Quinnipiac nat. poll: D: Clinton 61% Sanders 30%; R: Trump 28% Cruz 24% Rubio 12%

[Skye]

Nate thinks Marquette, the Wisconsin gold standard poll, is only a B? I guess that's yet more evidence that the guy is an idiot and a fraud.

I absolutely adore random Atlas posters saying how good they are at analyzing stuff like this.

[Lief 🗽]

You betray your profound ignorance by clinging to irrelevant nonsense like MOE and sample size, X. Run along.

Average sampling error and contrasting separate pollster results is what portrays a pollsters accuracy, are you honestly kidding me?

You can't jumble together hefty words to make your claims more true. What a joke you are, lol.

You are the one jumbling together words and thinking you're not coming off as incredibly ignorant, friend. As long as sample size is above an adequate level (400 or so), there's not that great an improvement in accuracy in going higher, and if anything I would be wary of polls with sample sizes that are too high because 1) it reeks of overcompensation (in an attempt to fool people like you who don't understand these things) and 2) good sampling methodology is expensive, and each additional respondent is expensive, and money does not grow on trees, so a quality polling firm may rightly spend their money on getting a good small sample than a not so good large one. Because at the end of the day, people who understand how this works (not you) are going to value methodology, sampling, etc. over mostly inconsequential things like MOE and the negligible improvement in accuracy a 1% lower MOE brings any day.

This is all pretty off topic though; let's all return to pretending this poll from a mediocre uni pollster isn't a huge outlier.

[Penelope]

I see Cruz is now the Anti-Trump flavor of the month. I wonder if he will generate as much hype as Walker, Bush, Carson, and Rubio before him? More importantly, how long do you guys wager it'll be before the media and the GOP establishment decides to hype up someone else? I hear Christie is on the rise.

[Cruzcrew]

Junk poll. Trump isn't leading by 30%. Quinnipiac clearly hates Trump.

[EliteLX]

You betray your profound ignorance by clinging to irrelevant nonsense like MOE and sample size, X. Run along.

Average sampling error and contrasting separate pollster results is what portrays a pollsters accuracy, are you honestly kidding me?

You can't jumble together hefty words to make your claims more true. What a joke you are, lol.

You are the one jumbling together words and thinking you're not coming off as incredibly ignorant, friend. As long as sample size is above an adequate level (400 or so), there's not that great an improvement in accuracy in going higher, and if anything I would be wary of polls with sample sizes that are too high because 1) it reeks of overcompensation (in an attempt to fool people like you who don't understand these things) and 2) good sampling methodology is expensive, and each additional respondent is expensive, and money does not grow on trees, so a quality polling firm may rightly spend their money on getting a good small sample than a not so good large one. Because at the end of the day, people who understand how this works (not you) are going to value methodology, sampling, etc. over mostly inconsequential things like MOE and the negligible improvement in accuracy a 1% lower MOE brings any day.

This is all pretty off topic though; let's all return to pretending this poll from a mediocre uni pollster isn't a huge outlier.

I simply pointed our Our Lord & Savior PPP's historical average polling error is actually the exact same as Quinnipiac. They are not a garbage pollster.

I'm not in college for statistics nor am I an expert on political polls, so I won't speak like I'm an expert. You are just over-discrediting something when there is a clear trend Cruz is catching heat. Once Ted wins Iowa, has his face/name running the airwaves for a week, and his campaign starts throwing cash nationwide on huge nationwide ad campaigns and more touring appearances it's going to get even more tight. If he then goes on to take SC then old Don' could be in the kahoots. Not that anything is a guarantee, but.

[emailking]

EliteLX, he's just trolling.

[Lief 🗽]

Er, no, I'm not. Quinnipiac is not a very good polling firm and this poll is an outlier. I don't think either of those statements are controversial at all.

[Obama-Biden Democrat]

Who would have thought that there would be a candidate out there to make the radical right wing extremist Ted Cruz look like a moderate. I bet if Cruz wins the nomination and losses the general, he will be accused of being not Conservative enough.

[EliteLX]

Er, no, I'm not. Quinnipiac is not a very good polling firm and this poll is an outlier. I don't think either of those statements are controversial at all.

Hm, you must of not heard correctly.

Quinnipiac has the same historical end-result polling error average as your beloved perfect polling messiah PPP. Quinnipiac University is not a garbage pollster, and if hypothetically they are, you can count out 93% of every poll result posted in the primary/general board.

And to be on topic, I'm not claiming this result is completely accurate. But it's trend is completely accurate, Cruz has nowhere to go but up from here on out w GOP primary voters, and the Cruz/Don gap will slowly close with that in mind.

[Senator Cris]

So, dear Lief, I guess you'd have said the same if Trump was leading by a lot. Sure.

[Gustaf]

Lief is correct in that things like MoE and sample sizes are not particularly important in determining pollster quality. I'm not as convinced that this is what is at stake in this discussion though because EliteX is being a bit confusing.

FWIW, my impression was that Quinnipac was ok but not great. But I don't pay this as close attention as I used to.

[Sorenroy]

Sorry for dragging this back from the dead, but I wanted to compare this with other polling organizations and Quinnipiac doesn't have anything about race in their methodology. Does anyone have any idea about whether or not their sample covers this, or is this just something they don't think is important enough to ask?

[Mr. Morden]

Sorry for dragging this back from the dead, but I wanted to compare this with other polling organizations and Quinnipiac doesn't have anything about race in their methodology. Does anyone have any idea about whether or not their sample covers this, or is this just something they don't think is important enough to ask?

They list the racial breakdown of the general election sample here:

http://www.quinnipiac.edu/images/polling/us/us12222015_demos_Uhkm63g.pdf

though they don't say what it is in the Democratic and Republican primary subsamples.