Results of National Deliberative Poll^® on Health Care and Education

Nearly one thousand adult Americans engaged in a national scientific experiment to see what the public would think about education and health care—if only people became more informed about the issues and talked about them together. For five weeks, participants engaged in weekly small group dialogues with trained moderators, discussed balanced briefing materials and posed questions to experts from different points of view. Their views changed significantly about policy options, about public officials and about each other.

The experiment, part of PBS Deliberation Week organized by By the People, was conducted by Stanford University's Center for Deliberative Democracy in conjunction with the public opinion research firm Polimetrix of Palo Alto, CA. 981 adult Americans completed pre and post-experiment questionnaires. 360 participated in three or more one hour discussions and the remaining 621 were part of a control group that did not deliberate. There were 150 discussion group meetings (thirty groups meeting for five weeks). The project was conducted online with a nationally representative sample who discussed the issues using voice rather than text, (employing microphones for interaction), and software designed to facilitate group discussion.

All of the changes are statistically significant comparing participant views at the beginning of the experiment with those at the end. Nearly all are statistically significant when deliberators are contrasted with those of the control group who did not deliberate. (Details are in the appendix listing significant changes.)

Education

The respondents moved toward local control and parental involvement as key factors in improving the nation's schools. However, they also lessened their support for vouchers, preferring to put more money into improving the public schools to lower class sizes and increase teacher pay. They also became less supportive of No Child Left Behind and statewide reliance on standardized tests.

Priorities

When asked which of several measures would be the single most important factor for improving the nation's schools, the top three answers —quality of teachers, academic standards, and parental involvement— were the same before and after deliberation. However, there were large changes in the percentage naming each of these three. Before deliberation, teacher quality was the most popular response, named by 34% of participants. But after deliberation, it became the third most frequent response, chosen by only 18% of participants. There was a big increase (14% to 25%), on the other hand, in the number of participants naming "parental involvement" as the single most important factor. The proportion naming aAcademic standards did not change significantly, moving from 33% before to 28% afterward. In addition, participants came to see class size as a more important issue (it was named as the single most important factor before by only 34% before, but by 12% afterwards).

Parental Involvement and Local Control

A follow up question asked which factor was second most important. When these are combined, participants who mentioned "parental involvement" as either the single most important or the second most important factor in improving the nation's schools rose from 27% before deliberation to 45% after.

There was also increased support for local control of education, with the percentage saying that tests for student achievement should be set at the state level decreasing from 62% to 56%, and the percentage saying such decisions should be made by local school boards in each community increasing from 31% to 38% (n=358, p=.02).

Increased Funding

Participants became more supportive of increasing funding to public schools, at least for certain purposes— even if that meant raising taxes. Before deliberation, 58% of participants supported, and 32% opposed, increasing funding for public schools to reduce class sizes, even if that meant increasing taxes. After deliberation, the percent supporting such a measure increased to 63%), while the percentage opposing decreased to 23%. Similarly, the percentage of participants who supported boosting funding for teacher pay increased from 59% to 66%.

Standardized Tests, "No Child Left Behind" and Vouchers

Participants became less enthusiastic about standardized testing in their community's public schools, with the percentage saying there was too much emphasis on standardized testing increasing from 58% to 65%, and the percentage saying there was not enough emphasis decreasing from 14% to 9%. Support for "No Child Left Behind" dropped from 39% to 31% and disapproval increased from 53% to 59%. Opposition to vouchers to pay for private schools with public money increased from 56% to 61%.

Health Care

Participants thought that the number of Americans without health insurance was the single most important problem facing the system and they were willing to support policies that would require at least some sacrifices on their part to deal with it. They were also very concerned with the cost of health insurance and prescription drugs. However, they were less concerned with issues such as medical malpractice, quality of health care for those with insurance and medical errors and mistakes.

Priorities

The top priority for health policy, both before and after deliberation, was dealing with thewith the number of Americans without health insurance (45% before, 49% after selected it as the most important problem). The other most important problem, both before and after, was "the cost of health insurance" (29% before and 30% after, selecting it as the most important). When the top two priorities are considered, the cost of prescription drugs becomes important (rising from 29% to 38% as either first or second most important).

Remedies

Participants moved in the direction of supporting major changes in the health care system, particularly those thatthose that might address the uninsured. When asked "would you be willing to pay more than you do now for health care if this meant that many more Americans would have health insurance coverage" the percentage answering yes went up from 52% to 62%. Support also increased for the US adopting a "single payer" system "where a government entity accepts all health care fees and pays out all health care costs for everyone." The number supporting this option rose from 51% to 57% after deliberation.

Another option which increased in support was the idea of the government requiring all individuals to find at least minimal coverage, perhaps funded by a tax credit. Support rose from 37% to 43% and strong opposition dropped from 37% to 25%. Deliberators also liked the idea of offering uninsured Americans on a voluntary basis "income tax deductions, tax credits or other financial assistance to help them purchase private health insurance on their own." Support for this proposal rose from 57% to 66%.

However, support for some other options either went down or remained stagnant. For example, support for the notion that the government should require employers to provide coverage for their workers declined with deliberation, from 51% to 44%. And support for the notion that the US must increase funding for Medicare and Medicaid, even if this meant increased taxes, was high but unchanged (58% before deliberation and 59% after).

Knowledge

Participants were asked information questions about the two issues and they became more informed overall (as judged by an index for all information items). Some questions showed a dramatic increase. For example, the percentage able to identify the number of Americans who are uninsured (45 million) rose from 30% to 50%.

Tolerance and Efficiacy

Deliberators became more tolerant of those with different points of view. Those agreeing that "people with views very different from mine often have good reasons for their opinions" went up from 67% before to 75% after deliberation, and the percentage disagreeing decreased from 21% before to 8% afterwards. Those who agreed with the statement that "People like me don't have any say about what the government does" went down, from 52% to 38% with deliberation.

Sample Selection and Weighting Methodology

The initial sample was drawn from Polimetrix's PollingPoint panel of approximately one million persons. Panelists were matched to a national sampling frame to be representative of U.S. adults aged 18 or higher. 6,053 persons completed the baseline survey and were randomly assigned to treatment (participation in online discussions) and control groups with probabilities 0.81 and 0.19, respectively. 63% of those assigned to the treatment declined to participate, leaving 1,830 participants in the treatment group and 1,158 in the control group.

The sample was weighted to have the same distribution of racial, gender, age, and education groups as the U.S. Census. In addition, respondents were weighted so that the distribution of party identification would be the same as in the 2004 American National Election Study. Weights were computed by the method of "raking" which attempts to make the sample distributions of each variable match its known population distribution. However, to avoid very large or small weights on individual respondents, weights greater than 6.0 were reduced to 6.0 and weights less than 0.167 were increased to 0.167.

Each control group member who participated in at least three sessions was matched to the most similar control group member (in terms of age, gender, race, education, and party identification). The treatment group was then post-stratified using the weighting system described above and the same weights were applied to matching observations in the control group. This subset is closely balanced between treatment and control group characteristics and was used for comparing treatment-control differences.

The following categories were used for weighting:

Gender	Male	Female
	49.4%	50.6%
Education	HS or less	Some College	College Graduate	Post-Secondary
	45.9%	28.4%	15.7%	9.9%
Race	Blacks	Hispanics	Asian	All other
	15.5%	6.7%	1.8%	75.9%
Party	Republican	Democrat	Independent/Other
	29.9%	30.4%	38.1%
Age Group	18-30	31-44	45-64	65+
	22.1%	26.1%	34.9%	16.9%

Contact Information

The Center for Deliberative Democracy

The Center for Deliberative Democracy at Stanford Univesity employs Deliberative Polling^®1 to study informed public opinion. Professor James Fishkin directs the Center and has conducted Deliberative Polls with various collaborators in the US, Britain, Australia, Canada, China, Denmark, Bulgaria, Hungary and other countries. Professor Robert C. Luskin of the University of Texas has collaborated with Fishkin on all of these projects.

By the People

By the People is a project in citizen engagement organized by MacNeil/Lehrer Productions, with major funding by the Flora and William Hewlett Foundation, with additional funding from the Rockefeller Brothers Fund and the Corporation for Public Broadcasting. It seeks to energize and enhance the national conversation on important issues of the day through a series of national and local broadcasts and events that demonstrate the relevance of national and global policy issues to local concerns.

Polimetrix

Polimetrix is a public opinion research firm in Palo Alto, CA.that employs Web surveys, large-scale databases, and cutting edge statistical techniques. In the 2004 election it conducted over two million interviews through its PollingPoint panel. Representative samples of respondents are drawn from PollingPoint and interviewed on the Internet to collect data with accuracy, speed and efficiency.

¹ Deliberative Polling is a trade mark of James S. Fishkin. Any fees from the trade mark are used to support research at the Center for Deliberative Democracy.