A little more than half of Americans favor allowing parents to send their children to any public, private or parochial school, with the government paying at least part of the tuition.

And the voucher idea is gaining, according to the people who run the Phi Delta Kappa/Gallup Poll survey. In 1996, the first year the question was on the survey, 54 percent opposed the idea; last year opinion was almost evenly divided.

The poll findings mean . . .

Well, according to a fascinating analysis by the Public Agenda, they don't mean diddly. Why? Because, when it comes to vouchers, charter schools and other choice initiatives, respondents in the national polls don't know what they're talking about.

"On Thin Ice," the Public Agenda's report, doesn't put it quite that bluntly, of course. Deborah Wadsworth, executive director of the New York-based public opinion and policy-analysis group, says the problem is that the "choice" debate has become a captive of the experts on both sides.

"But while leadership debate on these issues is thriving, most citizens have only the vaguest notion what terms like `voucher' and `charter' mean, much less how these ideas might affect their own lives. For most people, these issues are not much more than words in a newspaper headline. `Oh, yeah . . . I saw something about that.' "

"On Thin Ice" reports the startling discovery that even in places where "choice" options have been initiated -- Milwaukee and Cleveland, for instance -- "parents are virtually as unfamiliar with vouchers as everyone else." The result, says Wadsworth, is that most public opinion surveys on the subject are essentially worthless when it comes to informing public policy.

The new report is, at bottom, an indictment of two institutions that ought to know better. The first is the media. People don't know much about the specifics of "choice" because we don't tell them -- at least not with sufficient frequency, detail or clarity. We tend to report on the expert fights -- the mutual name-calling, the oversimplifications, politicizations and, yes, the poll results. We seem more interested in who's winning than in the educational implications of victory.

The second failure is the pollsters' -- not their inaccuracies but the dirty little secret that their questions often lead respondents to answer yes or no without establishing that they understand the question. We often respond to the sound of the question as much as to its content. Even changing questions from "public funding" to "government funding" creates a difference in poll results on vouchers, according to nonpartisan Public Agenda.

"It may seem odd that an organization that conducts public opinion research -- and promotes it as an important component in policy-making -- would advise leaders to step back, take a deep breath and view survey results with a skeptical eye," says Wadsworth.

"That's because not all survey situations are alike. Surveys are most useful when they plumb public attitudes in areas that people have worried about, chewed over in discussions with family and friends or at least observed from a distance for a reasonable length of time. Surveys tend to be mushy when they get into areas that people haven't focused on and may not understand clearly."

As for its own survey, Public Agenda makes only the most tentative of claims: People tend to view vouchers as at least a partial solution when the idea is explained to them; they have no clue as to whether the competition from vouchers will spur public schools to improve, and, though they worry about the prospect of dismantling the public school system, they are not at all happy with the present state of public education.

But unhappiness with the status quo is not necessarily a vote for any particular alternative. Polls can be helpful, but officials are skating on thin ice if they expect polls to tell them how Americans want their school systems organized, run and financed.

Most just don't know.