I would not depend on surveys measuring attitudes on an issue or two whose "ownership" is claimed by this or that political faction. There are numerous problems with such surveys - chief among them being that different people define issues in their own personal way regardless of the claimed 'political ownership." Thus, a person may support the right to abortion, but reject the feminist interpretation of it or even oppose the idea of women's rights to choose altogether.
For that reason, I would rather depend on the behavior as an indicator of political attitudes. Examples of such behavior include voting, purchasing publications, or interaction in everyday life.
Take voting. Conservative Republicans consistently receive nearly half of the popular votes if the opposing candidate is a centrist Democrat (cf. Bush vs. Gore), but that support skyrockets if the Democrat is even slightly leaning to the left. Such was the case even the supposedly "liberal" 1960s and 1970s cf. McGovern 37.5% vs. Nixon 60.7% in 1972. Another indicator is that solidly conservative or right wing "third parties" command a much higher popular support that solidly progressive or left wing "third parties. Thus Nader attracted 2.7% of the popular vote in 2000 and 0.7% on 1996; Anderson (nor exactly a leftie, to be sure) - 6.6% in 1980. By contrast Perot received 8.4% in 1996, and 18.9% in 1992. Wallace got 13.5% in 1968 - the heyday of the supposedly liberal era. http://www.uselectionatlas.org/USPRESIDENT/frametextj.html
These numbers suggest that regardless of how people answer opinion polls, they prefer conservative candidates to center-left ones, and the far right outnumbers far-left by the factor 5:1 (Nader in the "conservative" 2000 vs. Wallace in the "liberal" 1968).
What is more, the left leaning publications, such as The Nation, Mother Jones, or even New Yorker, are not very popular. If the Us population was so overwhelmingly supportive of the positions "owned" by the Left, one would expect the sales of these publications to be much higher, especially that there not that many titles on the Left.
Another area is interaction in everyday life. People may be publicly supportive (or not dismissive) of affirmative action - but how much inter-racial dating there in the US? Forget the dating - how many times employees of different racial backgrounds go to lunch together? Or take another issue - using public transit. In most urban (outside the NYC or DC where it is often the only viable alternative) people avoid it because it used predominantly by minorities. They also oppose transit line extensions for the very same reason.
The same can be said about schools. What is more, people tend to support ballot measures to fund police but oppose funding for public schools, especially in areas with high numbers of ethnic minorities.
To conclude, I would like to reiterate my argument that the conservative and right wing politics seem to enjoy a much higher popular support than left-leaning politics. I would use the Nixon-McGovern split as a proxy to estimate the popular support of center-left (McGovern) vs a conservative candidate (Nixon) which is 38% vs. 61%. I would use the Nader vs. Wallace split (2.7% vs. 13.5%)as a proxy to measure the popular support of left vs a far right candidate in the US.
PS. Most of the people with whom I came to a direct contact during my 24 or so year stay in the US (longer than in any other country) were fairly centrist, even by the European standards. But then I have worked mainly in government or academic institutions, and lived mainly in Northern CA and Northeast (NYC - DC area). Judging what I see in central PA (my wife's home state) - which is fairly representative of anything outside the SF Bay Area, NYC, Boston, DC areas - it is solidly conservative and often bordering on fascism.
Wojtek