This is an election season in the U.S. and many claims and counter-claims are being made about what "the American people" want. Most of it is fiction. Even polls with percentages this way and that are often unreliable because of any one of several factors.
- Opinions swing from day to day as the debate rages. What a survey showed as the percent who hold opinion X on Monday may not reflect the actual percentage on Wednesday. Etc.
- Many surveys are conducted by people who have an ax to grind. The way you ask the questions; the methods you use can tip a poll one way or the other.
- To speed up the process, there is the temptation to reduce the sample size. Sometimes polls are simply too small to be reliable.
- There is a growing percentage of Americans who vote on election day but refuse to participate in polling prior to the election. When elections are close, this non-response percentage may be the key to the outcome.
All of these factors are particularly true with the candidates or their partisans quote polls. It is not that they tell a lie; it is that they believe lies because they feel so strongly about the issues and/or the outcome. American politics have become so heated that there are very few of the politicians who are capable of actually giving room to the possibility that they might be wrong, even deeply wrong in their values and policies. That possibility, of course, exists whether they want to recognize it or not. And when things are desperate and volatile, it is a much larger factor than any of us would like to admit.
So is there any source of worthwhile survey data about the current attitudes of Americans? Yes, there are academic researchers who are objective enough to provide more reliable data. The best source that I have found at the moment is this article: "Public Opinion, the Deep Recession, and the 2010 Elections" by Robert Blendon and John Benson, Challenge53:5, Sept-Oct 2010, pp 14-33.