The title question can only deserve a quick response and obvious: Because you need some information, and that kind of information only you can deliver the surveys. Which is valid, but still leaves open the question of the uses. In other words, how much confidence can we have in their results? And given this, what is it that you can do with them?
Now, we know -because we have already had a choice relatively massive – that the surveys have serious problems. Demos for certain all the preventions that they did (or could do) of those who did surveys at that time: That there were electoral polls, which were clearly not asking for a vote of the primary, that primary is not the presidential, that could not be used to predict. However, one could assume that at least should have been in the major trends.
The case is that not what they were. Of the main results only that Piñera would be first, and with distance important, it was something that was already in the surveys. But, were there any survey that would Ossandón would have something less than half of the vote, Piñera? Have there been any survey to tell us that Ossandón would get a vote similar to the whole of the Front Wide? I repeat: I’m Not asking for the surveys to give numerically close results (I have not named any percentage for that matter), only that they were near the main trends.
Primary elections in July, and Surveys
| Candidate | Primary (2 July) | Adimark (June) | Cadem (30 June) | CEP (April-May) |
| Piñera | 47,4 | 57,4 | 45,8 | 69,1 |
| Ossandón | 21,3 | 7,4 | 14,6 | 15,7 |
| Kast | 12,5 | 5,6 | 10,4 | 0,9 |
| Total CHV | 81,2 | 70,4 | 70,8 | 85,7 |
| Sanchez | 12,7 | 24,1 | 22,9 | 14,0 |
| Mayol | 6,1 | 5,6 | 6,3 | 0,3 |
| Total FA | 18,8 | 29,7 | 29,2 | 14,3 |
NOTE: The percentages are calculated around what amounted to five candidates in the primary, not in relation to the general population, to make them comparable.
And if in that case were not even close, why would we think that they are close to now? And again, we are not talking about a topic of points, more or less, but of the big trends.
When we remember how feeble are the methodological bases of a number of surveys. There will be more to remember that not all surveys from which to speak and discuss meet even represent the whole of the territory; that which does not cease to be relevant when the electoral behavior differs between the various territories. We think that the survey CEP -although the most distant of the vote, and therefore with almost no presence of Mayol or Kast (that raise more later on)- is the one that best pointed to the vote of Sanchez. It is precisely the fact of being a national survey allows for that, when you leave spaces where Sanchez was better (more urban for example).
We can follow: we can Also remember that since we are to vote, volunteer models are needed for prediction of vote, and that we do not have-to my knowledge – studies that show that the models designed and applied to regard to effectively work in Chile (I think that only Mori attempted to do this). In the end, in Chile does not use a lot of the surveys out of the ballot box because this tool, in common use in many parts, was a dismal failure when he tried to apply in Chile.
It is true, moreover, that we have no other sources, and that you have to do is use the information you have. What is required is to use them with circumspection. But, at the same time, when used in this way, how much is what we actually provide? The fact that it is clear in all the surveys, and that we can give by the way, is that Piñera is the candidate in the first place, and to distance it from the rest, but do we need polls to know this?
The utility of the surveys, at the end, depends on the answer to that question. I have the impression that for that kind of result simply live in the society; but knowing, at the same time, the ease with which people cheat on these issues, perhaps yes -even that kind of general result already is sufficiently valuable.
FINAL NOTE: it Is true that this entry could have been done quite a few months before, but in the end, life and all that is not always allowed to write when the one I had first thought. It would be more appropriate probably to analyze the latest CEP, would at least be more responsive, but better late than never I guess.