CNN has an article claiming that the majority of Americans are against Trump's travel ban.
cnn.com
At the bottom it says the sample was 1,002. That's it, just 1000 people, used to represent a country of over 320 million people. How does that make sense?
I understand that the more answers you get, the more accurate it is. But aren't there so many variables when human beings are involved?
Other urls found in this thread:
youtube.com
i2.cdn.turner.com
purplemath.com
youtube.com
reason.com
reuters.com
twitter.com
not really it's a binomial distribution
this
also its cnn.
they probably interviewed only black women in new york city
random sampling means that every american has a chance to be selected for the survey. Of course this is assuming that the sample used random sampling, or at least a systematic stratified sample
They only ask the questions to urbanite Democrat cucks. That's also the reason why all the polls said Shillary would win.
>polled by phone during middle of the day
Gee, it's not like only niggers and other shitskins are home during the day.
>We sampled 1002 college kids that are overwhelmingly liberal
>They say travel ban is bad, thus most Americans say it's bad
(((CNN)))
Also archive the link, why the fuck do you faggots never do that
Reminds me of the focus groups CNN had for the debates. They claimed Hillary won when like 65% of the group was democrats.
you guys just came here to shitpost about how fake cnn is and how liberals are retarded without even attempting to answer OP's question. Nice job.
Also another question: I'm sure a lot of people say "no I do not want to take part in this poll". Many people probably said that.
Are those people just irrelevant? I mean I feel like that matters.
Fake polls
Fake news
Fake channel
Fake network
You would really have to dig into the study to find out how fake it is. Is 1000 too small a sample? Probably not, so long as it was randomly selected.
But were the questions like "You don't approve of Trump discriminating against people of color, do you?", then you're going to get some fucked up skewed answers.
When a poll gives a number, they try to make sure that it's at least 95% confident that the true number is within +/- the error margin of their sample.
In other words, if their study of 1000 people produces 60% do not favor the Trump ban, at a 95% confidence level, with +/- 5 percentage points, they are telling you they think the real number has a 95% confidence of being between 55 and 65%.
"Do you favor Trump not letting terrorists come into the country to murder Americans?" would get different results.
(((Statistics)))
...
You guys need to get your shit together, like, quick.
>CNN is fake news
False.
Source: They cite their sources in the actual link, you've even acknowledged the existence of said source by admitting the poll exists.
CNN is fake news.
So said God Emperor Trump.
>CNN has an article claiming that the majority of Americans are against Trump's travel ban.
The majority of White Americans are in favor of it. Who gives a shit what the coloreds think?
Yeah, that's actually the only reason people think it's fake news. Although I'm starting to think you're only memeing it as fake.
>Everything with sources is true
Why are you watching CNN = fake news.
How stupid are you nigger? Its divided, you can get that by looking at any poll. Its no ones fault but your own that you went to CNN instead of checking actual polling companies. You rag on others for adding nothing and all you did was come into this discussion and tell everyone how fucking worthless you are.
Basic statistics uses 1000 sample size, because the math is easy. Stats has 3 averages; mean, median, and mode. Median is a true average, but you are mathematically correct to use the other two. Would have to see the questions and who they called.
What if they only surveyed Muslims in Michigan? What if they only called California? How were the questions asked, open ended or yes/no?
Basically with a number of 1000, you can make a poll to say whatever you want by affecting the sample and the questions. The math is fudgeable also by the mis use of the word average.
Shit because it's CNN, they might get 100 responses they don't like, throw those out and call 200 more to get the answer they want.
>not understanding the difference between a news station reporting on a study and a news station reporting unsubstantiated fictions.
Kill yourself if you aren't shill.
They provided quality statistics during last year.
As long as the sampling is truly random (this is key), 1000 people is plenty sufficient. You're looking for a proportion, so if the sampling is random enough, your 1000 person sample will be proportional to the whole population.
The bigger issue I think is people being hesitant to voice their true opinion in light of social pressure to maybe disavow people like Trump. Brexit polls being way off were probably due to similar reasons.
Polling agencies maybe need to re-evaluate how they poll. Maybe less phone calls for polling, or have the polling calls made by automated systems (as opposed to actual people) so a respondent feels less pressure to give the pollster a socially or politically correct answer?
And I'm sure you would all still be calling the article horseshit if they said the majority were in favor of the ban since it was CNN.
Is it fake? Or fake fake? Only time will tell. But yeah, fake.
are you trolling or did you just miss what I said entirely? OP asked a question and half the posts became "Hurr durr muh retarded liberals."
All I did was point it out.
youtube.com
This is the reason they're fake news.
320 million people and you dumbfucks had to install a clown into the fucking White House who is 24/7 on fucking Twitter. How does that make sense?
1000 people is a fantastic sample size
"It's illegal to possess these stolen documents"
But yeah, you tinfoil hat's interpreted that as "It's illegal to know what was written there"
the key word in SRS is RANDOM... yes the math is basically fullproof if its fucking RANDOM. The only way they could actually do a random sample is if they got a list of every american citizen, shuffle it up and pull names out of the bag
So really I have no faith in polls like these desu
Huh. I heard on NPR (left-leaning US radio) that the majority of Americans favored it. Maybe it was phrasing: muslim ban vs travel ban or something. Or, maybe there's an error somewhere.
The statistical theory is that if you take a sample from a population, which is REPRESENTATIVE of the population with regard to some measure X, as you increase the sample size (n), the measure X(n) [X in the sample] approaches X(N) [X in the population].
Now, the first thing to consider is sampling bias. Let's say we wanted to ask Republicans how they feel about the Affordable Care Act. What initial factors would result in a sampling bias?
> Only people who are interested enough in this issue to respond to our question will be included.
> If our data source is public, only registered Republicans rather than ideologically aligned Republicans/Independents/Nonpartisans, will be targeted for our questions.
> Only people who are willing to answer an unknown caller ID / respond to the stranger at the door will be represented in our data.
> People who are on vacation will not be included.
> If the questions are verbal, deaf people and those with speech impediments will not be targeted for our questions.
You see, this list can go on and on indefinitely, each item being less important and representing a smaller fraction of the Population (big P) being intrinsically left out of the data.
The same thing qualifies for representation. Statisticians will run multivariate analysis on a Population to determine which are the most important coefficients to X. It goes like this:
> What is the structure and proportion of races in Republicans?
> Income Level
> Education
> Geographic Distribution
> Age
> Health Status
> Religion
> etc etc
Often, researchers find that there are levels of stratification along these lines, and to make the sample have any predictive value on the population, these strata have to be replicated in the sample.
>statistics
you're a fool to look at these and think it's speaks for the bigger picture. people say things, people, lie, and people change. I wipe statistics with my ass.
This. People with no math background often think that statistics is some complex field so they don't dare to question dubious surveys in fear of being called stupid, not educated enough etc. Granted, I'm "only" an EE that minored in statistics but even so I have enough of an overview to say that it's one of the most trivial fields of mathematics (yes QM related stuff exists I know, but that's something else.)
People have to learn that statistics is more humanities than math and that means that you can skew it as you please and it's VERY prone to personal biases.
jesus fuck what is that? what's going on here? is this some kind of butchered post op?
Trumpkins just want to think they're in the majority. They'll believe anything that comes out of orange-man's mouth, and deny anything else.
> 1000 sample size because the math is easy
Brainless - what are computers? Having said that, bigger samples are more expensive so companies don't do larger than they need to.
Some of the rest of what you say contains truth (phrasing and representativeness of sample affecting the result), some is just crap - especially the average vs median vs mode stuff.
You've got some half understanding of statistics there.
What, you've never had the privilege of fucking a nice, tight MtF tranny you shitlord?
You saying he doesn't get enough done?
>i2.cdn.turner.com
They are including more Democrats (29%) than Republicans (25%) in this poll.
The "independents or members of another party" make up 45%, and they most likely are left leaning.
We don't know if these people are pre-selected by CNN to give a desired outcome.
All 1,000 people were probably CNN staff. As a matter of fact, the entire study could be completely made up for all we know. It's fucking CNN, after all.
random sampling.
the methodology says they did calls nationwide.
they fucked up by not saying if they did random digit dialing though, or if the interviewers went off script.
they could do random digit dialing.
All your did was say, "no that's wrong", you can't even correct me. It's probably you that has a half ass understanding.
Here's a primer to prove you have no understanding of the math.
purplemath.com
If you go to areas known to be more heavily left than right you can easily make the results say what you want. I can go to a deep red god fearing area and get the exact opposite with the right questions.
Always keep in mind figures don't lie but liars can figure.
Fuck Globalism. You and your fucking fake news will be silenced. Plus anyone covering up the child pedophile sex ring just remember this. Covering up or if you are part of a conspiracy of a crime will get you jail time.
It's not stats. People that visit CNN are liberal by a large majority. They don't say where these people are, who they are, ages, race and so on. So you can just assume they polled CNN viewers in California or New York.
no. but this guy is a joke. everybody starts to follow politics just to laugh at trumps tweets and soon to be war with: iran, best korea, chayna etc,etc.
sorry user, but pack your stuff, you are going to iran this year
they could have used random digit dialing.
sure hope they did
>taking polls seriously at all after the election
They also said Obama had a 60% approval rating when he left office, which would mean the nation would highly support having his designated successor in office and not one elected to repeal all of his bullshit.
>he actually tries to argue that CNN is not fake news
>what a colossal faggot
read the methodology.
dont try to act smart on Sup Forums if you dont even know what you are talking about
I love how it's always the most irrelevant nations on the planet spewing this shit
So the goal for a good researcher is to cast a wide net and capture as many of these variables as possible. The data-sampling itself must be done with excruciating care to avoid capturing a sample which is weighted in a way the Big P Population is not. This is easier said than done. it's the unseen effects that are difficult to measure. If you conduct some sort of sampling op during the day, you are intrinsically excluding people who work night shifts. If there is a physical location involved, all of your data is shifted with a gradient moving away from that location and it's demographic sample (median income, age, etc). If you use some other data set to produce yours, you incorporate their sampling bias, which may not even be well documented.
The truth is,
ALMOST NO ONE IN STATISTICS DOES THEIR DUE DILIGENCE.
Especially when it comes to sociopolitical and economic studies which are heavily infested with SCIENTIFIC VERMIN who produce data as a means to an end and have no interest in statistical rigor. Remember the 1.8% chance of a Trump presidency? Remember the D+12 polls, among a DOZEN (((independent))) "research" "institutes"? They were "correcting" the sample to "accommodate historical levels of turnout" (massage the data until it says what I want by making the sample unrepresentative of the population we want to study).
Here are three things to ask about ANY statistical shit you see:
> What sampling bias was introduced through the actual data collection strategy? (Hint: none is not an answer)
> -HOW REPRESENTATIVE- (with numbers) is the composition of the sample compared to the population in question?
> The larger the population, the more difficult a target. Saying X about "pediatric dentists" or "desert tortoises" is literally orders of magnitude easier than "Americans". Americans are so rich in data structure, assembling a statistically significant sample which rejects a null hypothesis with n = 1000 would be an incredibly difficult project.
>will be silenced
Not by your crippled ass
>covering up
The fact that "pizza" is the primary code word should tell you a lot about how this was conjured in-house (read: on *chans by a smattering of would-be-redeemed Crusaders who created all their own leads).
But you won't be redeemed. Your HDDs are still getting seized, and you're still going to prison for looking at kiddie porn. Externalizing an enemy when the enemy is yourself is bizarre and stupid.
Don't understand how anyone can think these are in any way accurate anymore. Not just cause the election, but the fact that so little people take part, and the only people who will even see the poll are the already anti Trump people who browse their site.
well, i guess you will have to prove that you are relevant one more time. and again and again.
good luck user
>proceed to partly answer OP's question
>retard claims shit posting
I swear to God everyone on /pol gets more and more autistic by the day.
We are all Americans, dumbass.
Sure but if I call orange country CA, Denver CO, Seattle Washington, Washington D.C.and other places that voted against trump, it could still be nationwide. If I used the exact same list that said Clinton has a 99% chance of winning the POTUS, that's still nation wide.
It's impossible to tell the built in bias by what they report, outside of this If you ask questions like "is that orange cheeto fucking up this country?" It has a built in bias vs "one to ten, how would you rate the president?".
Thanks kikeburger.
>Because a poll exists it's true data
Are you retarded? Of course the poll exists, what all these retards are discussing is the results and the actual groups surveyed.
Because no one even has to?
You go to Harlem and ask a thousand people and you will get a fundamentally different answer than if you went to Billings and did the same thing.
Polls like this are fake and pathetic. Snowflakes are just losing their shit and burning their own neighborhoods, and now universities, down because they cannot handle their emotions with how the government is not going to be "daddy" and give them free money for sex change operations any longer.
CNN is just another tabloid that is pandering to advertisers for profit only. They never once gave a shit about reporting news or informing anyone at all. They just know that if they bitch and whine about Trump a lot, then they will get all the viewers that bitch and whine about Trump a lot and that will please the sponsors. They are the literal definition of fake news in it's simplest form.
Good points
The methodology says it called both landlines and cell phones but doesn't specify the amounts.
People who use landlines are likely to be older and get their news from the television vs younger cell phones users who use the internet.
As rogan says, that's 100% sampling of retards who take the time to answer polls. AKA old people. Who do nothing but watch the news all day.
My guess is CNN has some sort of directory of what channel old people watch tied to phone numbers, it's the same cable. With that simple data you could construct a poll of a reasonable sample size, such as 1000, that says exactly what you want down to the decimal point.
Because the illegals in California looking for handouts are not liking it too much. Everyone else that was born here thinks it is perfect and should have been implemented a decade ago.
Wrong
Half of the pollsters dont really call and fill out form themselces. Fact.
>trusting CNN fake news
This is a longer form version of "the polls are fake".
Important point well made about the sampling biases in the statistics you use to generate your own statistics carrying through in to your own.
But as we know the typical CNN poll is simply a piece of propaganda with a pre-determined conclusion.
And indeed, it's true that even with a sample size of 1000, it's difficult to properly sample the US and a few random outliers could potentially end up messing with things.
That was an NYT line of attack against the LA Times poll - the best by far before the election. Apparently the LAT had a black south side Chicago guy who was saying he'd vote Trump. The NYT never memed that line of attack with much enthusiasm though because the idiots they try to trick usually don't have the patience for that sort of thing, that was the only juicy example they could find and, probably, they didn't want to open up an official public debate about polling methodology.
Believing in (((CNN))) (((POLLS))) after 2016, Nigger - PLEASE!
>you have no understanding "of the math"
>posts a link for children
>unironically doesn't know the difference between distribution and frequency
In a well-designed poll, the sample size can actually be pretty small and still convey decent accuracy. The "margin of error" given in the poll (they should list it somewhere) is basically the mathmatical limits in between which you can be 95% sure the sample correctly reflects the population. So, assuming the 1000 were a good representation of the american public (i.e., matched every relevant variable to their real proportion among all US citizens), then the margin of error should basically tell you where you can be reasonably statistically certain.
The problem isn't how may people they select; its how they pick the people they do. They could survey 500,000 people but if there is a lurking variable (or the pollster is biased, of course) , it wouldn't make a difference. The data would still be shit. Meanwhile, a poll of 150 people with accurate sampling representation can be fine. They just give you the margin of error, and that's that.
Unfortunetely, as the election proved, pollsters suck at choosing the correct demographic makeup. The statistics/math aspect has nothing to do with it, just shitty polling.
You keep it up Globalist, you are the scourge of humanity. You think you still have a chance of destroying the west? George Soros will be wanted and people like you will end up becoming a shadow of the past.
Traitors is what you will be rememberd as when you sold our nations and tried to destroy human progress. Communism, progressiveism and marxism. All of it will be wiped out. You lost our revolution has won, the best thing you can do is surrender and just leave people the fuck alone and abandon these ideologies.
Gee, it's not like retired old people are home all day to take phone calls, you stupid niggerfaggot
The election proved all of CNN's polls were shit and bias.
Why should we believe this poll?
Some are only 3/5ths american
Better than 0% American
If you guys had to guess, what percentage of Sup Forums users have taken an actual statistics class you think?
that image is magical
The term is "fake news". Biased polls used to make "news" stories "proving" what the mainstream media wants you to believe. The election was not really an endorsement of Trump; it was a rejection of the fake news industry. According to the polls, Britain rejected Brexit, and Hillary won the election with 98% of the vote.
Google "Maximum Likelihood Estimate Binomial Distribution" and you can find your answer.
Assuming all American are either FOR or AGAINST (binary outcome), their opinions are independent of each other and identically distributed (each person has a same p). Then we can assume the sample to follow a Binomial(p,1002) distribution where p is, say, the probability of person being AGAINST and is a parameter we want to estimate.
The MLE of p# = k/n where k is the number of people that answered AGAINST. (Here I use p# because I need some way to mark that it is an ESTIMATE of the real parameter)
The Fisher Information IE(p) = n/(p(1-p))
A 95% confidence interval is:
p# +/- 1.96 * sqrt(p#(1-p#)/n)
p# = 531/1002 = 0.53
0.53 + 1.96 * sqrt(0.53*0.47/1002) = 0.5603
0.53 - 1.96 * sqrt(0.53*47/1002) = 0.4985
Unfortunately, CNN does not have significant evidence at 5% level that p does not equal 0.5 (i.e. that you have a majority as they claim).
This is the frequentest approach that gives CNN the benefit of the doubt (about their sampling being random and representative of the population).
Potential sources of bias:
>Reporting bias (people are more likely to openly admit opinions that fit the official narrative, this is the biggest one. Someone calls you in the middle of the day and asks you if you hate Muslims, lul)
>CNN might not sample randomly.
>CNN might fiddle with the data. (Analytical bias)
>Selection bias (CNN followers might prefer one opinion over the other).
We could use Bayesian approaches with arbitrary prior knowledge parameters to account for this, and we might even come to the conclusion that the opposite is true, i.e. you have evidence that you have majority FOR the policy. Ofc these parameters depend on the expert/statistician and are subjective - but that can't prevent you from eye-balling the 'real' result using your understanding of the world (CNN = Fake News).
Yup.
Fake news. And literally by definition. A news organization conducting or paying for a poll and then reporting on its own poll is the height of fake news.
Don't make yourself the fucking story, CNN. Ever. You are fake news.
don't trust statistical claims about a 320 million population with a sample size of 1000
this is the same kind of research that propped up their fantasy that she was going to win
Even if random sampling was done, a false majority could easily be selected by chance.
They would need to do many 1000 person samples and average the results together. All without overlap and it would still be guranteed to be off by a little.
>Even if random sampling was done, a false majority could easily be selected by chance.
You can make this "argument" for any sample size because it's always a possibility and "easily" doesn't actually mean anything.
We have elections because polling doesn't work. Trump won the elections so the policies he ran on get enacted.
Its happening youtube.com
For reference, the Brexit referendum is a sample with n = number of people who voted with valid ballots vs. the entire population.
99.999% confidence interval for p is (0.5186,0.5193). "Leave" definitely has a majority)
Thank you Nate Sliver.
Best answer.
The jew knows his statistics.
Remember.
All media polls are always accurate.
Fuck white people and fuck drumpf
Over sampling...same reason they said Hillary had a 99% chance of winning the Presidency.
Its what people in the "know" like to call "Fake News". Just ignore it and it'll go away.
Are you retarded or just pretending? Serious question.
Random sampling nigger. The catch is that the sample they use isn't really random.
Reuters source
reuters.com
49% agree with ban, 41% disagree
