In fact, the precision does not depend on the number of people in the state. In fact, with equal sample sizes, the confidence interval will be wider for a small state where the two candidates are both near 50% than a large state where one candidate is at say 70%.

The above assumes that the sample was randomly selected from registered voters.

The more likely problem with a sample is that to really get an estimate of how the election will turn out requires modelling who will turn out -- and that is very difficult.

the one number that jumps out to me is that a majority of respondents are moderate/conservative.

than in past years. Do you know how that matches with past NY polls?

I have wondered more about a different problem and have been surprised that polls are even as good at predicting results as they have been. I know how hard it is to get people to answer the phone when calling people in organizations I belong to. I did read one article back in the early 2000s that spoke of how non response was becoming a bigger concern. It used to be the practice that they would try more than once to reach almost all the people selected -- but by the time of the journal article I read, the nonresponse rate was becoming higher than it was. (No surprise in a day where most of us have caller id)

Statisticians do the only thing they can - which is to assume those they can't reach are like those similar to them that they reached. I have long wondered whether that assumption can work as the % increases.

any changes in the underlying electorate.

It's not only that there may be some lurking demographic or psychological difference in the responders vs non-responders, it also means that partisans may be more likely to be responders when their candidate is having a good day, even if their voting plans do not change.

there's also time of day/weekend differences, etc. in the city that never sleeps, there are a lot of people working during evening hours who sleep during the day.

just like the polling in Wisconsin felt off in favor of Clinton, to me the recent polling in NY feels off in favor of Sanders. We'll see soon enough.

I don't have any real reason to think one way or the other about the NY polling. In all states, the norm would be that HRC is favored and this was the state she represented in the Senate. I know that the Republicans had to search high and low to find the sacrificial lamb to run against her. I know that she has been a shoo in to win there - and no poll has really shown otherwise.

Still, it is not surprising that her numbers have fallen - still looking at them all, HRC is still likely to win by low double digits. I haven't seen any really detailed analysis that defines where Sanders is strong. I would suspect that he is in the universities - and that jibes with what I see on my two NY PHD student daughters' facebook pages. I have heard that he might be stronger in upstate NY, which though more conservative than NYC in some ways, might be more like VT than NYC. Some - here on DU - speculate that fracking could be an important difference there.

I would not be surprised to see a result where HRC wins by 12% - just as this poll shows. I really don't expect that her gain will be substantially bigger than that rather significant win.

That's Ithaca, where half the town is Cornell and the other half is practically a commune (they have their own local currency and everything).

Upstate is a big area that isn't at all consistent--Buffalo doesn't share much in common with people in the Hudson Valley. Hard to make generalizations.

In 2014, Zephyr Teachout won a lot of places in upstate. But that was an anti-Cuomo protest vote (Cuomo has been way, way, way too cozy with Republicans, to put it mildly). Sanders is a better-known and better-financed opponent, but Clinton is more popular than Cuomo.

The polling has Sanders outperforming Obama's 2008 numbers despite losing the African-American vote. Not sure I believe that. How many people were Clinton in 2008 but Sanders in 2016. Sure there are some, but enough to outweigh black voters transitioning from Obama to Clinton?

but a majority are moderate/conservative?

Hmmmm.

Clinton will also win a lot more than 48% in Brooklyn.

Also suspicious of 18-24 being equal to 65+ nt

Could that be a factor?

we have low turnout and in the city the turnout is driven by the machine (party and labor unions).

The sample skews young, so it doesn't seem that the landline thing is much of an issue.

this OP was touting:

http://www.democraticunderground.com/12511707207#post1

a well-known person in NY and has done so much for NY State. Why only 12? Well I'm assuming that lead will only expand over the next few days. Especially since this is a closed primary and the deadline to change registration was way back in October. It's surprising she is doing so poorly here.

against a candidate who promises free college, free health care etc..

let me be Queen - I Mean President - already".

So much more inspiring.

Oh - and send me your address. I'm sure I can find a dictionary to send you so you can look up the word "free". Or just ask your right-wing friends. They use that bs all the time.

In the poll was not wider. Asked and answered.

Save your dictionary for Bernie, he uses the word "free" more than anyone here. I'm sure you already know.

paying for things. Unless you, like some right-wingers - believe that poor people and young people just want free things. It's something I've heard many times in the right wing world. So strange to hear it from someone who claims to be a Democrat.

But maybe it's the fingers in your ears that prevents you from hearing anyone but those voices in your head.

Everyone wants giveaways. Hillary, unlike Bernie, isn't promising what she knows
she can't deliver.

saying that.

nothing is free - and to also know what what he promises is what we could have if we aspire to it.

You call it pie in the sky. I don't. I just repeated your inanity to make a point. Please don't agree with me. I will spoil my entire day thinking you believe you are making sense.

Shift a few dollars from the war machine.

involving the funding of our military.

experience, almost always guarantees plenty of mistakes.

When it comes to foreign policy, Bernie doesn't know shit from grape jelly.

credentials than Bernie. He's such a one trick pony. Likable yes, but he'll be in way over his head as President.

Got to go now. Haven't got time for immediate replys.

All that sniper fire, anyone would make mistakes.

I'll miss you.

tax loopholes for large corporations.

was running on, he wouldn't have gotten this far. His pie-in-the-sky promises are what filled up the stadiums.

concerned about a rigged economy, not getting free stuff.

Get real.

Guessing no.

He can't spin away delegate math.

If he loses 35-40 delegates here, and another 25-30 the following weekend, he's down by 270-280 delegates.

And the story is that the primary race is over.

51% of those polled say they are conservative/moderate. This is huge reverse from other states where majority (often 2/3rds) said they were very liberal or somewhat liberal (for reference in Illinois 64% said they were liberal).

Age breakout is actually favorable to Bernie (47% under 50) but IMO they over polled black voters. In '08 they represented 15% of total voters (this is Obama on the ballot) while here its 18%. I don't think there is more enthusiasm for Clinton than Obama.

This is a somewhat odd breakdown. I'm typically seeing either all landline or 50/50 cell/land.

Many studies have shown that the "cell phone only" population is closer to 40%, and I'd be surprised if it weren't higher in NY.

In any case, the lower numbers of cell phone participants might contribute to that conservative skew you picked up on.

on that basis.

The 12% difference would put Sanders at 109 delegates short by 32 delegates. Putting him further in the hole.

for gets beat fair and square

Its safe to say Bernie is done here in NY. Clinton is likely up by at least 12 and probably more like 15.

That explains some of the panic seen here.

In their populations demographics and impact in the General.

Mmmkay.

But there is no chance Bernie gets a 15 point swing in 8 days. Most people have their minds made up at this point. And its looking like Clinton will be dominating NY.

Hillary!

every state he won over the past 2 weeks he was expected to win. In reality he losing pretty badly. If this were a baseball game Hillary be up by 15 runs and Bernie just hit 2 back to back solo HRs. Bottom of the 7th

sorry to mix sports analogies all over the place, but this one looks closer to me.

HRC's got a big lead right now, but we've seen them evaporate before; she still has to close this one out.

of course, a small win or a "virtual tie" doesn't help Sanders much either, but i don't think he has much of a shot at winning big in NY.

And only 56%(?) of Clinton voters were "certain" of their vote.

So the possibility of a large swing is very much possible.

RealClear Politics keeps a running average of the New York polls. The average includes data from the Emerson poll conducted March 14-16 and the Siena poll conducted February 28-March 3.

CANDIDATE VOTE%

Hillary Clinton 63.0

Bernie Sanders 28.5