Katha:
Actually polls are often used in comparative research that way. They might be quite wrong when we try to use them to estimate the absolute lovel of support for a particuler cause, but they may give an idea about relative diference among different groups, on the assumption that errors are randomly distributed and thus cancel each other out (i.e. are same, on average, for all groups being compared).
To illustrate it with aner example, using polls to determine the total sum of money donated to various causes is a moot proposition, but we can use polls to determine how different causes rank in the support they receive. That is, if we ask a respondents how much money they gave altogether, the sum they report is likely to be off the mark. The same happens if we ask how much they gave to cause A and to cause B - but since what Rs reported about the cause A and the cause B is likely to be off the mark by the same amount (error), we can assume that the difference between what Rs reported for both causes reflect the actual diffrence. This can be demonstrated as follows. Suppose that the total amound reported by all Rs to cause A is $5,000 and to cause B - $8,000. Then:
5,000 + E + 8,000 + E = T
where E is the unknown quantity representing error in reporting, an T is the unknown quantity representing the amound of total giving. It is easy to see that the equation above cannot be solved - i.e. the value of T cannot be found - becaue there are too many unknowns. We can, however, calculate the relative difference between giving to A and giving to B as 5,000/8,000, or saying that the cause A receives only 63% support the cause B gets.
The same holds for the acceptablity of abortion. We do not know what, say, a 5 means in terms of support for abortion, but we can say that a group that averages 5 is more supportive of abortion than a group that averaged 4.
Of course, there is a way of 'calibrating' those measures by comparing them to the level of R's approval of other items pertaining to sexual morality that were included in the survey. For example, we can compare the R's responses to the abortion item to her reponse to the items on acceptablity of divorce, extramarital sex, and underage sex, and then express her opinion on abortion in terms of the difference from her support of the other three items. I might do that in th efuture, but this is a time consuming analysis and unfortunately I also have to do some work in this research factory :).
>ps. As for talking about abortion in ways that appeal to working class
>people -- why do you assume pro-choicers and feminists don't try to do
>this? I assure you, the political campaigns and ad campaigns of the big
>pro-choice organizations are rigorously focussed grouped.(that's why,
>for instance, they fight anti-choice measures by talking about govt
>intrusion on privacy rather than by talking about women's rights). And
>believe it or not, abortion activists and grass roots feminists, people
>like suzanne Pharr for example, spend a lot of time on this issue. some
>of them are even working class themselves!
No doubt. My point is, however, that the US Left has the tendency to either (i) speak in incomprehensible jargon, inside-speak etc. that is alien to most outsiders, (ii) talk to the academic-intellectual community and in their languare, or (iii) both. Hence my suggestion to be more attentive to what common folks are saying and use the language that is on _their_ wavelength to get our message across. There is an old Sufi proverb "A fool tries to convince me with his arguments, a sage - with my own." It is about tactics rather than principles.
Regards,
Wojtek Sokolowski