Board 8 > Politics Containment Topic 120: A Nationally Lampooned Vacation

(+) Yeah! (+)

I mean, that is exactly what the Commander in Chief does so he's just parroting him.
---

Peace___Frog posted...

Corrik posted...

if you have a masters in stats and can't conclude that then I am baffled.

Perhaps you should just admit that mathematics isn't your strong suit

Did pretty damn dominant in math in HS, ACT, SAT and in college (including stat). So, I am good.

The poll is obviously flawed. I do not have much more to say about it.
---

My Immortal posted...

I can't read the article, because WaPo, but did they give the numbers for the democrats that answered?

---

My Immortal posted...

My Immortal posted...

I can't read the article, because WaPo, but did they give the numbers for the democrats that answered?

They don't, unless there's supplemental information somewhere else that I'm not seeing.

Also, you can access WaPo if you refresh the page and stop the refresh before the adblock message comes up.
---

Metal_DK posted...

Corrik posted...

Was not a 1325 sample size. As already stated, they threw out over half the survey for being invalid.

They did not poll registered republicans by their own admission by saying they used those who "identified with or leaned republican". Which is a big difference.

And they do not state in any way the metrics of the poll satisfactorily. You polled based on ratios of sex, age etc... based on your initial 1325. Which you threw out more than half of.


The poll is beyond flawed and if you have a masters in stats and can't conclude that then I am baffled.

I didn't read the entire poll so I don't know about any sampling biases like age or something. But if you think 1325 is not large enough of a sample (which i believe you said), then you are being ignorant. Which was what you are originally saying. You are kinda backtracking here because you went from "too small" to "well its this thing".

I was talking specifically about sample size. Yes, 1325 (and even only half or whatever crap you are claiming) is large enough.

The poll literally states they used responses from 650 respondents while claiming a 1325 sample size and corresponding ages to america from aforementioned 1325 which is invalid to the 650.

I never once mentioned 1325. I said the sample size was too small. And, I was referring to 650. The poll has a ton of problems with it.

You can't claim whatever percent of Republicans would do this in this scenario due to this poll when you didn't even confirm registered Republicans nor when your sample size possibly does not even correspond to the metrics when you made the whole size correspond and threw out over half of it.
---

Corrik posted...

Would you consider postponing elections to only eligible Americans could vote if you thought it could be attained?

That is basically your question.

So what's your answer to this question?

I honestly want to know how you would have answered the polling questions as asked
---

"You can't claim the poll is valid because I don't understand it".
---

Jakyl25 posted...

Corrik posted...

Would you consider postponing elections to only eligible Americans could vote if you thought it could be attained?

That is basically your question.

So what's your answer to this question?

I honestly want to know how you would have answered the polling questions as asked

My answer is no. If they could legitimately ensure that, they can have it done by the next election after. No need to postpone an upcoming election for any reason ever.
---

And you think it's absurd to state that 52% of respondents leaning Republican would say yes?
---

Corrik posted...

Metal_DK posted...

Corrik posted...

Was not a 1325 sample size. As already stated, they threw out over half the survey for being invalid.

They did not poll registered republicans by their own admission by saying they used those who "identified with or leaned republican". Which is a big difference.

And they do not state in any way the metrics of the poll satisfactorily. You polled based on ratios of sex, age etc... based on your initial 1325. Which you threw out more than half of.


The poll is beyond flawed and if you have a masters in stats and can't conclude that then I am baffled.

I didn't read the entire poll so I don't know about any sampling biases like age or something. But if you think 1325 is not large enough of a sample (which i believe you said), then you are being ignorant. Which was what you are originally saying. You are kinda backtracking here because you went from "too small" to "well its this thing".

I was talking specifically about sample size. Yes, 1325 (and even only half or whatever crap you are claiming) is large enough.

The poll literally states they used responses from 650 respondents while claiming a 1325 sample size and corresponding ages to america from aforementioned 1325 which is invalid to the 650.

I never once mentioned 1325. I said the sample size was too small. And, I was referring to 650. The poll has a ton of problems with it.

You can't claim whatever percent of Republicans would do this in this scenario due to this poll when you didn't even confirm registered Republicans nor when your sample size possibly does not even correspond to the metrics when you made the whole size correspond and threw out over half of it.

Judging from the people ITT, i was assuming that (i havent read it just was talking about needed sample sizes - something tons of people get wrong - and not the actual poll) but:

The poll was 1325 republicans, about half said theyd be fine with postponing the 2020 election. I was assuming they didnt "throw out" anything, they just looked at the fucking people who said theyd be fine with Trump/government as a whole postponing the 2020 election.

You then came in saying it was too small (whether the ~1300 or ~650 or whatever), which is false for getting a fairly accurate representation of a population the size of the republican party.

Now if it turns out they only polled say...Texas border Republicans, perhaps its biased. That was not the original argument. I am talking about sample sizes. Which this poll did fine.
---

Jakyl25 posted...

And you think it's absurd to state that 52% of respondents leaning Republican would say yes?

I think its absurd that even 20% would say that.



Btw, I would like to point out while SEP sits here and calls me a troll now for whatever reason that he has been asked by I believe 2-3 people to answer the kmart full time employee insurance question.

And, I believe, has responded to the posts asking it but has never actually answered the question.
---

Corrik posted...

Jakyl25 posted...

And you think it's absurd to state that 52% of respondents leaning Republican would say yes?

I think its absurd that even 20% would say that.



Btw, I would like to point out while SEP sits here and calls me a troll now for whatever reason that he has been asked by I believe 2-3 people to answer the kmart full time employee insurance question.

And, I believe, has responded to the posts asking it but has never actually answered the question.

Fake news, fake news!

I already answered Lasa.
---

Nobody is calling you a troll Corrik. You not being a troll is why this is so frustrating.
---

"The survey interviewed a sample of 1,325 Americans from June 5 through 20. Respondents were recruited from the Qualtrics online panel who had previously reported identifying with or leaning toward one of the two major parties. We focus on the 650 respondents who identify with or lean toward the Republican Party."

They polled 1325 Americans. 650 of the respondents identified with or said they lean Republican


Of the 650, the percent listed said they would agree to postpone elections if yadda yadda.


The age, sex, etc were done corresponding to America for the 1325 sample size.
---

Metal_DK posted...

The poll was 1325 republicans, about half said theyd be fine with postponing the 2020 election. I was assuming they didnt "throw out" anything, they just looked at the fucking people who said theyd be fine with Trump/government as a whole postponing the 2020 election.

No, the 650 is the Repulicans surveyed like people were saying.

The survey interviewed a sample of 1,325 Americans from June 5 through 20. Respondents were recruited from the Qualtrics online panel who had previously reported identifying with or leaning toward one of the two major parties. We focus on the 650 respondents who identify with or lean toward the Republican Party. The sample has been weighted to match the population in terms of sex, age, race and education.

---

Corrik posted...

Jakyl25 posted...

And you think it's absurd to state that 52% of respondents leaning Republican would say yes?

I think its absurd that even 20% would say that.

What if you're wrong?
---

Jakyl25 posted...

Corrik posted...

Jakyl25 posted...

And you think it's absurd to state that 52% of respondents leaning Republican would say yes?

I think its absurd that even 20% would say that.

What if you're wrong?

Then that's embarassing.
---

So out of the 650 who identified as republican, half said they would be fine with postponing the 2020 election? Ya...still not a sample size issue.
---

Also, yeah, Corrik... I specifically said you aren't a troll, so...

I think the issue is that you don't actually read anything the entire way through. You read bits and pieces and then fill in blanks based on your own assumptions.
---

Yeah, I'm not disagreeing, just clarifying.
---

kevwaffles posted...

Yeah, I'm not disagreeing, just clarifying.

i didnt think you were fwiw. Thanks for clarifying.
---

Metal_DK posted...

So out of the 650 who identified as republican, half said they would be fine with postponing the 2020 election? Ya...still not a sample size issue.

As having a Masters in Stat, this question is entirely to you.

Does this poll not have serious flaws in how it was conducted?
---

A sample of 650 isn't too bad either (assuming there was no bias in sampling), the error would be understandably larger (around +/- 5-6% IIRC).
---

LapisLazuli posted...

Nobody is calling you a troll Corrik. You not being a troll is why this is so frustrating.

My bad.
---

charmander6000 posted...

A sample of 650 isn't too bad either (assuming there was no bias in sampling), the error would be understandably larger (around +/- 5-6% IIRC).

The sampling was from online, not confirmed republicans, and was set for the 1325 to match correspond norms and not for the 650 to.

I can't see how that is not flawed inherently, whether the results are 0%, 47%, or 100%.
---

statistics has a lot of inherent issues, and it's good to be mindful of its limitations.

But their result being anything above 5% should be cause for a reasonable person to berate those who feel that way, instead of immediately dismissing any results whatsoever.

At the very least, it's far more credible than the Twitter popularity poll that our president retweeted.
---

Corrik posted...

charmander6000 posted...

A sample of 650 isn't too bad either (assuming there was no bias in sampling), the error would be understandably larger (around +/- 5-6% IIRC).

The sampling was from online, not confirmed republicans, and was set for the 1325 to match correspond norms and not for the 650 to.

I can't see how that is not flawed inherently, whether the results are 0%, 47%, or 100%.

Online polling has been used successfully for a long time and besides how would telephone polling confirm respondents to be Republicans, why would a significant amount of people lie when being polled? Heck why don't you argued that Republicans lied about their answers for this poll or any poll ever done for that matter?
---

That's a pretty large-scale effort over a longer period of time than just this one poll by more than just those 650 people to fuck with statistics for shits and giggles for that scenario to be of any meaning, Corrik.
---

Some states don't even have people register by party, so I think asking people which party they identify with is fair
---

Corrik posted...

Metal_DK posted...

So out of the 650 who identified as republican, half said they would be fine with postponing the 2020 election? Ya...still not a sample size issue.

As having a Masters in Stat, this question is entirely to you.

Does this poll not have serious flaws in how it was conducted?

On sample size? No not really.

Internet sampling is always a problem though.
---

Metal_DK posted...

Corrik posted...

Metal_DK posted...

So out of the 650 who identified as republican, half said they would be fine with postponing the 2020 election? Ya...still not a sample size issue.

As having a Masters in Stat, this question is entirely to you.

Does this poll not have serious flaws in how it was conducted?

On sample size? No not really.

Internet sampling is always a problem though.

I am asking you if the poll itself in all aspects has significant flaws in it.

The sample size to me is a flaw because we don't know the dynamics of the sample size. We do not know how their ages, sex, incomes, and location actually correspond. Or if they are even actual Republicans.
---

charmander6000 posted...

Online polling has been used successfully for a long time and besides how would telephone polling confirm respondents to be Republicans, why would a significant amount of people lie when being polled? Heck why don't you argued that Republicans lied about their answers for this poll or any poll ever done for that matter?

This is a bit unfair. Most online polling is done in a location where people choose to go, which causes sampling issues.
---

Metal_DK posted...

charmander6000 posted...

Online polling has been used successfully for a long time and besides how would telephone polling confirm respondents to be Republicans, why would a significant amount of people lie when being polled? Heck why don't you argued that Republicans lied about their answers for this poll or any poll ever done for that matter?

This is a bit unfair. Most online polling is done in a location where people choose to go, which causes sampling issues.

Which is why polls are normalized. I'm assuming this was done by a polling firm and not some GameFAQs like poll of the day.
---

Corrik posted...

I am asking you if the poll itself in all aspects has significant flaws in it.

The sample size to me is a flaw because we don't know the dynamics of the sample size. We do not know how their ages, sex, incomes, and location.

It has the flaw of being an online poll, which is the majority of polling these days.

Dynamics of the sample size is not a sample size issue. Its a selection bias issue. There could be a selection bias, again every online polling goes through this sadly (and the internet has allowed polling to exponentially increase in how often we do it due to what the internet is). Its another reason why the internet being the main aspect of our lives is a problem. We had better balance in the past imo, but thats another topic because people hate when i bring up you know what of you know when.
---

charmander6000 posted...

Metal_DK posted...

charmander6000 posted...

Online polling has been used successfully for a long time and besides how would telephone polling confirm respondents to be Republicans, why would a significant amount of people lie when being polled? Heck why don't you argued that Republicans lied about their answers for this poll or any poll ever done for that matter?

This is a bit unfair. Most online polling is done in a location where people choose to go, which causes sampling issues.

Which is why polls are normalized. I'm assuming this was done by a polling firm and not some GameFAQs like poll of the day.

Normalizing the poll is better than nothing, but it still doesn't really do enough to fix the self selection issue.

Normalized weights consider the survey weights, but not the other aspects of the design like cluster sampling or stratification. It can still underestimate variance results.

The point is. The poll overall might overstate it, but its accurate enough to say that plenty of republicans would support holding the 2020 election. Even if that number is only half the 50% that the poll result says, thats pretty fucking telling.
---

You are correct, but that's an issue with polling in general. Polls are never meant to be taken as 100% true. What we can say is that only 20% of Republicans believing that is much more unlikely than 50% of Republicans.
---

charmander6000 posted...

You are correct, but that's an issue with polling in general. Polls are never meant to be taken as 100% true. What we can say is that only 20% of Republicans believing that is much more unlikely than 50% of Republicans.

i would agree with that, yes. I would argue that the internet has a bigger problem with self selection bias than other ways we have conducted polling in history (and also how often due to how cheap polling on the net can be), but ya.
---

People are way too trusting of studies, especially the ones who use a 95% confidence interval. It means based on the data, X is 95% probable to be within some range.

But consider, what happens when a model does not produce significant results or produces boring ones? It doesn't get published and the author is perfectly allowed to try again. If you run 20 models with no real validity, chances are decent that 1 of them will appear significant using a 95% confidence threshold.

If the study results conflict with your intuition, you should be even less trusting of the study. Especially when the methodology (polling people) has directly underperformed Republican intuition in very recent history (the 2016 election).

And remember, this entire question is a matter of feelings. Whether one supports delaying the 2020 election is entirely a question of feeling. It is inherently not fact-based.
---

The Mana Sword posted...

I think my eyes just rolled all the way in the back of my head

---

The Mana Sword posted...

The Mana Sword posted...

I think my eyes just rolled all the way in the back of my head

---

For people who want a more in depth look at this problem, read Paul Romer's paper, The Trouble with Macroeconomics. He's the 2012 Nobel winner in Economics.
---

Peace___Frog posted...

The Mana Sword posted...

The Mana Sword posted...

I think my eyes just rolled all the way in the back of my head

---

If you guys have any reason to disagree with me, you are free to make your argument. Otherwise, I think it's almost fair to conclude you don't have any.
---

Or we're smart enough not to take the bait.
---

LapisLazuli posted...

Peace___Frog posted...

The Mana Sword posted...

The Mana Sword posted...

I think my eyes just rolled all the way in the back of my head

---

I kinda doubt anyone who responded has read Romer's paper yet. But uh, this is the great thing about the modern left I suppose. They see useful information as "bait" while being totally unable to stop themselves from taking the bait, every time. They judge people more on orthodoxy than the content of their argument. They ignored the people on the ground who said trouble was coming in the Midwest, and....

Yes, #This Is Why You Lost.
---

honestly, the argument of 52% or 40% or whatever doesn't really matter. even if it's off by some and isn't quite over half i'm not going to doubt that a very significant chunk of republicans feel that way. similarly i wouldn't have doubted that a vast majority of democrats would've preferred a third Obama term over Hillary or Trump. that wouldn't shock me if it were 60%+ quite frankly, with a good chunk of that being fine with the same thing proposed here. so the idea of Trump having a good chunk willing to suspend? well yeah, 70-80% of the republican base supports him, he is the representative of the party. they may not like everything he does but they like enough (way more than Mitch, that's for sure). net result, they'll have no problem with more years with him: it means no dems and he continues to dominate.


in the end it's a silly poll to have mind you, it's still 2017. run that poll again in 2019, when it'd actually represent something after most of a presidential term. right now there's still a "Trump high" because he's constantly in the news and doing things many of them like. it's the impression that he's doing stuff when he's really not.
---

Fyi, hashtags only capture the first word
---

You can put on as much bait as you want, no bites.
---

That's kind of sad. My post was not bait at all . Nor is Paul Romer's paper, as far as I can tell.
---