Posted on 08/08/2016 10:03:19 AM PDT by xzins
APL is a macro-language used to fine-tune insurance models. If used properly, very powerful. Insurance companies used it for years with high accuracy to compute profitability of their products, based on social, economic and other assumptions.
The point is that they are telling us how they do what they do. Better than the others. How do we know the others aren’t pulling their responses from NY, CA, NJ, MA, etc.?
https://www.longroom.com/polls/
These people have a lot of experience, history and transparency.
They expose the disinformation and bias in the other polls.
TWB
We heard this same thing about Romney. The polls were bias against him by oversampling Dems. And we know how that turned out.
Trump, and his supporters, need to realistically access his position and adapt accordingly, and not pretend that all is well, as it is not.
I don’t want to give the impression that the election is not based on electoral college votes.
However, there is a historic correlation between a national sample and the state samples.
What this poll suggests is that Georgia, Texas, South Carolina, etc. are simply not going for Clinton.
It suggests that Trump has an excellent chance in all the battleground states.
Are you surprised? Of course he is.
I did not know that. Thanks!
These people accurately forecast the Obama/Romney race. They accounted for bias from either side.
That said, polling firms want to look more accurate the closer to an election that they get UNLESS they are getting big bucks to be manipulative. They sell their polling expertise to most firms that use survey data.
However, many of them will use manipulative polls during a campaign because of their own bias or that of their clients.
THIS is the time when manipulative polls are really significant political tools to shape opinion and energy.
Apply the biases to state polls. In the end. If you take bush 2004, swap NH for NM and PA for VA, you would have a win. Even losing CO.
The difference. The liberal media pollsters were not afraid of Romney or McCain. They certainly are much afraid of Trump and no doubt they will use incorrect models and polling to reflect their favoritism for Shillary. You can see their fright manifesting everywhere.
Thanks for the reply. By “these people”, do you mean this poll accurately forecast the last election? How does their methodology differ from “major polls”?
The predicted turnout was D+2. It was D+6. That is where things went wrong. It will be between Even and D+2 this time. Our side will will. Gut feeling: Anti Mormon conservatives stayed home. Presbyterians don’t have that kind of bias against them.
The predicted turnout was D+2. It was D+6. That is where things went wrong. It will be between Even and D+2 this time. Our side will win.. Gut feeling: Anti Mormon conservatives stayed home. Presbyterians don’t have that kind of bias against them.
So why does Yougov on this site have Trump ahead by 2.4% on August 1st, but on RCP, Yougov has Hillary ahead by 3% on August 2nd?
And why does Ipsos/Reuters on this site have Trump up by 3.8% on August 4th, but on RCP it has Hillary up by 4% on the same day?
Then CNN on this site has Hillary up by 1.9% on July 31st, but on RCP has her up by 9% on August 1st.
I stopped after this. It’s completely off. I’ll stick with RCP.
We have developed our analytical model using the programming language that we and other actuaries have used for the last 30 yearsYou've never heard of actuaries?
Never heard of them before this post.
This is just more “unskewing” polls. It is the same nonsense people fell for in 2012.
You can’t “unskew” a poll like this site and others in the past attempted to do.
A poll is a poll, accept it or reject it. Trying to apply your own special sauce to it does not work.
The best bet is to simply go by the RCP average. It isn’t perfect, but it averages all the polls that are regarded as scientific. Last time people complained that the D+ sample was too large, but the polling turned out to be fairly accurate - especially on the state level.
This is relevant because it's producing somewhat different results from state-run media polling. And also because the national popular vote is usually the same as the electoral result. It's an interesting indicator, three months out.
This is relevant because it's producing somewhat different results from state-run media polling. And also because the national popular vote is usually the same as the electoral result. It's an interesting indicator, three months out.
Using CNN, your last example, if you look at the lower data on this site, you’ll see that they have CNN bias recorded at 7.1%.
That would be claiming that CNN had decreased their Hillary bias and then increased it again.
It would seem to be, in my opinion, looking at the chart, an attempt to show a Trump convention bounce and then to take it away again.
I’ve never heard of “Long Room.” Their internet presence just popped up.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.