When will we get pollsters who ask information-seeking questions about the answers they get, and discount the "opinion" of non-answerers?
"Vice President Kamala Harris' approval rating is 28% – even worse than Biden's. The poll shows that 51% disapprove of the job she's doing. One in 5, 21%, are undecided."
When will we get pollsters who ask information-seeking questions about the answers they get, and discount the "opinion" of non-answerers?
"Vice President Kamala Harris' approval rating is 28% – even worse than Biden's. The poll shows that 51% disapprove of the job she's doing. One in 5, 21%, are undecided."
It would be nice to know what action or absence of specific action forms the basis for the disapproval vote and why the undecided are undecided. Might it be because they know nothing about what VP Harris has done or not done but don't want to admit their ignorance?
The primary purpose of these polls is not to provide useful information... at least not in their presentation to the public. Their purpose is to act as fodder for particular narratives. I have serious doubts about the usefulness of many of these polls that are so often touted for various political purposes or for political horserace commentary.
A long time ago (around 25-30 years ago) when I was a grad student I did a lot of work with polling and data analysis in social science (a lot of political polling, actually). HOW you ask and word the question is just as important as the content of the question (actually more). I used to like to joke saying that I could construct a poll that would make Hitler or Stalin look like fantastic people and leaders--except it isn't really a joke.
A carefully constructed and validated poll can tell you lot of interesting things. Political polls are not those types of polls, in most cases--at least not as they are presented to the public... and used by the media. As an aside, we used to get data from exit polling--this was basically the raw data (on which we did our own analysis)--we are talking about computer printouts (yes that was how long ago it was) that was several inches thick and very detailed.
Progressives like to throw out that X% (forget the actual number) support progressive environmental policy--well, that is both true and not true. Here is an example (using example questions to make the point):
1) The environment is important and we need to act to preserve it (this will be very popular);
2) The environment is important and we need to tax rich people so that we can preserve it (not as popular but still popular);
3) The environment is important and we need to increase everyone's taxes so that we can preserve it (probably not very popular); and
4) The environment is important and we need to increase gasoline tax to $6 a gallon so that we can preserve it (not going to be popular at all).
Do you see what I did there? General ideas and aspirational goals are popular (they are also generally open to interpretation). Specific policies are usually unpopular--the more specific 9and higher obvious cost) the less popular.
Everybody wants to save the planet until they find out how much it will cost THEM.
I can kill your idea or policy or popularity in a poll by how I ask the question, when I ask the question (something as simple as time of day) and who I ask. It becomes a question of rigor in construction and performance.
When will we get pollsters who ask information-seeking questions about the answers they get, and discount the "opinion" of non-answerers?
"Vice President Kamala Harris' approval rating is 28% – even worse than Biden's. The poll shows that 51% disapprove of the job she's doing. One in 5, 21%, are undecided."
It would be nice to know what action or absence of specific action forms the basis for the disapproval vote and why the undecided are undecided. Might it be because they know nothing about what VP Harris has done or not done but don't want to admit their ignorance?
The primary purpose of these polls is not to provide useful information... at least not in their presentation to the public. Their purpose is to act as fodder for particular narratives. I have serious doubts about the usefulness of many of these polls that are so often touted for various political purposes or for political horserace commentary.
A long time ago (around 25-30 years ago) when I was a grad student I did a lot of work with polling and data analysis in social science (a lot of political polling, actually). HOW you ask and word the question is just as important as the content of the question (actually more). I used to like to joke saying that I could construct a poll that would make Hitler or Stalin look like fantastic people and leaders--except it isn't really a joke.
A carefully constructed and validated poll can tell you lot of interesting things. Political polls are not those types of polls, in most cases--at least not as they are presented to the public... and used by the media. As an aside, we used to get data from exit polling--this was basically the raw data (on which we did our own analysis)--we are talking about computer printouts (yes that was how long ago it was) that was several inches thick and very detailed.
Progressives like to throw out that X% (forget the actual number) support progressive environmental policy--well, that is both true and not true. Here is an example (using example questions to make the point):
1) The environment is important and we need to act to preserve it (this will be very popular);
2) The environment is important and we need to tax rich people so that we can preserve it (not as popular but still popular);
3) The environment is important and we need to increase everyone's taxes so that we can preserve it (probably not very popular); and
4) The environment is important and we need to increase gasoline tax to $6 a gallon so that we can preserve it (not going to be popular at all).
Do you see what I did there? General ideas and aspirational goals are popular (they are also generally open to interpretation). Specific policies are usually unpopular--the more specific 9and higher obvious cost) the less popular.
Everybody wants to save the planet until they find out how much it will cost THEM.
I can kill your idea or policy or popularity in a poll by how I ask the question, when I ask the question (something as simple as time of day) and who I ask. It becomes a question of rigor in construction and performance.