• Trump tweets fake poll about the popularity of the travel ban that doesn't even exist
    81 replies, posted
[QUOTE=TornadoAP;51797738]This is fucking terrifying news if this is true. A demagogic, racist, sexist, power-hungry president in charge of a completely corrupt party that controls the entirety of a nation's government is already scary as hell, but when the populace refuses accept the real truth, but accepts the one manufactured by that government, that's when you know shit is absolutely fucked.[/QUOTE] There is no such thing as truth anymore. Only alternative information.
[QUOTE=Helix Snake;51797907]Can we please have a sticky explaining how polls work or something, this shit is in almost every topic involving polls[/QUOTE] How polls work as we currently are okay with is shit. You can't actually expect an accurate representation of a country of 360,000,000 people with a polling of even 3,000. I don't know how people even actually fool themselves in to thinking this is remotely rational, especially with the multicultural and extremely large in area country that is America. 2,000 fucking people do not represent the whole country or even serve as a good polling size. Polls are a joke. [QUOTE=Psychokitten;51797202]"I reject your reality, and substitute my own!"[/QUOTE] While I think Trump is going to be a below average president, I think the American Liberals are doing exactly this right now.
[QUOTE=Revelificent;51801624]How polls work as we currently are okay with is shit. You can't actually expect an accurate representation of a country of 360,000,000 people with a polling of even 3,000. I don't know how people even actually fool themselves in to thinking this is remotely rational, especially with the multicultural and extremely large in area country that is America. 2,000 fucking people do not represent the whole country or even serve as a good polling size. Polls are a joke.[/QUOTE] Oh for fuck's sake... Go take a course in statistics.
[QUOTE=_Axel;51801671]Oh for fuck's sake... Go take a course in statistics.[/QUOTE]Can you actually explain how they work or do you just believe everything people tell you?
[QUOTE=Ridge;51797765]Wait, if the travel ban doesn't exist, what is everyone getting upset over?[/QUOTE] Isn't it still being enforced though?
The party of anti-science and anti-intellectualism has reigns on the narrative now, and even when things are explained AD-NAUSEUM to people in a scientific and objective fact-based way, posts like that are the result. Why is this where we are now. How can people see non-biased, long term, well tested scientific evidence as to something like how sample sizes work, posted repeatedly, and then just throw it all out and make up some shit on the spot as if they know better. I am sick of the feels before reals world.
[QUOTE=Xion21;51801696]The party of anti-science and anti-intellectualism has reigns on the narrative now, and even when things are explained AD-NAUSEUM to people in a scientific and objective fact-based way, posts like that are the result. Why is this where we are now. How can people see non-biased, long term, well tested scientific evidence as to something like how sample sizes work, posted repeatedly, and then just throw it all out and make up some shit on the spot as if they know better. I am sick of the feels before reals world.[/QUOTE]You do know that Science can be wrong and doesn't have everything figured out yes?
[QUOTE=Revelificent;51801624]How polls work as we currently are okay with is shit. You can't actually expect an accurate representation of a country of 360,000,000 people with a polling of even 3,000. I don't know how people even actually fool themselves in to thinking this is remotely rational, especially with the multicultural and extremely large in area country that is America. 2,000 fucking people do not represent the whole country or even serve as a good polling size. Polls are a joke.[/QUOTE] Because this has to be answered in every thread: [quote]3. How can a sample of only 800 or 1,200 truly reflect the opinions of 200 million Americans within a few percentage points? Sampling methods and measures of sample reliability or precision are derived from a mathematical science called statistics. Statistics is a subject taught in colleges and some high schools. Text books on the subject are available in most libraries. At the root of statistical reliability is probability; i.e., the odds of obtaining a particular outcome by chance alone. As an example, the chances of having a coin come up heads in a single toss is 50%. Heads is one of only two possible outcomes. The chance of getting two heads in two coin tosses is less because two heads are now only one of four possible outcomes; i.e., a head/head, head/tail, tail/head and tail/tail. As the number of coin tosses increases, it becomes increasingly more likely to get outcomes that are either very close to half heads or exactly half because, as with two coin tosses, there are more ways to get such outcomes. Sample survey reliability works the same way -- but on a much larger scale. As in coin tosses, the most likely sample outcome is the true percentage of whatever it is we are measuring across the total population. Next most likely are outcomes very close to this true percentage. A statement of potential margin of error or sample precision reflects this and often appears in poll stories. Using a sample of 1,000 as an example, the statement could read: the chances are 95% of coming within +/- 3% of a hypothetical survey conducted among all members of the population. This means that 95% of all samples that could possibly be drawn will yield an outcome within 3 percentage points of the true percentage among the population. Keep in mind that estimates of potential sample error always assume random samples. But even in true random samples, precision can be compromised by other factors, such as the wording of questions or the order in which questions were asked. There is no single ideal sample size. Samples of any size have some degree of precision. The question is always whether there is sufficient precision to draw conclusions as determined by statistical formulae.[/quote] [url]http://www.pollingreport.com/ncpp.htm[/url]
[QUOTE=KillerJaguar;51801797]Because this has to be answered in every thread: [url]http://www.pollingreport.com/ncpp.htm[/url][/QUOTE]First thanks for answering, I've never seen it mentioned before and I don't browse SH often. As nice as that post is however, it doesn't address the whole cultural differences problem I mentioned based on where you are in the United States and what the area is like. A coin toss has two single options each throw with no complicated philosophy or opinion behind them. A coin also can't purposely vote the opposite option from what they believe to influence the stats, nor is it particularly easy for an organization to hide political bias in their statistic gathering, unlike what they could probably do with a coin. The complexity of a country and its people already makes a poll of 2-3000 people very poorly indicative of the entire country wide thought patterns.
[QUOTE=Revelificent;51801836]First thanks for answering, I've never seen it mentioned before and I don't browse SH often. As nice as that post is, however, it doesn't address the whole cultural differences problem I mentioned based on where you are in the United States and what the area is like. A coin toss has two single options each throw with no complicated philosophy or opinion behind them. A coin also can't purposely vote the opposite option from what they believe to influence the stats, nor is it particularly easy for an organization to hide political bias in their statistic gathering, unlike what they could probably do with a coin. The complexity of a country and its people already makes a poll of 2-3000 people very poorly indicative of the entire country wide thought patterns.[/QUOTE] I apologize for being snarky, I am just sick of good people getting banned in this section because they are getting tired of arguing with people that use milkandcooki level "debate" tacticts and get angry. While what you are saying about cultural differences is certainly a valid point that I understand your stance on, that is basically the entire point behind an anonymous poll. The more you tamper with the process, like by making sure a certain requisite number of X religion or X race get in, the less random and more engineered, and thus less scientific, the results are. The idea behind a truly scientific poll is basically "randomness," let people poll, and as long as the sample size is correct for the population intended, and it wasn't deliberately tampered with, like by excluding, OR including certain groups purposefully, the results will bear out correctly every time. There can be no political bias in the gathering of the statistics if the gathering process, and the people involved, do not directly interact with the population involved, like trying to get a bunch of a certain race in the poll that they know would oppose a certain issue etc. That's the point essentially. Also, numbers in the several thousand are known to give SEVERE diminishing returns on accuracy and are often considered a waste of time. If ~1,200 can give you in the high 90 percents, is an extra thousand people really worth the effort for an extra like ~.5%?
[QUOTE=Revelificent;51801676]Can you actually explain how they work or do you just believe everything people tell you?[/QUOTE] It's part of my background in maths as an engineering student. Essentially the width of your confidence interval doesn't depend on the size of the polled population. Provided the sampling is correctly done, it's equal to k/sqrt(N) where k is the constant that corresponds to amount of confidence you want (for instance if you want to have a 95% chance of having the actual distribution fall within your confidence interval you take k=1.96) and N is your sample size. If you plot N->1.96/sqrt(N) you can see that the CI width gets quite narrow pretty quickly, and increasing the sample size gives diminishing returns once you get past a few thousand people. [IMG]https://i.imgur.com/tXCDmxi.jpg[/IMG] Considering polling people costs time and money, there's no point in having a sample size any bigger than that. For a sample size of 1,000 people you get a margin of error of k/2sqrt(N)=0.98/sqrt(1,000)=3% If you quadruple the sample size you divide the MoE by 2, so for a sample size of 4,000 you get MoE=1.5%, MoE(16,000)=0.75% etc... That's provided your sample is representative of course, quality polling requires appropriate sampling methodology more than it does raw numbers. For instance if 40% of Americans are Republican, 40% of your sample should be too. If you want to bring the legitimacy of a poll into question, it would be much more relevant to criticize the sampling method than the sample size, especially if it's within that ~1000 sweet spot. [editline]10th February 2017[/editline] A quick Google search brings up demonstrations if you're interested: [url]http://stats.stackexchange.com/questions/166/how-do-you-decide-the-sample-size-when-polling-a-large-population[/url] (2nd answer) [editline]10th February 2017[/editline] [QUOTE=Revelificent;51801701]You do know that Science can be wrong and doesn't have everything figured out yes?[/QUOTE] I also suggest you look into epistemology.
When I was in high school, I was also taught to pay attention to "selective placement" for physical polling places. My teacher gave the example of a public polling place that was used for opinion polls like these by certain people, and would always be touted under that exact phrase. "Public Opinion Polling." To make it sound more legitimate. However, as it turns out, the results were always heavily right leaning, despite there being no evidence of direct tampering with who could vote there. Why? Even though it was "public" (as in, you couldn't be turned away from voting there) the center was in a fire station in an extremely wealthy gated community. Which did not provide a balanced cross section of people at all. A better place, my teacher said, would be to set up a booth in a mall or something like that. Just an anecdote to add to the idea of population representation.
[QUOTE=Xion21;51801894]I apologize for being snarky, I am just sick of good people getting banned in this section because they are getting tired of arguing with people that use milkandcooki level "debate" tacticts and get angry.[/QUOTE] To be fair, I posted earlier with the best of intentions and thought that was a legitimate thing to point out. I wasn't aware that I was, by definition on these parts, shitposting, nor I wasn't at all planning on getting people pissed off and "getting good people banned". Just because I only have double digits to my posts doesn't mean I exist to piss people off. Again, yes I agree that it's the sampling that is important.
[QUOTE=Revelificent;51801836]First thanks for answering, I've never seen it mentioned before and I don't browse SH often. As nice as that post is however, it doesn't address the whole cultural differences problem I mentioned based on where you are in the United States and what the area is like. A coin toss has two single options each throw with no complicated philosophy or opinion behind them. A coin also can't purposely vote the opposite option from what they believe to influence the stats, nor is it particularly easy for an organization to hide political bias in their statistic gathering, unlike what they could probably do with a coin. The complexity of a country and its people already makes a poll of 2-3000 people very poorly indicative of the entire country wide thought patterns.[/QUOTE] Look, as somebody who tried to make the exact same arguments you are making a few weeks back, trust me when I say you're absolutely wrong. Polls around 800-1200 are ridiculously accurate, even if at first common sense makes it seem like that is impossible. I looked into this extensively when somebody pointed this out to me, and I had to admit that I had no idea what I was talking about. Trust me, the sample size is fine.
[QUOTE=Xion21;51801894]I apologize for being snarky[/QUOTE] Oh it's completely fine. I guess I should stop acting like I'm 100% correct when in reality I'm still very ignorant on what I'm discussing. I've just always questioned the accuracy of polling. I guess I should have actually done the proper research instead of just assuming I know more than the poll scientists who have dealt with this for a long time and DO deal with it from day to day. Fair enough guys, it does seem you are all definitely more knowledgeable than I am on it. My main issue though is that because people can vote 85% liberal in one place and 85% conservative (examples) in another area it seems like that would throw the whole poll off its rocker. I'm guessing there are averages and these sorts of things are accounted for in the actual polling itself? I would like to point out that Hillary Clinton lost despite polls showing in her favor. I'm just thinking about the complexities, like for example people who would vote for Trump despite being liberal, or the people who say they will vote for Hillary but only go that way out of fear of being judged. Maybe these complexities are less of an issue in polling than I was thinking. Edit: I never try to act like I'm right even in the face of massive evidence. I'm just really ignorant and I can come off as a knowitall sometimes because of my past and how I used to be.
[QUOTE=Revelificent;51805677]Oh it's completely fine. I guess I should stop acting like I'm 100% correct when in reality I'm still very ignorant on what I'm discussing. I've just always questioned the accuracy of polling. I guess I should have actually done the proper research instead of just assuming I know more than the poll scientists who have dealt with this for a long time and DO deal with it from day to day. Fair enough guys, it does seem you are all definitely more knowledgeable than I am on it. My main issue though is that because people can vote 85% liberal in one place and 85% conservative (examples) in another area it seems like that would throw the whole poll off its rocker. I'm guessing there are averages and these sorts of things are accounted for in the actual polling itself? I would like to point out that Hillary Clinton lost despite polls showing in her favor. I'm just thinking about the complexities, like for example people who would vote for Trump despite being liberal, or the people who say they will vote for Hillary but only go that way out of fear of being judged. Maybe these complexities are less of an issue in polling than I was thinking. Edit: I never try to act like I'm right even in the face of massive evidence. I'm just really ignorant and I can come off as a knowitall sometimes because of my past and how I used to be.[/QUOTE] Sure that COULD happen, but every single person you add halves the number of people that could be in a "split" region that is unaccounted for. Think of it this way; if you have 100 marbles in a bag that are either red or blue, and you pull out 50 of them completely at random, and they're all red, it is extremely probable that the bag contains mostly if not entirely red marbles. You can't be [I]certain[/I] of it, but you can repeat the same test, and if you get similar results, the probability of accuracy also increases. In a huge population, say, 300,000,000, the ratio of your sample size actually shrinks by a lot- because the curve's logarithmic, mostly only the sample size matters, and not the total population size. If your sampling is completely random, and you have over 500 samples, almost no matter the population size this can be considered to be a fairly accurate cross-section. For comparison, to get a representative sample with a 99% confidence rate and a +- 5 % confidence interval of the entire population of the world, (7,000,000,000), you would need only ~666 individuals, if they were selected randomly. [editline]10th February 2017[/editline] check [url]http://www.surveysystem.com/sscalc.htm#one[/url] for a calculator
[QUOTE=Radical_ed;51805969]Sure that COULD happen, but every single person you add halves the number of people that could be in a "split" region that is unaccounted for. Think of it this way; if you have 100 marbles in a bag that are either red or blue, and you pull out 50 of them completely at random, and they're all red, it is extremely probable that the bag contains mostly if not entirely red marbles. You can't be [I]certain[/I] of it, but you can repeat the same test, and if you get similar results, the probability of accuracy also increases. In a huge population, say, 300,000,000, the ratio of your sample size actually shrinks by a lot- because the curve's logarithmic, mostly only the sample size matters, and not the total population size. If your sampling is completely random, and you have over 500 samples, almost no matter the population size this can be considered to be a fairly accurate cross-section. For comparison, to get a representative sample with a 99% confidence rate and a +- 5 % confidence interval of the entire population of the world, (7,000,000,000), you would need only ~666 individuals, if they were selected randomly.[/QUOTE]See my problem from this comes from like I said earlier, that if you completely pick a random group of people from all over, what about the areas you select from that are almost going to be completely liberal or completely conservative? If you pick 1000 people for example but you pull those mostly from the liberal areas then your numbers are going to show a preference in the liberal area. I don't understand how a random 1000 people that you pull from various areas is going to accurately give you a good representation, especially when some areas have far more liberals or far more conservatives. The 50 people you poll in a liberal place that all vote liberal could skew the whole poll no?
[QUOTE=Revelificent;51806226]See my problem from this comes from like I said earlier, that if you completely pick a random group of people from all over, what about the areas you select from that are almost going to be completely liberal or completely conservative? If you pick 1000 people for example but you pull those mostly from the liberal areas then your numbers are going to show a preference in the liberal area. I don't understand how a random 1000 people that you pull from various areas is going to accurately give you a good representation, especially when some areas have far more liberals or far more conservatives. The 50 people you poll in a liberal place that all vote liberal could skew the whole poll no?[/QUOTE] That's the entire point of sampling properly and actually being random. The 50 people you poll in a liberal place won't skew the poll, because if it's truly random you're going to get 50 people in a conservative place as well. Statistics are very, very interesting when it comes to this. I recommend reading up about it, the problem is always going to be whether or not you sampled properly, not the sample size when you get around 800 people.
[QUOTE=Revelificent;51806226]See my problem from this comes from like I said earlier, that if you completely pick a random group of people from all over, what about the areas you select from that are almost going to be completely liberal or completely conservative? If you pick 1000 people for example but you pull those mostly from the liberal areas then your numbers are going to show a preference in the liberal area. I don't understand how a random 1000 people that you pull from various areas is going to accurately give you a good representation, especially when some areas have far more liberals or far more conservatives. The 50 people you poll in a liberal place that all vote liberal could skew the whole poll no?[/QUOTE] Sampling is like a whole field of its own, and there are numerous strategies to get around those problems. Look into [URL=https://en.wikipedia.org/wiki/Sampling_(statistics)#Stratified_sampling]stratified sampling[/url]. In short, they have for probably considered and accounted for your, and probably many other less obvious, concerns.
[QUOTE=danjee;51796418]I wish politicians were ever held accountable for lying[/QUOTE] that's the electorates job.
Thanks guys!
Sorry, you need to Log In to post a reply to this thread.