Well, I see your mindset in regards to eliminating double voting, but even without double voting, an online poll is still useless. The entire concept of a scientific poll is randomization. If you are trying to get a statistical sample of the people that watched the debate on Monday, you have to have a random sample of people that watched the debate. The only way to get that random sample is for the poll maker to randomly contact people in a systematic way. If you have an online poll, that isn't a sampling of all the people that watched the debate, rather it is a reflection of people that watched the debate that decided to go online to the xyz website and cast a vote. The people that go to xyz website and cast a vote do not represent all people that watched the debate, they are instead a subset of that larger group. A subset of the larger group does not represent the larger group.
the only polls that accurately get the pulse of the American people, are random scientific polls. people have studied this for years and make a lot of money doing these polls. internet polls, based on user interest and can be easily manipulated, are pure bull(*)(*)(*)(*).
I have coursework in statistics to econometrics and ALL POLLS can be manipulated for a desired outcome. Steve
oh, so you think an internet poll that says HIllary has only 10% support nationwide, is legit?? LOL!!!!!
That's NOT WHAT I SAID! I said I have coursework in statistics to econometrics and ALL POLLS can be manipulated for a desired outcome. Steve
yes, but if Qunnipiac committed such fraud, they would be shut down and face prison time. if a bunch of right-wingers gather together to commit fraud on an online poll, there is no punishment.
I suggest you start looking at the tabs and methodology. rather than just the headlines and spin. For example CNN's overnight poll on the debate. Just looking at the headline they say clinton 62% and Trump 27%. But look at the source on the sample weighting and you find the poll consisted of 15% more democrats than Republicans and 8% more democrats than Independents. According to Gallup the country consists of 27% GOP, 31% democrat, and 38% Independents. The race is if yoiu assume hillary pull obamas 2008 numbers. I seriously doubt that she can even get 2000 or 2004 percentages of democrats to the polls, let alone a wave like obama got in 2008 source for polling demographics http://i2.cdn.turner.com/cnn/2016/images/09/27/poll.pdf 26% of the respondents who participated in tonight's survey identified themselves as Republicans, 41% identified themselves as Democrats, and 33% identified themselves as Independent source for political party affiliation http://www.gallup.com/poll/15370/party-affiliation.aspx
I understand how statistics work, lol. I'm just saying, a poll that requires registration (with email verification or captcha) is much harder to manipulate than an online poll that just requires clicking a button. Even kiddie programmers know how to use cURL through proxies to be able to vote as many times as they want.
That's pure BS. You have no clue of what you are talking about. Wait until a few days before the election and all will be giving true results to save face. Steve
months ago, Bernie supporters would openly post on Facebook that they should flood the CNN polls and make Bernie look more popular. that's how worthless these online polls are. hell, last night I voted 10 times in the CNBC poll. any poll that can be soo easily manipulated is (*)(*)(*)(*).
You may understand, but obviously there a lot of people that do not. I understand how you are contending that double voting will really skew results, but my point is that it is useless to begin with, so double voting is only skewing what is already useless.
when an online national poll says Hillary has only 10% support nationwide, you know the poll is bull(*)(*)(*)(*). Trump and Hillary are about evenly split, give or take 5 percentage points. any poll that significantly strays from this, is a very isolated regional poll or is being manipulated.
The LA Times is always an outlying poll but at least they make note that their methodology differs somewhat from other polls.
It's not useless at all depending on the poll. Random sampling doesn't have to only deal with the entire population. You can get very useful information by taking polls within groups and comparing them to previous polls.
And if you look at the average of the polls, Trump has only been ahead on two, brief occasions: immediately after winning Indiana and immediately after the Republican convention. Link
It doesn't really make it much more accurate. It does keep the poll from being brigaded to some degree by making it harder for people to vote multiple times. Alot of these results are merely from a few script kiddies hammering polls like they actually mean something. http://www.dailydot.com/layer8/trump-clinton-debate-online-polls-4chan-the-donald/ Also, freerepublic.com, a well-known bastion of far-right values was trying to "freep" online polls after the debate was over.
You may be able to get useful information by taking polls within groups, but that information is only useful in assessing that group. When you try taking that information about that group, and using it to represent the whole, then that information is useless. For example, if FOX or MSNBC conducts an opt in poll about who won the debate, the sample is only representative of people that watched the debate, AND went to FOX or MSNBC in order to cast a ballot. The group of people that watched the debate AND went to FOX or MSNBC to cast a ballot are not a representative sample of ALL people that watched the debate. In fact they aren't even close to being a representative sample of the whole. That is a very unique type of voter that would bother to go online to cast a vote. For example, people that would go online to cast a vote would inarguably be a younger subset than the whole. When people quote an opt in poll from FOX or MSNBC, they are not quoting an accurate representation of all people that watched the debate. In the same way, if you were to conduct a poll on this website, that would in no way represent the sentiment of the entire population in this country. It would only reflect the sentiment of political junkies that frequent this political chat room specifically.
Accuracy is questionable at best among so-called "scientific polls." Pollsters have a problem making contact with people that screen phone calls for one thing. Another is the honesty of the people being polled, not many will admit they will vote for Trump with the media bashing him. Another is that Trump is bringing new voters into the voting population that never voted or rarely voted before, so they aren't even in most polling databases. Wait until a few days before the election to check the "scientific polls" and you will see their best attempt to tell us what the outcome of the election will be. They want to at least be known for how accurately they picked the winner and a few days before the election is what will be the comparison. Otherwise, it's a waste of time. Steve
Again, the LA Times daily tracker did show Trump over Clinton by around 7 points after the convention. It has also showed a similar bump for Clinton. In fact, that poll showed Clinton inching back up over the last few days - until the debate. Do suspect Trump may have bought the LA Times and USC recently?
a true poll is scientifically randomized. online polls, are no such thing and can be EASILY manipulated to feign the results.
really? how would Qunnipiac manipulate a randomized poll? and if evidence came out that they did this, who would ever hire them or trust them again?
sorry, but if evidence came out that Qunnipiac or any other polling company was manipulating their data so as to alter the results, they would NEVER he hired or trusted again. why would they be?