Do you belive the polls? - Alpha Centauri Forums

  Alpha Centauri Forums
  Non-SMAC related
  Do you belive the polls?

profile | register | prefs | faq | search

Author Topic:   Do you belive the polls?

OhWell posted 05-21-99 09:28 AM ET
Does anyone really believe the poll or survey results that the talking heads on American TV are so fond of quoting? I don't. Here is why:
1) There are over 270,000,000 people in the USA. A typical survey may talk to 1000 people. So they talked to about .00037% of the population. IMNSHO a .00037% sample is statistically insignificant. To top that off they claim accuracy of +- 3%!
2) There is no way to validate the assertions of a poll. It is very easy to make claims that no one can refute.
3) The question and sequence of questions asked can have a profound effect on the answers given. Kind of like quantum physics; the act of observing an event can have effect on the event.
A typical survey call might go something like this:
Phone: ring, ring
Citizen: "Uh, howdy there?"
Pollster: "Good afternoon. I'm conducting an opinion poll on the performance of Bill Clinton. May I have a moment of your time?"
Citizen: "Bill who? I don't know any Bill."
Pollster: "You know, in the White House."
Citizen: "No, Bobby Jo McCoy lives in the white house down the road."
Pollster: "No, the White House in Washington!"
Citizen: "President Washington?"
Pollster: "Yea, the President. Now, how do you feel about his admission that he had inappropriate sex?"
Citizen: "Was he messing with the livestock again?"
Pollster: "No it was an Intern."
Citizen: "Well taken turns is good. I always make my kids take turns 'cause I got 14 of 'em."
Pollster: "So you approve then?"
Citizen: "Uhh, yea."
Pollster: "Now, how do you feel about his performance?"
Citizen: "Uhh, is he an actor?"
Pollster: "No, that was two Presidents back."
Citizen: "So what does he do then?"
Pollster: "He runs the country!"
Citizen: "Uhh, which country?"
Pollster: "The USA!"
Citizen: "Oh, I love the US of A!"
Pollster: "So you approve then?"
Citizen: "Sure!"
Pollster: "Thank you for your time. Have a good day."
What do you think? The survey says....

Cameleon posted 05-21-99 09:36 AM ET
There is a great quote about lies and statistics that would be very fitting, but I can't remember what it was....Can anyone enlighten me?
Cameleon
-who cannot see anymore HTML with two threads gone to waste-
GaryD posted 05-21-99 09:38 AM ET
Something along the lines of
"there are lies, damned lies, and statistics"
Anyway stats is a black art. you shouldn't go dabbling.
Rex Little posted 05-21-99 11:12 AM ET
The phrasing of the questions are indeed crucial. One survey will ask, "Do you favor keeping guns out of the hands of small children and mental patients?" and conclude that 80% of the people favor gun control; another will ask, "Do you think patriotic Americans should be armed against a communist invasion?" and report that 80% are against gun control.
And no matter how the questions are phrased, I'd ignore any survey that's based on asking people about their sex lives. People will lie about that just to make themselves feel better, even if they believe their answers are confidential.
JohnIII posted 05-21-99 01:23 PM ET
I mess with stats so I look good. It's the only way...
John III
Spoe posted 05-21-99 01:39 PM ET
Yeah, in the polls I'd be worried about how the questions are asked and not the sample size.
To close this thread, click here (moderator or admin only).


Hop to:

Contact Us | Alpha Centauri Home

Powered by: Ultimate Bulletin Board, Version 5.18
© Madrona Park, Inc., 1998.

Author	Topic: Do you belive the polls?
OhWell	posted 05-21-99 09:28 AM ET Does anyone really believe the poll or survey results that the talking heads on American TV are so fond of quoting? I don't. Here is why: 1) There are over 270,000,000 people in the USA. A typical survey may talk to 1000 people. So they talked to about .00037% of the population. IMNSHO a .00037% sample is statistically insignificant. To top that off they claim accuracy of +- 3%! 2) There is no way to validate the assertions of a poll. It is very easy to make claims that no one can refute. 3) The question and sequence of questions asked can have a profound effect on the answers given. Kind of like quantum physics; the act of observing an event can have effect on the event. A typical survey call might go something like this: Phone: ring, ring Citizen: "Uh, howdy there?" Pollster: "Good afternoon. I'm conducting an opinion poll on the performance of Bill Clinton. May I have a moment of your time?" Citizen: "Bill who? I don't know any Bill." Pollster: "You know, in the White House." Citizen: "No, Bobby Jo McCoy lives in the white house down the road." Pollster: "No, the White House in Washington!" Citizen: "President Washington?" Pollster: "Yea, the President. Now, how do you feel about his admission that he had inappropriate sex?" Citizen: "Was he messing with the livestock again?" Pollster: "No it was an Intern." Citizen: "Well taken turns is good. I always make my kids take turns 'cause I got 14 of 'em." Pollster: "So you approve then?" Citizen: "Uhh, yea." Pollster: "Now, how do you feel about his performance?" Citizen: "Uhh, is he an actor?" Pollster: "No, that was two Presidents back." Citizen: "So what does he do then?" Pollster: "He runs the country!" Citizen: "Uhh, which country?" Pollster: "The USA!" Citizen: "Oh, I love the US of A!" Pollster: "So you approve then?" Citizen: "Sure!" Pollster: "Thank you for your time. Have a good day." What do you think? The survey says....
Cameleon	posted 05-21-99 09:36 AM ET There is a great quote about lies and statistics that would be very fitting, but I can't remember what it was....Can anyone enlighten me? Cameleon -who cannot see anymore HTML with two threads gone to waste-
GaryD	posted 05-21-99 09:38 AM ET Something along the lines of "there are lies, damned lies, and statistics" Anyway stats is a black art. you shouldn't go dabbling.
Rex Little	posted 05-21-99 11:12 AM ET The phrasing of the questions are indeed crucial. One survey will ask, "Do you favor keeping guns out of the hands of small children and mental patients?" and conclude that 80% of the people favor gun control; another will ask, "Do you think patriotic Americans should be armed against a communist invasion?" and report that 80% are against gun control. And no matter how the questions are phrased, I'd ignore any survey that's based on asking people about their sex lives. People will lie about that just to make themselves feel better, even if they believe their answers are confidential.
JohnIII	posted 05-21-99 01:23 PM ET I mess with stats so I look good. It's the only way... John III
Spoe	posted 05-21-99 01:39 PM ET Yeah, in the polls I'd be worried about how the questions are asked and not the sample size.