Acts20.com

Recently when Brexit occurred it was a shock to the world. Polling data had shown that Britain would stay in the EU. I am wondering if all of the polling data is as accurate as it once was. With so many people who don't have a land line, ( www.digitaltrends.com/home/you-can-still-call-about-40-percent-of-u-s-households-on-a-landline/probably less than 40% now since this article is a year old)how are these folks being polled. Doesn't that skew results. I believe it does and I doubt the validity of some of the polling. If the sample is down to those who have a land line and will speak to a pollster, polls are becoming more and more reliable. Perhaps this election will tell the tale. We shall see...I know what NBC, CBS, ABC and CNBC say about this. Nah, they(polls) are accurate. They won't be wrong. But they invest a lot of money on these polls and they need them to be right. They also have influence. They need older voters to say, the polls are right, what is the need of my participation? and younger voters to show up.We shall see...www.cnbc.com/2016/07/04/why-the-majority-of-brexit-polls-were-wrong.htmlI believe Trumps biggest asset is the folks that will show up in a hurricane to vote for him. The Hillary supporters (many young voters who have never voted before) won't need much to keep them away.We shall see...Here is another interesting article on how polling is done...www.electoral-vote.com/evp2016/Info/polling-faq.html

Re: Are the polls accurate now?

back to the topic, yes I think there are some inaccuracies. There are wildly different results and most likely because their sampling is heavily their own readership. ABC is the bastion for all liberal thought and their poll has Hillary double-digits ahead while nearly ALL others have either candidate within the margin of error. I heard all kinds of newscasts saying Texas is now purple. I don't believe that for a second. We'll see on Nov. 8 I guess.

very possible! Some facts but mostly just my [email protected]/

I hope it don't hit here again, I just got cleaned up at our new-to-us house. Some facts but mostly just my [email protected]/

Diakoneo, I had not commented before on this thread, and what I was going to say wouldn't make nearly as much sense now, lol, but I was going to say, no it wasn't like Brexit, and that our polls are more accurate, and that they do poll cell phones.I answered a poll the other day, and after all those questions, the young lady on the other end of the line wanted to verify that this number is associated with a cell phone? I told her I was on a landline. I wondered then if she had to call a certain number of cell phones and that perhaps she had already reached a quota on landlines and that we had both just wasted each other's time.One other thing I wanted to mention was the margin of error. Some people don't seem to accept that even if they are giving mental assent to the concept.It's like when our weatherman says 60% chance of rain. Some people think that means it WILL rain and that if it doesn't rain on their own head, the weatherman was wrong. In a similar way, if a candidate is ahead, but within the margin of error ahead, it doesn't mean the poll was wrong if the candidate loses by a small percentage of votes.For example, NC was a battleground state. As election day neared, the polls I was reading showed the race getting closer and closer. A few days out, I read an article on HotAir that cited a WRAL (Raleigh TV station) poll that had Trump up by a point or two. I distinctly remember reading the poll the day before the election and it had Trump up by about 3 which I mentioned when I was on the phone with my father in the morning of election day.

A lot of the accuracy of the poll is also the degree of confidence. For example, if the there is a 3% margin for error with a 95% degree of confidence, a candidate could be predicted to have 47% of the vote and the actual could be between 44 to 50% and the poll would have still been accurate. People tend to zero in on the initial prediction without the tolerance, thus they think if the candidate was predicted to get 47% and they actually got 44, they think it was wrong.One of the polling problems is that in order to decrease the margin for error to be .001 percent (versus the typical 3%) , the increase in data points (sample size) would have to be increased almost geometrically. the same applies if the degree of confidence was increased to 99.99 percent. From a financial viewpoint, that type of accuracy is not likely to be affordable in many cases. In a close election, even a 1% margin is far too large to be acceptable for projecting an actual winner. That said, from what I've heard, it seems the primary polling error in this case is the over-reliance on using the 2012 results of actual voters. First, that election was won by a democrat, thus the percentage of likely voters were also more likely to be democrat, up front. Second, with the unusually high turnout in 2012 were blacks voting for a black president, those numbers would again be skewed (unless Hillary could get the same number of black voters, which she did not do). Therefore, pollsters should have realized the odds of blacks showing up for Hillary would surely be less than blacks showing up for Obama. By relying on the flawed logic of using 2012 likely voters, and ignoring the very real possibility that trump was attracting voters who vote very infrequently, they underrepsented that element of the population. I suspected the election could be closer than projected because of those flaws, but I also realized that the margin of error was such that Hillary could have had 50% of the vote versus what may have been an expected 47%, thus potentially eliminating the offset of the new voters Trump attracted. Nevertheless, I would have also been fooled and my internal feeling was Hillary would win. So, I wouldn't have done any better than the Pollsters, but I'm not a statistician either. I think part of the rigged election claims by Trump was based in part on the attempt by the media to discourage Trump voters that it was a done deal and Hillary would absolutely win, therefore, don't bother to vote. I think that was definetly a legit complaint by Trump. Moreover, the extent of literal election day fraud is well known in Chicago, Philadelphia and many other democratically controlled cities. Naturally, the historical cheating doesn't guarantee current year cheating would occur, but it is natural assumption to make. Demo's don't have a good track record in voting honesty. IE. President Kennedy in 1960 versus the actual winner Richard Nixon.