WASHINGTON -- Six months after President Obama signed landmark legislation that will extend health care coverage to millions of people, Americans still do not really know what the law does.

More than half mistakenly believe the overhaul will raise taxes for most people this year, an Associated Press poll finds. That would hold true only if most people were devoted to indoor tanning, which the law hit with a sales tax.

The uncertainty and confusion amount to a dismal verdict for the Obama administration's campaign to win over public opinion. Before the final votes in Congress, Obama personally assured wavering Democrats he would take the case to the American people after the law passed. It has not worked. And in the final stretch before congressional elections, scheduled for Nov. 2, Republicans are united by their demands for repeal.

The dire circumstances seemingly bode ill for Obama, who campaigned vigorously in 2008 on a promise to bring Americans universal health care. Having staked his presidency on health care, growing voter distrust of administration promises of cheaper and more effective care is a bad sign for Obama's Democrats in the coming elections. They could lose their majority in the House of Representatives and possibly even in the Senate.

Many who wanted the health care system to be overhauled do not realize that some provisions they cared about actually did make it in. And about a quarter of supporters do not understand that something hardly anyone wanted did not make it: They mistakenly think the law will set up panels of bureaucrats to make decisions about people's care, to which critics gave the evocative name "death panels."

The law does include a prohibition on health insurance companies denying coverage to people because of pre-existing medical conditions.

The AP poll was conducted by Stanford University with the Robert Wood Johnson Foundation.

The poll's questions included a true-or-false quiz on 19 items, some of which are in the law and others not. People also were asked how confident they were about their answers.

For the most part, majorities picked the right answers. But a sizable number also got things wrong. Right or wrong, people were unsure of their answers. Two-thirds or more were uncertain about their responses on eight of nine core provisions of the legislation.

Analysis of the findings indicated a split as far as the impact of accurate knowledge, between Democrats and independents on one side and Republicans on the other.

Accurate knowledge of the law made no difference in overwhelming opposition from Republicans.

However, for Democrats and independents, the more accurate knowledge people had of the bill, the more they liked it.

"Among Democrats and independents, the lack of knowledge is suppressing public approval of the bill," said Stanford political science professor Jon Krosnick, who directed the university's participation. "Although the president and others have done a great deal to educate people about what is in this bill, the process has not been particularly successful."

The White House is staging an event Wednesday to mark the six-month anniversary of Obama's signing of the bill. The president and top administration officials will be joined by people from around the country who are already benefiting from such popular provisions as allowing adult children to remain on their parents' insurance until they reach 26 years old.

Will it make a difference?

The poll shows Obama has yet to find the right wavelength for communicating even information that is relatively straightforward. One question stood out as an example:

People were asked whether the Congressional Budget Office had ruled that the legislation probably would increase the government's debt, or whether the nonpartisan budget analysts found that the health law would reduce red ink. (Correct answer: CBO found it would reduce the federal deficit over time.)

But 81 percent in the survey got the wrong answer, including majorities of both supporters and opponents -- even though Obama seldom misses a chance to remind audiences of CBO's favorable report.

Overall, three out of 10 in the poll said they favored the law, while four in 10 said they were opposed. Another 30 percent were neutral. The findings on support and opposition differ from another recent AP poll, but the two surveys cannot be compared because they were drawn up and carried out differently.

The other survey, an AP-GfK political poll, found 41 percent supporting the bill and 46 percent opposing it, with only 12 percent neutral.

The new survey was conducted Aug. 31 to Sept. 7, and involved interviews with 1,251 randomly chosen adults nationwide. It has a margin of sampling error of plus or minus 3.9 percentage points.

The survey was conducted by Knowledge Networks, which first chose people for the study using randomly generated telephone numbers and home addresses. Once people were selected to participate, they were interviewed online. Participants without Internet access were provided it for free.

Stanford University's participation in the project was made possible by a grant from the Robert Wood Johnson Foundation.