Hang about, you said that about Washington too, and it wasn't true there. What's your source for this claim?

http://www.realclearpolitics.com/epolls ... -3764.html
http://projects.fivethirtyeight.com/ele ... emocratic/

Both show not only Sanders winning, but a slight upward trend for Sanders as well, with the trend line finishing at around 3-4%. Given Sanders was actively campaigning in Wisconsin and outspending Clinton on advertising by around $2-$1, this isn't remotely the upset that overturning a steady 10+% point deficit would imply. He finished up about 7% better than the polls, not 17+%, it's a good result, but a far cry from being outside of the range of predictions. The errors on these things (with the smaller sample sizes and polling representation issues that arise) are normally around 6%, which if you look at things like the demographics of who showed up to vote (in particular mediocre black turnout) can be accounted for without evoking a seemingly spurious starting point or quasi-religious attitude towards evidence when it doesn't suit the conclusion.

Wasserman is trying to use math to explain a psychological/sociological process, and missing the mark.

He's using the best techniques he has. We all know how Nate Silver the fivethirtyeight founder got to where he is. We all know how he applied math to baseball, and to politics and put aside these supposedly better psychological/sociological measures and proved substantially more accurate at predicting outcomes.

[NineOneFour]

Sanders is going to lose New York.

Sorry, but he is.

[OlivierK]

Wasserman is trying to use math to explain a psychological/sociological process, and missing the mark.

He's using the best techniques he has. We all know how Nate Silver the fivethirtyeight founder got to where he is. We all know how he applied math to baseball, and to politics and put aside these supposedly better psychological/sociological measures and proved substantially more accurate at predicting outcomes.

Wasserman is not using the best techniques he has, he's doing statistics badly. He's introducing an error into a popularity measure that biases towards winners of primary states, where turnout is high, over winners of caucus states, where turnout is lower. We already know Clinton has a better record in primaries and Sanders in caucuses, so Wasserman is merely showing that when you introduce an error that favours primary winners, that favours Clinton. No fucking kidding. Nate Silver should be taking Wasserman out the back for a quiet word, because Silver would never make such a hash of the maths.

FiveThirtyEight built its reputation precisely on removing these sort of errors - house effects, demographic effects, etc - to get a clearer, more accurate picture from the data. Poll aggregation, FiveThirtyEight's political stock in trade, requires the data to be pushed into an apples to apples state so that the aggregation is meaningful, and it's what Silver is a genius at. Wasserman is doing the same dumb shit that RealClearPolitics do with their polling averages - combining data without thought for whether that;s a valid thing to do. That can often get you close, but Silver's approach is superior, and has a superior track record to show for it.

Adding total votes from primary and caucus tallies is a demonstrably less accurate measure of popularity than the delegate count, so insinuating that it's a good thing for Clinton to have a bigger lead when looking at a less accurate measure is just nonsense. As a mathematician, I find such an approach pitiful.

[Teague]

Sanders is going to lose New York.

Sorry, but he is.

...but what if he wins New York....

[GT2211]

Wasserman is trying to use math to explain a psychological/sociological process, and missing the mark.

He's using the best techniques he has. We all know how Nate Silver the fivethirtyeight founder got to where he is. We all know how he applied math to baseball, and to politics and put aside these supposedly better psychological/sociological measures and proved substantially more accurate at predicting outcomes.

Wasserman is not using the best techniques he has, he's doing statistics badly. He's introducing an error into a popularity measure that biases towards winners of primary states, where turnout is high, over winners of caucus states, where turnout is lower. We already know Clinton has a better record in primaries and Sanders in caucuses, so Wasserman is merely showing that when you introduce an error that favours primary winners, that favours Clinton. No fucking kidding. Nate Silver should be taking Wasserman out the back for a quiet word, because Silver would never make such a hash of the maths.

Well I think that is kind of Wasserman's argument, that if we want to argue the will of the people, we should favor high turnout ones events over low ones. Giving equal weight to caucus events and primary ones as far as delegates goes is essentially giving far more weight to voters in caucus states. If you rack up a narrow victory and the victory relies on racking up delegates in complicated vote systems which produce low turnout, its rather questionable as to argue that is the 'will of the people'.

Adding total votes from primary and caucus tallies is a demonstrably less accurate measure of popularity than the delegate count, so insinuating that it's a good thing for Clinton to have a bigger lead when looking at a less accurate measure is just nonsense. As a mathematician, I find such an approach pitiful.

Again I'm not sure that it's less accurate. The delegate count itself is an inherently inaccurate as a measure of overall popularity because it has a fairly strong bias towards caucus voters.

[GT2211]

Sanders is going to lose New York.

Sorry, but he is.

...but what if he wins New York....

We've had 5 polls in the last 2 days and he hasn't come closer than 11 in any of them. Winning NY is pretty improbable at this point. And even if he were to make up that gap and its likely to look more similar to Michigan which would be kind of a loss because that would be another 250 delegates off the board and his time is ticking to make up that gap.

[Thommo]

Well I think that is kind of Wasserman's argument, that if we want to argue the will of the people, we should favor high turnout ones events over low ones. Giving equal weight to caucus events and primary ones as far as delegates goes is essentially giving far more weight to voters in caucus states. If you rack up a narrow victory and the victory relies on racking up delegates in complicated vote systems which produce low turnout, its rather questionable as to argue that is the 'will of the people'.

I don't think that is really his argument to be honest. What he's saying is fairly clear:

1) Bernie Sanders’s supporters believe that superdelegates are going to bow to the “will of the people” if Sanders ends up winning more pledged delegates than Clinton by June.
2) If Hillary has received a larger share of the public vote than Bernie, then she will appeal to superdelegates that it's not "the will of the people".
Therefore:
3) Even if Sanders were to draw even in pledged delegates by June, Clinton could be able to persuade superdelegates to stick with her by pointing to her popular vote lead.

Notwithstanding the opacity of the exact method used by the greenpapers source (and bearing in mind it makes little different to his point anyway), Wasserman has then simply done some arithmetic to show how big of a chance Sanders has to be able to eliminate that argument, and concludes that if he doesn't, then he might have yet another hurdle to overcome to secure the nomination. Superdelegate votes matter too.

It's fairly obvious that he has got a point. After all, if the superdelegates were obliged to go with the majority of pledged delegates they'd serve absolutely no function at all. Their very existence is precisely to allow the party elite to tip a finger on the scales and move it away from their formula of half weighting by democratic voter turnout numbers in presidential elections and half weighting by electoral college amounts.

[OlivierK]

Yes, that's a fair argument. As I said in the other thread that Wasserman's article was posted on, it's obvious that no matter how poor the statistic of "popular vote" attained by adding the apples and oranges of primary and caucus vote tallies is, whoever is leading in that measure will interpret it as meaningful if they can get some political advantage out of doing so.

If it turns out that the only way to get Clinton over the line is to introduce this highschool statistical error, then that's what her campaign will do, and the innumerates amongst the public (most of them) will lap it up, as will the superdelegates (who aren't looking to understand the maths, but rather to justify and already made decision). Wasserman has a point that no matter how poor a measure this hybrid total is, Clinton would wring it for all its emotional/political worth, which for her is not inconsiderable.

If I was a DNC official from a caucus state whose results effectively got weighted down to about 25% of those of primary states, I'd be pretty pissed off if a naive (nicest word I can think of) popular vote total was used as an argument to overturn the more accurately weighted delegate total. If that's how it goes down (and I don't think it will this time around as Clinton's well on track to win the delegate count), I could see states switching from caucuses to open primaries for the next cycle so that their voice would not be diminished in the same way again.

[NineOneFour]

Sanders is going to lose New York.

Sorry, but he is.

...but what if he wins New York....

What if unicorns could fly?

[GT2211]

Yes, that's a fair argument. As I said in the other thread that Wasserman's article was posted on, it's obvious that no matter how poor the statistic of "popular vote" attained by adding the apples and oranges of primary and caucus vote tallies is, whoever is leading in that measure will interpret it as meaningful if they can get some political advantage out of doing so.

Isn't that essentially what the pledged delegate system is doing? Its a way of trying to measure support, comparing the same apples and oranges of different voting systems.

If it turns out that the only way to get Clinton over the line is to introduce this highschool statistical error, then that's what her campaign will do, and the innumerates amongst the public (most of them) will lap it up, as will the superdelegates (who aren't looking to understand the maths, but rather to justify and already made decision). Wasserman has a point that no matter how poor a measure this hybrid total is, Clinton would wring it for all its emotional/political worth, which for her is not inconsiderable.

Its not like this is the first time people have discussed the overall vote in regards to primary elections. You seem to be attempting to smear Hillary for a newly invented position, that we should just ignore popular votes or superdelegates.

If I was a DNC official from a caucus state whose results effectively got weighted down to about 25% of those of primary states, I'd be pretty pissed off if a naive (nicest word I can think of) popular vote total was used as an argument to overturn the more accurately weighted delegate total. If that's how it goes down (and I don't think it will this time around as Clinton's well on track to win the delegate count), I could see states switching from caucuses to open primaries for the next cycle so that their voice would not be diminished in the same way again.

This strikes me as a benefit, not a drawback. We should encourage states to switch to systems which maximizes participation. Think about all the whining about voter lines we hear about in primaries as suppression, then think that is standard procedure for caucuses. Those things take hours. And as I remember reading about in HI, a guy showed up as the polling location was opening and after a few minutess went across the street and came back an hour later was denied entrance for being too late.

If you have kids, a work schedule that isn't 9-5, commitments, poor health, caucuses aren't for you.

The party rightly opposes “Voter ID” laws as a contrived burden that depresses voting by race and class. So then this same party makes people come stand in a gymnasium for two or three hours on a weekend?

Two political scientists from Brigham Young University studied these events, resulting in a paper called “Who Caucuses?” Mostly it’s “the wealthy, educated, white and interested.” This fits with The Seattle Times portrait of one caucus in the city’s most nonwhite neighborhood: “While the caucus was located in the racially diverse but gentrifying Rainier Valley, most of those who turned out were white.”
http://www.seattletimes.com/seattle-new ... emocratic/

[Thommo]

This strikes me as a benefit, not a drawback. We should encourage states to switch to systems which maximizes participation. Think about all the whining about voter lines we hear about in primaries as suppression, then think that is standard procedure for caucuses.

Indeed, there's a reason that turnout is 3-4 times higher (compared to eligible voting population) in primary states than caucuses. It knocks all the other voter suppression problems into a cocked hat.

And then you've got the disaster that was Nevada to consider.

[OlivierK]

@GT2211: I'm not trying to smear Clinton for pointing out that if the opportunity arose, she'd make political hay out of a naive popular vote total. Every candidate would do so if it were to their advantage, including Sanders, and all past and all future candidates, I'm sure. It's just a bad measure compared to pledged delegates, so I'd always consider using it to sway superdelegates away from the pledged delegate winner in the name of "respecting the people's will" as dodgy. As others endlessly, and correctly, point out, there's plenty of exploitable dodginess in the US electoral system. Doesn't mean a blind eye should be turned, though.

[Teague]

Sanders is going to lose New York.

Sorry, but he is.

...but what if he wins New York....

What if unicorns could fly?

Yeah, I knew that would be meaningless to you.

[NineOneFour]

Sanders is going to lose New York.

Sorry, but he is.

...but what if he wins New York....

What if unicorns could fly?

Yeah, I knew that would be meaningless to you.

Because I go by verifiable reality, not what I wish would happen.

[Teague]

...but what if he wins New York....

What if unicorns could fly?

Yeah, I knew that would be meaningless to you.

Because I go by verifiable reality, not what I wish would happen.

It was a quote.

[Scot Dutchy]

What if unicorns could fly?

Yeah, I knew that would be meaningless to you.

Because I go by verifiable reality, not what I wish would happen.

It was a quote.

You are obsessed with the word 'if'. Not surprising when you are a Sander's supporter.

[NineOneFour]

What if unicorns could fly?

Yeah, I knew that would be meaningless to you.

Because I go by verifiable reality, not what I wish would happen.

It was a quote.

What was a quote? Are you high?