nimh wrote:Huh? Those differences are even all well within the margin of error, so how does that "point to" anything?
With sampling errors of three to five percentage points, those differences could vary widely, or not vary at all. In fact, the last poll conducted among those which you cited and linked, in June, could have been completely meaningless based on the margin of error. Which is why i say it points to the unreliability of polls. If the possible sampling error is factored in, either your point or mine could be supported (the 5% difference in "self-identification" of partisan affiliation could in fact be a lot wider, making your point, or it could vanish altogether, making my point). All of which shows how unreliable such polls can be. It is useful to keep in mind, as well, that polls rarely ask people exactly the same questions, from one poll to another, or from one sample to another over time.
Quote:No poll claims that it has the result precise to the individual percentage; they tell you that each of the (candidate's, party's, etc) numbers is submissive [<not the right word] to the margin of error, and that even then the results are reliable just 95% or 90% (etc) of the time.
Which is a fancy statistical way of admitting that the methodology is sufficiently imprecise for the margin of error to be meaningless.
Quote:Ive never quite gotten the bipolar response to opinion polls... Either people take each individual number for gospel truth and gloat because their candidate is shown 1% or 2% up over the other, or, if things turn out to be nothing quite that exact, bounce to the other extreme and say that polls are just good for nothing.
That would not be me, however, among those "people" to whom you refer. I have consistently been skeptical of poll results, and have said as much in threads by people from either side of the partisan divide. Most recently i took on the poll results which Zippo referred to when he alleged that the Shrub had lower approval ratings than Nixon. He compared the Shrub's approval rating on the handling of the war to Nixon's overall approval rating in the early summer of 1974, just before he resigned. I pointed that out, and he got predictably huffy with me about it. A comparison of the overall approval ratings of the Shrub and Nixon actually showed the Shrub slightly ahead of Tricky Dick, and i cited and linked the relevant polls.
Quote:Whereas they're pretty good in doing what they only ever claim to do: pointing out general trends and getting reliable numbers within the range of error margins and the potential for outliers.
It is precisely because they are indicative of general trends rather than being able to infallibly pinpoint public opinion that i objected to Cyclo's claim about the relative poll results with regard to Cheney and Clinton. Quite apart from his linked sources not supporting his claim, taking into consideration the margins of error, and that fact that it were possible that the error made Clinton look more popular than he was and Cheney less popular than he is, it were entirely possible that in the poll which found a 45% approval for Clinton's impeachment, and this which finds a 54% approval for Cheney's impeachment could actually be as little as 3 percentage points apart, or even less, depending upon the size of the samples. Therefore, i consider it a suspect basis upon which to make allegations about the popularity of an "impeach Cheney" movement.
Quote:If you dont expect more of them than is realistic; focus on the average or trend across several different polls or pollsters, keep in mind that they have margins of error, house effects, and the occasional outlier, and cant predict future developments, they're pretty useful - they're still the best way we have available to gauge collective/public opinion, which is why both politicians and businesses make such use of them.
They don't use them to the extent that polling organization would like the public to think. When either commercial operations or politicians want to make crucial decisions about their respective advertising campaigns, they don't rely on a survey by telephone with a handful of questions, the biases of which could skew the results--they use focus groups which spend hours refining just how people see the product (and with politicians the product is the man or woman herself) in order to fine-tune their marketing strategies. And, they continually go back to the focus groups to eliminate errors and mistaken impressions. Polls are most useful to news media who want to find a hook to hang a story on.
Quote:Well, thats your opinion.. though I dont know what it is based on exactly.
I disagree, obviously. Take, for example, the opinion polls for the 2004 Presidential elections. The average
of all the last polls out, in the very last days before the election, had Bush winning with 48,7%, Kerry trailing with 47,9%, and Nader at 0,9%. Not bad, compared to the actual result.
Compare those results to the same types of polls conducted many months before an election. In the last days before an election, a large proportion of the voters will have made their minds up--even if all potential voters have not done so, far more will have made up their minds than would have been the case during the primary season. Look at the "impeach Clinton" numbers in the poll summary which Cyclo linked. They fluctuate between about 25% and 38%, until it reached the point that the House was actually considering the indictment, the impeachment. At that point, approval of the process shot up. It might, of course, have plummeted, too--but the point is that people in the days just before an election are more likely to have made up their minds, and similarly, people who know the Judiciary Committee is writing a bill of impeachment will also be moved to make up their minds. Right now, the Cheney data is unlikely to be as conclusive as it would be if the Judiciary Committee actually began to write a bill of impeachment.
Quote:True. But the general point he was making - that there was significant public support for impeachment, a greater support than you'd think or pundits make out, with even an actual majority sympathetic to an impeachment of Cheney - is well taken even if the "double than" specification was incorrect.
Of course, you don't know what i'd think of the matter, because you can't read my mind, and i've not stated my opinion on the matter. My remarks were only concerned with the issue of the comparison of support for the impeachment of Cheney and the support for the impeachment of Clinton. What the "pundits" think on this, or any other issue, is a matter of indifference to me.
I don't consider comparisons in such complex matters to be valid. Nixon was caught with his hand in the cookie jar, and despite having high poll ratings in January, 1974, because the public were convinced that he was getting us out of Vietnam, the relentless work of the Senate Watergate Committee torpedoed his political career in the course of the later winter and spring of that year. By contrast, the Shrub's low approval ratings depend on dissatisfaction with the conduct of the war, and it is by no means certain that a significant number of the electorate consider him corrupt and criminal. So i don't put much credence in comparisons of Nixon and the Shrub, or Cheney for that matter. Similarly, the situation which Clinton faced in 1998 and the one which the Shrub and Cheney face today are sufficiently dissimilar that i don't consider comparisons in those two cases to be valid either. Even the impeachment of Andrew Johnson doesn't compare realistically with this situation, as a Republican controlled Congress was out to get him, and specifically passed the Tenure of Office Act to prohibit him from firing Stanton, which was the action he took which precipitated the bill of impeachment against him.
The situation as it now stands is very much like that which Nixon faced before the Watergate committee sat--no evidence against the Shrub or Cheney for any form of malfeasance, just a general suspicion. It would require pretty solid evidence that one or both knowingly and willfully lied about the justifications for this war, or that one or both knowingly and willingly conspired to "out" Miss Plame in order to punish Mr. Wilson, to establish a realistic basis for an impeachment. As it stands right now, there is no historical precedent solid enough to justify a prediction about whether either Uncle Dick or the Shrub will be impeached.
As i remarked to Cyclo, we all have to wait to see how this falls out--i think any predictions about impeachment are premature, and i don't see the poll data Cyclo provided as being good evidence to make the case he would so dearly like to make.
This is definitely not a case of me complaining about polls who don't say what i would like to hear, and praising polls which do. I could think of nothing i would like better than to see Uncle Dick go the same way as Spiro Agnew--i just don't see any convincing evidence that it is likely based on the poll data to which Cyclo referred.