4 ms·

Why would someone reject the fundamentals of applied statistics?

dominotw 9y ago
Because inputs are not tightly controlled. we don't know who exactly is answering these polls and we have no idea if preferences of poll answering population can be extrapolated to "republicans". There is nothing "fundamental" about these polls, throwing a web widget in front of a self selected group of audience is not scientific or fundamental. Why do you continue to assert that these polls are "fundamental" when they are proven to be to be totally bogus time after time.
- ZeroGravitas 9y ago
  "The CBS News 2017 Alabama poll is conducted by YouGov using a representative sample of 1,067 registered voters in Alabama between November 28–December 1, 2017. The margin of error based upon the entire sample is 3.8 percent and for likely voters is 4.8 percent." This is a poll done by a standard polling company. You seem to be mischaracterising how they are done and, two comments up the chain, you attack even professional polls based on their performance in the last presidential election, where they were very accurate in predicting the popular vote.
  - dominotw 9y ago
    >This is a poll done by a standard polling company. Ok thats fair. I might have dismissed them without good reason. I have a blocker for all news sites so i can't access them without going through like 20 steps. >where they were very accurate in predicting the popular vote. My comment was about their predictions about number of seats each candidate will win. What does that have to do with "predicting the popular vote." .
    - ZeroGravitas 9y ago
      The "polls" did not predict the electoral college, they ask people how they are going to vote and how likely the are to vote i.e. the popular vote. Other predictions are a seperate step done afterwards using the polls as input, generally not by polling companies. And at least one such prediction by 538 had it as a 1/3 chance of Trump winning, which considering how narrow the eventual gap was, is a pretty darn accurate prediction. A couple of other people did say 99%, but they were clearly wrong even before the result came in and were called out by, for example, 538.
    - heartbreak 9y ago
      > My comment was about their predictions about number of seats each candidate will win. What does that have to do with "predicting the popular vote." . Why are you using “seats” to refer to Electoral College delegates?
    - bigchewy 9y ago
      >> I have a blocker for all news sites so i can't access them without going through like 20 steps. Then consider no longer commenting on news related articles, sources etc that you don't know anything about. Wild speculation isn't helpful.
      - dominotw 9y ago
        ok. my bad.

Because inputs are not tightly controlled. we don't know who exactly is answering these polls and we have no idea if preferences of poll answering population can be extrapolated to "republicans". There is nothing "fundamental" about these polls, throwing a web widget in front of a self selected group of audience is not scientific or fundamental. Why do you continue to assert that these polls are "fundamental" when they are proven to be to be totally bogus time after time.

ZeroGravitas 9y ago
"The CBS News 2017 Alabama poll is conducted by YouGov using a representative sample of 1,067 registered voters in Alabama between November 28–December 1, 2017. The margin of error based upon the entire sample is 3.8 percent and for likely voters is 4.8 percent." This is a poll done by a standard polling company. You seem to be mischaracterising how they are done and, two comments up the chain, you attack even professional polls based on their performance in the last presidential election, where they were very accurate in predicting the popular vote.
- dominotw 9y ago
  >This is a poll done by a standard polling company. Ok thats fair. I might have dismissed them without good reason. I have a blocker for all news sites so i can't access them without going through like 20 steps. >where they were very accurate in predicting the popular vote. My comment was about their predictions about number of seats each candidate will win. What does that have to do with "predicting the popular vote." .
  - ZeroGravitas 9y ago
    The "polls" did not predict the electoral college, they ask people how they are going to vote and how likely the are to vote i.e. the popular vote. Other predictions are a seperate step done afterwards using the polls as input, generally not by polling companies. And at least one such prediction by 538 had it as a 1/3 chance of Trump winning, which considering how narrow the eventual gap was, is a pretty darn accurate prediction. A couple of other people did say 99%, but they were clearly wrong even before the result came in and were called out by, for example, 538.
  - heartbreak 9y ago
    > My comment was about their predictions about number of seats each candidate will win. What does that have to do with "predicting the popular vote." . Why are you using “seats” to refer to Electoral College delegates?
  - bigchewy 9y ago
    >> I have a blocker for all news sites so i can't access them without going through like 20 steps. Then consider no longer commenting on news related articles, sources etc that you don't know anything about. Wild speculation isn't helpful.
    - dominotw 9y ago
      ok. my bad.