Do political polls accurately represent the views of a population?
It depends on how the question is phrased, you can get almost any answer you want if you ask it the way you want it to be answered.
It depends on how the question is phrased, you can get almost any answer you want if you ask it the way you want it to be answered.
April 3rd, 2010 at 10:18 pm
with in the margin of error yes
References :
April 3rd, 2010 at 11:08 pm
no, it represents the views of the small fraction of the population asked.
References :
April 3rd, 2010 at 11:24 pm
It depends on how the question is phrased, you can get almost any answer you want if you ask it the way you want it to be answered.
References :
April 3rd, 2010 at 11:58 pm
Kind of, but with any statistical methods there is a standard deviation from the mean. This is the plus or minus interval which says that the true value lies within the interval with a certain level of confidence. Usually it’s calculated at 95-99% confidence intervals. It’s usually best to take an average of the liberal and conservative polls and use propagation of error by partial derivatives method to find the error of the average of all polls.
References :
April 4th, 2010 at 12:43 am
I’d think yes if it was done randomly and across a wide area. No if it’s done from a control group of any sorts.
References :
April 4th, 2010 at 12:58 am
Sometimes they are wrong ie picking Gore to beat Bush but most of the time they seem accurate
References :
April 4th, 2010 at 1:26 am
Well, Rasmussen does- it polls actual voters.
I’d say the vast majority of polls hand-picks who they want to answer, though, in order to skew the results to be whatever "point" they want to prove.
References :
April 4th, 2010 at 2:02 am
WHICH polls?
IF they use sound methodology, and ask questions in ways that don’t bias answers, yes.
If they get non-random samples (for instance, self-selected samples from a skewed sub-group of the population), use wacky computing methods (Rassmusen, for instance, doesn’t give approval ratings, it takes the STRONG approval, and subtracts the STRONG disapproval percents and reports that, which skews toward the negative), or ask slanted questions (again Rasmussen is famous for this), then, no.
BTW Fox "News" uses all of the above in its polls.
Gallup, the group started by the guy who INVENTED the accurate public opinion poll is consistently reliable. Another good thing about Gallup is that they give full information about their polling, including WHEN they asked, how many responces they got, the exact wording of the questions asked, the margin of error, and level of confidence. Their website even explains why polling works.
If you want to see what Americans are thinking, go to their website.
Edited after reading other answers: More people voted for Gore in 2000, so the polls predicting his win were accurate. The reason he lost the election was that Florida election officials (headed by a Bush campaign manager, Kathleen Harris) had removed registered Dems from the lists, so they were prevented from voting.
That’s why Bush got Florida’s Electoral College votes.
As for "take the polls from the right and left, and conclude the true result is in the middle" — no. Use pollsters that don’t represent any side.
It’s simply a lie that everyone but Rasmussen pre-selects their sample to prove their point.
References :