The media will used biased polls to push the narrative that Trump lost the third debate. This will discredit his campaigning methods and allow cuckservatives to fill the airwaves with muh pr, instead of talking about wikileaks and other pieces of actual news. This has to be stopped. Pic related (from debate 2) and others were likely engineered to give a certain result by their publishers. This has allowed people to retrospectively characterise Trump’s rhetoric as failing, when nothing could be further from the truth. I’m going to explain some common errors with polling so that anons can explain to normies precisely what these luegenpolls are doing to manipulate the information and by extension them. If you can actually explain the reason why you claim a poll is wrong it gives your arguments extra weight. I saw the Brexit vote close up, and there are parallels with that and the USA presidential election.
This change of situation will make the polls more wrong than usual. I will use the stats from that vote, as opposed to the 2012 election, to predict what may happen in November. I firmly believe Trump will win, and maybe I’ll go through some reasons why another day. Here I just want to explain why polls are so bad. I’ll write about a few common and easy to communicate errors or obfuscations. If you’re just too busy/lazy to read more than one part, read section 2 on sampling.
Every poll has raw information, most polls will manipulate this information, sometimes even for honest reasons. I’ll explain a couple of these manipulations so you can spot a bad poll. To begin, I will list the most important pieces of information: The first lesson is this: If you see a poll that withholds any of the above, ignore it. I think this is self-explanatory.
Cooper Thompson
Note that sampling aims to recreate voter demographics, not necessarily population demographics. If there is a strong relationship between likelihood to answer a poll and likelihood to vote, the poll will be more accurate. In the past, they could get away with this because those who didn’t vote didn’t answer polls either. That’s changed this year, one of the main reasons that the polls will be shown to be wrong – like in the Brexit vote - is that the unpoll-able demographics – particularly working class men – also happen to be a core component of Trumps’ support. Even if they were being entirely impartial, they would be underrepresenting Trump’s support due to systematic polling flaws.
As an example, consider the number of people of each sex who vote. In 2012, the vote split (according to the Roper centre archive.is/StIkr) was 47/53 in favour of women. Polls will attempt to sample a gender split that matches the 2012 vote. This won’t be the case this time, I can guarantee you. Polls should be looking at closer to 50/50, though probably still in favour of women (maybe by 1-4%). Imbeciles like Nate Silver hide behind the shield of `this happened last time’ while talking about this being a `change election’. Also look at the party they declared they voted for in 2012 since we know the precise results for that election. It is astonishingly common for polls to claim that in 2012 Obama got 60%+ of the vote. There is also a very common trend in under-sampling independents, based on the idea that they didn’t vote often in the past, they hence are unlikely to vote now, despite the Trump (and Sanders) effect.
Liam Turner
To look at a specific poll, let’s look at a recent Arizona poll (10-15 October) (assets.documentcloud.org/documents/3144924/The-Arizona-Republic-Morrison-Institute-Cronkite.pdf). This puts Clinton 5 points ahead in a fairly solid red state. A rudimentary glance at pic 2 tells us there are more than 200 democrats too many. Though there is some weighting going on, it cannot close that natural gap Clinton is provided. However, don’t ignore the info entirely. Shit polls can still provide useful information. For example, if a slew of polls that are equally bad shows a trend for one candidate, you can begin to think that that trend is real. Here, the bombshell is between the second and third pictures. Whereas there are 413 registered democrats, when asked about party allegiance only 316 say they feel connected to the democrats. AT LEAST ONE QUARTER OF REGISTERED DEMOCRATS SAMPLED NO LONGER VIEWS THEMSELVES AS STILL DEMOCRAT That’s fucking crazy. Trump landslide is on. Pic 3 suggests they mainly went in the `unsure’ category, but some may have switched to the Republicans. There is a reasonable cohort of anti-Trump republicans, but they are themselves more than cancelled out by Democrats who now associate with the Republicans or Independents who will vote for Trump, not even counting those who are `unsure’ but are in actuality shy Tump supporters. If there is a shy Trump effect going on, it will most likely be with coy registered democrats who simply won’t vote Clinton.
Austin Thompson
Weighting is raising or lowering the value of a person’s vote in a poll to better represent the true population. For example, if I ended up with a poll consisting of one third men and two thirds women, I could give each man’s vote double the value and re-calculate the results. This would make the sample `act’ like it was 50/50. Obviously, when we talk about many different demographics, this process can be especially dubious, completely reliant on the subjective opinion of the pollster to guess how likely a demographic is to vote. The most common is weighting to represent a 47/53 W/M balance.
This is seen in a recent Monmouth University poll (archive.is/Zhsd2). This gave Clinton a 12 point lead with 50% of the vote. Buried within the raw data, we have the weighting process laid out for us, so at least they are fairly transparent. The unweighted data had around 15% 18-34 year olds, probably a bit lower than my own rough guess at what the total vote share of this demographic will be on Election Day. However, weighting gave them 25% of the total. This corresponds to an effect close to simply doubling the weight of their vote. Generally, you don’t want one person’s vote to have a far greater weighting than others in a poll.
Why does this give Clinton an advantage? Firstly, the youngest cohort are more likely to vote democrat. Secondly, they are very unlikely to vote. The Roper center tells us that 18-29 year olds made up 19% of the total vote in the 2012 election, meaning 18-34 was possibly 25%. That was then, however. To get a sense of what proportion of the total vote will be 18-34 for this `change election’, let’s compare with the more similar Brexit vote. In this large scale voter survey after the vote (archive.is/yVZrc) (see P5 of data table linked at the bottom) 18-34 year olds made only around 19% of the total respondents who claimed they voted, hence it is likely that in the similar vote in America they will make a similar proportion of the total. Add to that extra weighting that lowers the frequency of deplorables in the final sample (e.g., working class men, people without degrees) justified by the same subjectivity that made the pollsters wrong in the Brexit vote, and you get a Clinton lead literally conjured out of thin air.
Evan Thomas
I think that plenty for now. I could write about likely voter filters or weaknesses in methodology but this is long enough as it is. When CNN proudly states that 500 people say Clinton wins the third debate and Trump is `dangerous’, sift through the data, find the errors, tweet at e-celeb whores or media types and ensure that normies will know better than the media lets them. Remember to check the demographics, and if it looks dodgy, it probably is. Even the most blue-pilled of normies can be convinced of the flaws of these polls with a bit of effort.
One final note - there’s no such thing as a scientific poll. When people refer to scientific poll, what they mean is `sampled and weighted to match my assumptions’. People who use the term `scientific poll’ don’t know what they are talking about.
Anthony Brown
Good thread OP. Until now I've just been using "muh undersampling of independents". Have a bump
Juan Ramirez
I'd love to see a normie friendly article along these lines. People need to fucking understand this.
Tyler Collins
Yes, This.
Carson Howard
Statistics is a particularly difficult subject for people without a mathematical background to wrap their heads around
Landon Powell
Some added info.
Brody Price
I had been thinking about writing something that easy to understand but many normies just can't cope with numbers beyond 100. I have more stuff I could write about anyway that I'll probably post later.
True enough. Though as someone with a mathematical background, I don't consider this to be particularly mathematical. The really mathematical part is where they calculate margin of error using exceptionally complicated statistical techniques that not even they understand and usually aren't relevant or rely on assumptions. I think they throw in fancy equations to confuse readers without understanding it themselves.
That `bootstrap confidence interval' is a case in point. I have no idea how that is supposed to spot a systematic error in methodology. Maybe there is a theorem buried in some textbook that states if margin of error of mean is low, theres a high chance of the poll being accurately sampled, but even then they need to be vigilant about methodology errors. Every single one of those adjustments is not taking into account the increase in republican turnout or the increased likelyhood of independents voting. These kinds of assumptions were a big reason why the brexit polls were wrong, for example there was at least one poll where they assumed undecideds broke 2/1 for remain because at some point they collectively decided that was what would happen.
Xavier Wilson
We need to be careful we don't end up doing the enemy a favor.
When you think your candidate is winning your effort drops significantly. It may not drop in high energy places such as this but normies energy drops rapidly. If people think Clinton is winning they are actually doing a disservice to their own normies and making them try less. The Clinton normies will get complacent and then start to assume they don't need to put any more energy in.
I feel there's a chance this could spread to our Trump normies if we're not careful.
Tyler Cook
I guess that by raw information you mean raw data. If we could get the raw data of those polls, then we'd be in business.
Charles Fisher
The only way to stop it is sleuthing to show a conflict of interest, then calling the cunts out.
Check the names of people associated with the polling agencies in wikileaks, veritas videos and other places on the internet.
Charles Sanders
They're getting really desperate.
Mason Bell
Just make sure that along with showing this data, you remind people that if polls are manipulated, it gives them more cover for rigging the vote as well, we regardless of what the real numbers are, we need a massive voter turnout to overwhelm any rigging or fraud attempts.
An alibi for their rigging is the reason they are doing this, which is why they do it despite the potential for it to de-energize the Clinton voters. Hopefully it will backfire and their lower turnout versus our monster vote turnout will make Trump win despite rigging and fraud just like he won the GOP nomination despite rigging and fraud.
Parker Carter
In what world is that third picture accurate? 40 vs 25? They cannot expect people to swallow that.
Benjamin Morris
Truth. I have ramped up to evangelizing Trump to the normies on a daily basis. I've networked with known supporters at my employment and try to whitepill them anyway I can (and subtly shitlord/repdill them in the process).
Aaron Richardson
They staged all of it, and have been planning it for years.
First, they have been setting up the male vs female dynamic in media since they knew they were going to be putting Hillary in. Its been everywhere, movies, news reports, commercials, tv series, for those who know to look. Its as if we have had 4 years of a single scene where the white male character is depicted as being incompetent, evil, cucked, foolish, nepotistic, perpetrator etc and the female/sexual deviant/minority is cast as the hero, savior, righteous, intellectual superior, victim, etc.
If they were simply anti Trump they would have released the "Trump's a sexist" tapes in the republican primaries, but they didn't, that sat on them waiting until the October presidential debates. Then they start tweaking the polls, pretending its a reaction.
Don't ever trust polls in this election. The media narrative is that to support trump is racist and sexist. They did this to trigger the majority of the voters against him. So a large portion of people are unwilling to publicly admit that they are voting for Trump, because they know they will be branded.
The whole thing is fucking gay, just like their gay-ass skits.
Jonathan Walker
Trump normies will battle bears and wrassle alligators to get to the polls and vote for him. Dont underestimate the trump normies when he holds a rally ANYWHERE it is always a packed house and there are always tons that don't get in. This is gonna be great.
