Exclusive - Nate Silver Extended Interview

November 14, 2016 - Nate Silver & Adrian Grenier 11/14/2016 Views: 42,731

FiveThirtyEight founder Nate Silver discusses the complexities of polling, explains the importance of understanding uncertainty and weighs in on a Donald Trump presidency. (8:57)

Watch Full Episode

Please welcome Nate Silver,

-everybody!-(applause and cheering)

-How are you?-Good, man.

You're good?

All right, let's, uh,make it really simple.

-Uh, heads or tails?-(laughter) -Tails.

-You're losing it, man.-Yeah. -(laughter)

You're losing it.

-The gall, Nate Silver.-Yeah.

To show your facein the streets.

-(laughter) -To wear the cap... What?

Some guy yelled at meon the subway, actually.

-He did?-Yeah.

-And rightfully so. -He's like,"You blew it, Silver."

You did. You blew it, though.

You're wearing the...you're wearing the cap.

People who don't know thisdon't understand

how ballsy this is.

You, um... you said

that Donald Trump hadthe same chance of winning

-as the Chicago Cubs...-Which was...

...when they were four-one down,right?

Which was meant to be a...Three-one down.

Three-one down, yeah.

Meant to be an illustrationof uncertainty in the race.

And the Cubs...

I mean, there is somethingweirdly Karmatic,

I guess, about 2016.

Like, all this weird...

Can I say (bleep) on TV?

-(laughter)-All this weird stuff happens.

You can say (bleep)in Trump's world, my friend.

-(laughter)-(Bleep) is no longer a thing.

Censorship is nota thing anymore.

Let's-let's talk about polling.

-Because this entire race...-Yeah.

This entire race,as Nate Silver...

I mean, you're not a pollsterper se.

-You aggregate the polls.-Yeah.

All right? And you tryand figure out what they mean.

-This entire raceyou were wrong. -No.

-(laughter) -And I knowit's, like, not an easy

television explanation,but our job is to take...

You're sayingmy people are stupid, Nate?

-I am not...-(laughter)

I'm saying the medium.

-For television, it's veryelevated. -Okay. Okay.

-What I'm saying is that...-He said you guys were stupid,

-just so you know,but carry on. Yes. -(laughter)

We're taking...I mean, everyone could see

-that Hillary Clinton was aheadin the polls. -Yeah.

'Cause she was, right?

Our job is to take that dataand translate that

-into an estimate of risk.-But was she ahead, though?

Because there were polls...

-In the polls, she... I mean,the polls... -No, but which...

But which polls?That's the thing. Which polls?

Some of the best polls, in fact,and-and, you know...

No, no, no, but I mean,

do we still call them the bestpolls if they were wrong?

Why are people still lookingat the polls? Like, I'm gonna...

-Well, it's like...-Honestly, as a person

-who is experiencing thisfor the first time... -Yeah.

...I am confused by why

people in Americaare so obsessed by the polls.

Well, first of all,we would say

that if you havea bunch of polls

-that show the racewithin three points... -Yes.

...as they did on election day,

a bunch of swing statesthat are close,

people should characterizethat race

as being extremely competitive.

People should go out and vote,and, you know...

That's what we're trying to do.

We're trying to take that dataand translate that

into an estimate of probability.

-Our probability was 70% forClinton, 30% for Trump. -Yes.

Other people had Trumpat two percent.

And so, if we're sayingthere's a higher risk of rain,

higher risk of an earthquakethan you might think,

um, be wary.

At the very least,don't take anything for granted.

And then, Trump wins, then...

It is hard, right?

It's much easier to beon the right side of things.

But our whole thing is about,

there's uncertaintyin the world.

-The world...-But now, let me ask...

Let me ask you this.Let me ask you this.

-With that uncertainty...-Yeah.

...do you maybe acknowledgethat maybe things have to change

in the way you measure it?

And I'll say this.

Because up until now, you know,in terms of Nate Silver,

FiveThirtyEight,if you look at the years

that you hadto aggregate the polling,

one could argue it wasn'tthe hardest thing to calculate.

-Barack Obama, everyone saidwas going to win. -Right.

Barack Obama, again, everyonesaid was going to win.

You called it correctly,all 50 states,

but it was an easier raceto call.

Donald Trumphas defied all expectations.

And one of the biggest things

that people didn't seemto acknowledge

was the Bradley Effect,all right?

In that, people are ashamedto tell pollsters

that they're votingfor somebody like Donald Trump.

So I'm not sure this is true,because Republican Senate

candidates over-performtheir polls, also.

And so, what I think you hadis a higher turn out

of Republicans at the end,I think...

You know, Clinton had a badtwo weeks of the campaign

-with the Comey letter,with the WikiLeaks stuff. -Yes.

With Trump being sort of

halfway approaching normaland quiet, I guess.

-Um, and you saw...-Well, approaching less crazy.

Let's not say normal, becausethat was far from normal.

-Yes. -But you sawRepublicans come home,

especially,um, white Republicans

without college degrees,came home to Trump

in a big way.

And if you have a lot of unde...

One reason why we were lesscertain than we were in 2012

is 'cause you hadfar more undecided voters.

If Clinton's at 46%and Trump's at 43%,

well, she's gonna get her 46%,she'll win the popular vote,

-Yes. -get 47%,but if Trump goes from 43 to...

well, actually, 46-- 'causehe'll lose the popular vote

but not-- he wins,um, Michigan, Wisconsin,

Pennsylvania and Florida, andgoes from a narrow Clinton win

to a pretty impressive Trump winin the electoral college.

Okay, let's talkabout the turnout, then,

because what you are saying--

you did, I willgive you credit for this--

you did mention it beforehand,

we even spoke about iton the show--

-Yeah. -you said Democratsshouldn't be comfortable

with early turnout numbers,because you showed

that overall turnout is lower.

And so that's somethingthat should be more worrying.

And that ended upbeing the truth,

is that fewer people came outfor Hillary Clinton.

But my questionstill goes back to...

what is the purpose of the poll?

Why do people need to knowabout a poll?

Like, why don't they just votebased on what is happening?

I don't understand...-And again, I think any race

that's within three points,six points,

depending on the type of race,people should treat that outcome

as beingsubstantially uncertain.

But here's the thingWe hear all this talk now

about how reportersin New York and Washington

need to be in touch with peoplein the rest of America,

-Yes.-which I totally agree with.

Um... I thinkone of the best ways to do that

is to take a telephone pollwhere you randomly sample people

from all parts of America

and not just the peoplein your social network,

not just the peoplewho make for clich├ęd stories.

I mean, we have reporterswe send out to different corners

of the country,and it helps to confirm

what we saw in the polling,because it was a close race.

But now let's talkabout things that don't confirm.

So for instance,one of the things

we saw repeatedly in this race,one of the problems with polling

was they do not knowhow to poll Hispanic voters.

-Mm-hmm.-They struggle with that.

And I saw a lot of Hispanicjournalists saying no,

it's not that Hispanic votersare difficult.

You don't know howto speak to these people,

-Yeah.-you don't know how to engage

-Well, literally, in some sense,-with Hispanic voters.

is that there area lot of Spanish...

or Hispanic voters whosepreferred language is Spanish.

A lot of voters...a lot of polls

don't haveSpanish language interviewers.

And they're different thanEnglish-speaking Hispanics.

And by the way,people talk about the Midwest,

-where Clinton underperformedher polls. -Yeah.

There were also stateslike California and New Mexico

where she overperformedher polls.

-And so, you know, it definitelyis true... -But then I go back

to-- and forgive meif this sounds like

-the dumbest questionin the world. -Mm-hmm.

What is the point of the thing,what is the point

of the prediction ifthe prediction doesn't happen?

It's a forecast,not a prediction.

-It's a forecast?-It's an estimate of risk.

But then what is...what is the point of that?

What I'm trying to say,fundamentally,

like, you say with rain,you go like,

-there's a chance of rain.-So we would say if, you know,

if the news media had acted

like Trump hada 30% chance of winning,

-which is whatour forecast said, -Yes.

then I think they would haveacted very differently.

Frankly, if the Clinton campaignhad believed that Trump

had a 30% chance of winning,they might have spent more time

in Michigan and Wisconsin,and not Arizona and Ohio

and states like that.

So this isactionable intelligence

to prepare for, um,for different eventualities.

'Cause a 30% chanceis way different

-than a three percent chance.-I still find it weird

that it sounds-- this iswhat it sounds like to me.

Depending on who's gonna win

determines on who'sgonna go out and vote.

And yet, shouldn't it just beyou go out and vote

because you want the personto be president?

-(applause, whooping) -Thatsounds like a very strange world

to be living in.

Fundamentally,it doesn't make sense.

People... people don't vote

when they thinkthe race is a blowout,

-and that's why...-But it's a blowout

because theywere gonna vote that way.

Well, it can get circular, yeah.I mean, this is why...

this is why it's importantto be careful around...

I'm gonna choke you,Nate Silver.

And I really like you.I'm gonna choke you right now,

-Nate Silver.-Here's what I'd say.

I'd say that polling

is the least--it's like democracy,

the least worst systemever invented.

-There are lots of problemswith the polls, -Yes.

but if you want to tryand represent

all the different partsof America,

then, in theory at least,

and over the long run,in practice,

polling is a pretty good wayto have a representative sample

of the population in a way whereour media often doesn't do that.

That makes a lot of sense.

Let's do this--I've got to wrap this up.

Uh, Donald Trump,off the top of your head,

because, I mean, what's theworst that could happen here?

What are the chances thathe will follow through

on building a wall,locking Hillary up,

and draining a swamp?

I think, I'm not gonna makeany more predictions,

-I do think that Trump...-(laughter)

What's the worstthat can happen, Nate?

You're already at the bottom;you might as well go up.

-Get yelled up.-Just do it.

Come on, just throw it at me.What are the chances?

I think Trump'sgonna be very obsessed

with his polling numbers,ironically,

and his approval rating,in that,

if he detects public pushbackagainst his agenda,

he might abandon thingspretty quickly.

So just tell himthat numbers are low,

and then he'll go, he'll go--he'll be a better president.

If he, I mean, maybe...

No, I'm telling you, just tellhim the numbers are low.

And then he'll becomea better president.

-Nate Silver.-Thank you.

Thank you so much for beinghere, man.