The following was posted in Chilesoc to purpose of a discussion on polls that appeared slightly in the press (the same of always criticizing CERC for use fees; CERC criticizing Adimark for the questions used). And as you would be to leave this poor blog as a repository of all things, of greater or lesser importance, that it occurs to me to write publicly, I proceed to copy it. Here is the text (published in Chilesoc 30 October)
Hmmm, I have the impression that these controversial and I come from listening to such a long time ago. It is a matter more ritual than anything else at this point, I think. In order to.
Yes, it is true that the fees do not allow to measure with the formula accuracy. It is also true that it is irrelevant in reality. The political polls have the advantage of being able to actually be compared with reality (elections) , and the samples fees -while not the height of accuracy, and neither have said things directly idiots or wrong. If this is so, then the methodological issue not real is that the survey-fees are bad but to explain sociologically why surveys that do not follow methodology random do, however, regularly good results. That tells us something about how society works, that should be the main thing that determines good methodologies.
In any case, that resolves a difference of methodologies whose results can be contrasted it is precisely this contrast, is not determined by certain formulas, which, in turn, have certain assumptions about how society works. If these formulas say that the fee is not served, but in reality they serve-and in my humble opinion, at least you can say that useless are not-then the problem is in the formulas.
That Gallup came up with a way to ask 50 years ago that others have used doesn’t say anything. In itself, the fact of presenting 3 alternatives is not distortionary. Is another way, different, question. And obviously should produce different results. In the end, in the ‘reality’ (in the opinions that effectively kept the people on the presidency) 2 or 3 alternatives, it is also distortive: the real reviews are much more varied. If we want to reduce this complexity to a limited set of alternatives, well, that is our prerogative and depends on what we want to know. But, in any case, there is no way ‘single and correct’ to ask questions.
And the idea that the label of independent not there is because it does not appear on the ballot of election does not work too. The electoral options and the opinions of the individuals are not identical in their structure or in their divisions. For some cases we intereserá ask for electoral options (say, if we are talking of an election) and not in others we intereserá (because we are not talking of elections).
BTW, the theme that the Alliance has the 50% because it takes that in elections and you can’t believe in a 14% not applicable. The universe that is taken out, the 50% -of the emitted valid votes – is not the universe from which a roll of 14% -which is the universe of the survey, which is basically everyone over a certain age. The second is broader and includes those who have not enrolled, enrollees who do not vote, those who vote and give votes invalid. I would not be surprised that the same number all that is 50% of the universe election was not so far to 14% of the total universe of adults.
Juan Jiménez A.