Is Deliberation Neutral ? Patterns of Attitude Change During " The Deliberative Polls TM "

Though deliberative theory has a bias toward rigorous argument and democratic social relations, it presumes that an ideal discursive process otherwise has a neutral stance with respect to particular ideologies and cultural values. This essay provides a preliminary test of that assumption by examining attitude change across a wide range of Deliberation Polls held across the globe. We analyzed 65 questionnaire statements on which Poll participants significantly changed their views on a wide variety of issues. By coding each of these survey items on various value dimensions, we were able to look for any obvious patterns of attitude change. Despite its small size and the exclusion of items showing no attitude change, this sample showed that Poll respondents tend to move toward more cosmopolitan, egalitarian, and collectivist value orientations. Further analysis showed the strongest value-laden shifts were on empirical statements, with public opinion on such questions shifting moderately toward cosmopolitan and collectivist beliefs. The conclusion considers the implications of these findings for deliberative theory, research, and practice.

Critics of deliberative democracy come from both sides of the political spectrum, and many suspect that behind the celebration of reasoned and respectful public debate lie more sinister motives.In an essay critiquing the Deliberation Day proposal (Ackerman and Fishkin 2004), Posner (2004) closed his remarks on a harsh note: I have difficulty suppressing the uncharitable thought that there may be an element of bad faith in the deliberative-democracy movement generally...I think that what motivates many deliberative democrats is not a love of democracy or a faith in the people, but a desire to change specific political outcomes...I sense a power grab by the articulate class whose comparative advantage is-deliberation.
Part of what spooks Posner (2004) is the fact that deliberative event organizers and advocates -are coy about indicating what policies they dislike but would accept.‖Though some have a lingering suspicion of liberal Democratic motives, progressives have also doubted the motives of deliberation projects.Dean Baker, the Co-Director of the Center for Economic and Policy Research, challenged the Our Budget, Our Economy forums convened by AmericaSpeaks in July, 2010.After critiquing the discussion materials and noting a prominent conservative funder, he concluded that the forums were merely a -propaganda exercise‖ (Baker, 2010).
At some point, it becomes possible to subject such concerns to empirical analysis.After all, if the ‗hidden agendas' of organizers are real and effective, they should reveal themselves over time and show that one or another deliberative process lacks neutrality.Even if deliberation has a more Is Deliberation Neutral? -2 subtle bias toward one or another side of the ideological spectrum-one not designed and managed by event organizers, that should also become apparent when one looks at the outcomes of many different deliberative events.
In this essay, we suggest a way to test for underlying ethical, ideological, or political patterns in attitude changes across multiple deliberative events by focusing on the Deliberative Poll, a kind of event that has been repeated numerous times and provides survey measures of opinion before and after deliberation.As much as we hope to address the substantive question of bias in deliberative polling, we principally aim to introduce an original methodological approach to testing for bias in large samples of deliberative processes-a method that could be used with even more diverse and far-reaching samples.

Attitude Change and the Neutrality Assumption
There is a widespread belief among scholars of deliberation that it can improve the quality of democratic decision making (e.g., Chambers 2003;Fishkin 1991Fishkin , 1995Fishkin , 2009;;Gastil 2000Gastil , 2008;;Gutmann and Thompson 1996;Mendelberg 2002;Ryfe 2005).Regarding the mass public, the idea is that -opinions on many issues would be different, were people better informed‖ (Sturgis et al. 2005, 32).Essentially, deliberators arrive at more -enlightened‖ understandings of their own and the public's best interests (Dahl 1989).This is precisely the idea of the Deliberation Poll, the particular form of deliberation examined herein (Fishkin 1991(Fishkin , 1995(Fishkin , 2009;;Luskin and Fishkin 2002).

Is Deliberation Neutral? -3
This leaves open the question of whether those new, reflective opinions might tend to change in the direction of a particular political point of view, moral value, or cultural orientation.Deliberative event organizers' certainly intend to provide complete information that participants from diverse viewpoints conscientiously weigh in the course of their discussion (Fishkin and Farrar 2005).To promote such deliberation, Deliberative Polls use briefing materials loosely based on the model of the National Issues Forum (NIF) (Melville, Willingham, and Dedrick 2005), which provided assistance with the first U.S. Deliberative Poll in 1996.Past research on the NIF showed that even if individuals interpret issues in partisan terms and even reinforce their preexisting views at times, this need not result in an aggregate pattern of change that shifts participants to the left or right of the ideological spectrum (Gastil and Dillard, 1999).Moreover, unlike many informal processes, the Deliberative Poll aims to select a representative sample of the public to ensure corresponding ideological balance among the members, lest one side or the other exert undue influence owing simply to its strength in numbers.This is not to say that aggregate attitudes do not ever change during Deliberative Polls.There exists ample evidence that they can, indeed, change (Fishkin 2009;Fishkin and Luskin 1999), though critics question the frequency and size of such changes (Merkle 1996;Mitofsky 1996a).Rather, the presumption here is that the Deliberative Poll and other deliberative practices like it favor neither the political left nor the political right on one issue after another.Overall, then, the process remains ideologically neutral.

Is Deliberation Neutral? -4
If there were a strong and consistent tendency for aggregate opinion to slide leftward or rightward after deliberative polling, that would suggest that the Poll itself was causing that skew, since it happened across ideologically diverse populations and a variety of issues.Even if deliberation cannot shake people's core commitments, some critics contend that temporary opinion changes at Deliberative Polls could well result not from information and enlightenment but from the participants having a momentary lapse from their underlying values or even their rationality (Mitofsky 1996b;Posner 2003).In effect, the deliberative process itself would be the cause of the shift.

Inherent Bias in Deliberative Process
Is it possible that well-structure deliberation violates the neutrality presumption by systematically favoring an ideology or set of deeper values?Once again, our principal aim is to test for consistent patterns of aggregate opinion movement-not to ferret out the causes thereof.Nonetheless, we will consider, at least theoretically, how deliberation itself might privilege seemingly procedural values that become reflected in the balance of substantive policy opinion itself.

Liberalism
The first bias we consider is the most general-an orientation toward liberalism in the structure of deliberation itself.Though the introduction showed criticisms coming from the left and right, the more common concern is that deliberation has built into it a liberal bias (Kuran 1998), owing to the emphasis on social equality and justice traditionally associated with participatory and deliberative democracy (though the two are distinct; Hauptman 2001).Moreover, some conservative principles, including critical judgment of moral deviants and personal responsibility for misfortune (e.g., Sowell 2002), are simply more uncomfortable to express in a diverse forum (Gastil 2000).

Cosmopolitanism
More compelling than this generic hypothesis of liberal bias are the more precise conceptions of values that might violate deliberation's neutrality assumption.First, consider the orientation toward cosmopolitanism (Vertovec and Cohen 2002), by which we mean the tendency to see issues globally, across national and cultural borders.A cosmopolitan outlook sees it as necessary to relate to other countries and other societies cooperatively whenever possible.Particularistic or nationalistic approaches to public decision making stand as its opposite.
Given deliberative theory's emphasis on hearing different points of view and considering the experiences of all, it is plausible that in the aggregate, Deliberative Polls could tend to promote cosmopolitanism.The orientation toward finding broad public consensus or common ground in deliberation (Chambers 2003;Cohen 1989) is designed to yield -the kind of eye-opening, horizon-widening impact that advocates of deliberative democracy have predicted‖ (Melville et al. 2005, 46).Or, as Chambers writes, deliberation may Is Deliberation Neutral? -6 tend to -broaden perspectives‖ and -promote toleration and understanding between groups‖ (Chambers 2003, 318)-all of which would tend to yield a more cosmopolitan viewpoint.

Sustainability
Another value that may be expressed repeatedly in the attitude shifts that occur during Deliberative Polls is sustainability.The aim of sustainability is to ensure the welfare of future generations, which leads to an emphasis on conserving and carefully managing current resources (Tolba 1984).At a minimum, the principle of sustainability requires that the decisions taken today do not end up being an obstacle for further development and welfare, but in its most expansive form, it promotes the idea of helping future generations thrive by sustaining the conditions for general human and environmental flourishing.
It is a small step to then link sustainability with deliberative democracy, as Dryzek (2000) has done theoretically and others have done through specific deliberative designs (Kashefi and Mort 2004;Ward et al 2003).To the extent that deliberation looks beyond -the most narrow and immediate constructions of their self-interest‖ (Ackerman and Fishkin 2004, 55) and toward optimal long-term solutions to deep-root problems (Burkhalter et al. 2002), it could turn out that aggregate attitude changes that occur during Deliberative Polls favor the value of sustainability.

Egalitarianism and Collectivism
The last two values we consider come as a pair-the two orientations prominent in survey-based versions of -cultural theory‖ (Douglas and Wildavsky 1982;Wildavsky 1987).In this view, people are oriented toward either a hierarchical or egalitarian view of social relations and toward either an individualistic or collectivist/communal vision of public life.Every day, citizens use these cultural worldviews to orient themselves for or against different policies and choices in a complex world (Kahan et al. 2006).
Deliberation is not designed to privilege any particular cultural orientation (though see Gastil et al. 2008).Rather, it is designed to transcend such differences.The hope is that -when individuals of diverse cultural persuasions have the opportunity to engage one another face to face, they become more disposed to listen to one another and to reflect on what they are learning‖ (Gastil et al. 2006a).In this sense, deliberation operates as -an unconstrained exchange of arguments that involves practical reasoning and always potentially leads to a transformation of preferences‖ (Cooke 2000, 948).After all, the culturally-oriented positions with which they began the discussion were simply the results of heuristic processing, such as following credibility and source cues, rather than systematic derivations of principled positions chosen to line up with one's core beliefs (Kahan et al. 2006).
Nonetheless, it is still ambiguous whether asking people to engage in intensive deliberation implicitly promotes one or another of these cultural values.The root of the problem is deliberation's explicit orientation toward the Is Deliberation Neutral? -8 -common good,‖ an undefined position that aims to transcend competing values but may inadvertently promote some over others.As Cohen (1989, 21) argues in one of the earliest formulations of modern deliberative theory, [Deliberation] focuses debate on the common good.And the relevant conceptions of the common good are not comprised simply of interests and preferences that are antecedent to deliberation.Instead, the interests, aims, and ideals that comprise the common good are those that survive deliberation.
The public good promoted by deliberation does not inherently prevent individual values from being promoted, as long as they do not undermine the key principles on which deliberation is built.Nonetheless, deliberation does come with a presumption of shared purpose or at least a public enterprise.This orientation toward shared values and a common good leads us to hypothesize that deliberation could have a -priming‖ effect that leads to a cultural bias.Political psychologists and public opinion scholars have found that such -priming‖ effects commonly occur when particular values are foregrounded or repeated prior to asking people their opinion (e.g., Valentino, Traugott, and Hutchings 2002;Zaller 1992).These primed considerations become more -salient‖ and -recent‖ in the pool of cognitions in one's head.In precisely this way, deliberation's explicit pursuit of a common good could prime participants to seek out policies more consistent with a collectivist (versus individualist) orientation toward public policy.Also, the egalitarianism celebrated in deliberation's procedural equality (Burkhalter et al. 2002) could prime participants into a general orientation toward equal social relations (versus hierarchism) in policy solutions.

Testing for Patterns of Bias Across Deliberative Polls
Such potential biases remain hypothetical to this point, but we believe soon it will be possible to test for them systematically, and in the remainder of this essay, we show one means of conducting such a test by way of publiclyavailable data about the Deliberative Poll.Before describing our research design, however, we wish to say more about the Poll itself.

Deliberative Polling
In 1995, political scientist Jim Fishkin (1995, 162) wrote, The deliberative poll is unlike any poll or survey ever conducted.Ordinary polls model what the public is thinking, even though the public may not be thinking very much or paying much attention.A deliberative poll attempts to model what the public would think, had it a better opportunity to consider the question at issue.Since that time, the Deliberative Poll has become one of the most successful and widely practiced modern deliberative processes (Fishkin 2009;Fishkin and Farrar 2005).1 A Deliberative Poll gathers a representative microcosm of the population to discuss one or more current public issues.The participants receive background material about the issue they are going to discuss, and they participate in small groups discussions which are facilitated by a moderator.Typically, their discussions generate questions that are then posed to expert panels, representing a range of views and knowledge on the topic at hand.Participants are surveyed both before and after the Deliberative Is Deliberation Neutral? -10 Poll has taken place to investigate whether and how their opinions have changed as a result of the deliberative process (Luskin, Fishkin, and Jowell 2002;Fishkin 2009).
Most of the research about Deliberative Polls has focused on the process itself, or on the informational gains for the participants (e.g., Denver, Hands, and Jones, 1995;Fishkin 2009;Goodin et al. 2003;Luskin et al. 2002;Hansen and Andersen 2004;Sturgis et al. 2005).No study, however, has looked for any pattern of attitude change across the many different Polls to detect any of the biases hypothesized above.
From a methodological standpoint, the Deliberative Poll is uniquely well-suited for assessing the cross-issue impact of deliberation on public attitudes.The use of random samples means that the participants represent a broad cross-section of the respective populations, which makes the results more generalizable.More importantly, the Deliberative Poll has collected preand post-deliberation data from a wide variety of settings and on an equally diverse set of issues.This facilitates generalization across a wide range of political contexts and issues.Finally, the Deliberative Polls are always structured in roughly the same way, which makes the obtained data comparable across the different Polls.

Collecting a Test Sample
Unfortunately, a problem with studying Deliberative Polls is that the raw data from the Polls are not, as of yet, in the public domain.Thus, in our methodological illustration, we were not able to examine a complete data set of all the items from all the surveys, which would simultaneously permit an assessment of the direction and frequency of attitude change.At the time this study was conducted, however, the Center for Deliberative Democracy had made available on its website all the survey items that underwent significant change during previous Deliberative Polls.Thus, it was possible to assess the pattern of significant attitude changes but not the rate at which deliberation yielded opinion shifts overall. 2n total, our basic dataset consisted of 65 attitude statements (see Appendix A), each of which yielded a significantly different mean response after a Deliberative Poll when compared to the same item asked beforehand.
With the significance threshold set at a one-tailed alpha of .05,this sample size provided approximately enough power to detect what Cohen (1988) labels as -medium‖ effect sizes equivalent to r = .30.Moving the threshold to .10 adds a modicum of power, making even effects of r = .25sufficiently detectable (i.e., power >= .80).The practical upshot of this is that any nonsignificant findings reported herein can only be used to rule out medium or In the earlier theoretical discussion, we noted how some of these value dimensions are related, and we expected that to result in correlated coder ratings for the different dimensions.To make certain that the value codings were sufficiently distinct, zero-order correlations were conducted among them.
Table 2 shows that the strongest similarity was between Liberalism and Egalitarianism (r = .612).Expressed in the regression terms of an R 2 , this means that roughly a third (36%) of the variance in Liberalism was identical to Hierarchism.The other correlations were more modest, showing overlapping variance of less than 20%.The chief implication of this was our reluctance to interpret these values' unique effects were they to be treated as a set of independent predictors in a regression equation.Instead, we chose to assess each variable separately in relation to the corresponding degrees and directions of attitude changes that occurred during Deliberative Polls.

Measuring Attitude Change
To test the bias hypotheses, we analyzed the relationship between the 65 survey items' value codings and the magnitude and direction of attitude change that occurred during their respective Deliberative Polls.We calculated net attitude change (Change) as the difference between an item's pre-and post-deliberation survey result.Thus, if during a deliberative poll, the Is Deliberation Neutral? -15 participants' attitudes towards an item shifted from 36% support to 51% support, the Change score for that item would be +15 (i.e., 51 minus 36).6

Main Results
Given this simple study design, the basic results are straightforward and can be seen in Table 3.With regard to liberal-conservative ideology, the neutrality presumption received support.In other words, the degree to which a statement embodied liberal versus conservative ideology was not predictive of the direction and magnitude of attitude change.
Looking at more clearly defined value distinctions, however, revealed some significant results.Cosmopolitanism had a positive relationship with Change (r = .21,p = .04).This suggests that Deliberative Poll participants were, on average, likely to change their attitudes in favor of a more cosmopolitan positions on public issues.Sustainability, however, did not yield any significant effects.

Post-Hoc Analysis by Statement Type
Again, these findings are only meant to illustrate the potential for a future study to provide straightforward tests of bias on larger samples including a wider array of survey items (i.e., including those where no significant change appeared).In that same spirit, we wish to demonstrate how additional coding of the data can yield even more refined tests that clarify substantially the precise kind of attitude change taking place in deliberative events.
Though not built into our original analyses, we later chose to break This straightforward post-hoc coding (Krippendorf's α = .78)yielded three subsamples of items, 25 empirical belief statements, 19 opinion/value statements, and 21 policy recommendations.These smaller subsamples could only detect relatively large effect sizes (Cohen 1988), yet even so, they showed that the strongest effects came for the empirical statements in the Deliberative Polling surveys.Changes in responses to these items had large correlations with Cosmopolitanism (r = .43,p = .02)and Collectivism (r = .53,p <.01), two of the three values that yielded significant results in the full sample.
Is Deliberation Neutral? -18 The only two other noteworthy associations ran in opposite directions.
Recalling that in the analysis of all 65 items, Liberalism had no significant association.For empirical belief statements, though, Liberalism was positively associated with Change (r = .32,two-tailed p = .12),whereas policy recommendation statements yielded a negative but non-significant association between Liberalism and Change (r = -.26,two-tailed p = .26).Quite reasonable conventions of statistical significance prohibit indulgent interpretation of such results (but see Ziliak and McCloskey, 2008), but we mention them here simply to demonstrate that they countervailed (and thus concealed in the overall results) the a potential link between Liberalism, Conservativism, and attitude changes.

Discussion and Implications
Though it is customary to begin this final section with a recitation of results, we feel obliged instead to acknowledge our study's limitations at the outset.Once we have made these plain, we will discuss the main findings, then their implications for deliberative theory, research, and practice.

Limitations
As mentioned at the outset, we have presented this investigation as much to encourage a particular approach to bias detection as to glean particular findings.Our motivation stems, in part, from the limited sample with which we undertook this investigation.The sample size we had at our disposal was exceedingly small, and splitting it for post-hoc analyses stretched the credulity of the sample even farther.As a result, any non-significant results reported herein cannot be taken to rule potential value-attitude change relationships, particularly ones of modest size.
Second, though the set of survey items and polling contexts in this study was diverse, they still only reflect the small subset of political settings in which Deliberative Polls were deployed.As explained earlier, our results are also limited by the constrains on obtaining Deliberative Poll survey data.
Using only those items with significant attitude changes meant that our results have no bearing on the question of how often these events actually change attitudes; rather, we only speak to the direction of those changes when they occur.
Finally, it should be clear that the Deliberative Poll represents just one approach to deliberation, which presumes that one-to-two days of loosely structured meetings between citizens, experts, and advocates can yield

Findings
Having made these qualifications, we believe there are three main results that merit notice.This lack of a significant correlation between political ideology and the full set of attitude change measures suggests that, on balance, Deliberative Polls clearly promote neither liberalism or conservativism.In effect was clearest for those pre-and post-deliberation statements concerning matters of empirical belief.
Finally, it is interesting to notice that the liberalism coding correlated with the codings for both collectivism (r = .36)and egalitarianism (r = .61).
This is consistent with prior public opinion research, which has shown roughly the same associations between detailed cultural worldview measures and selflabeling as liberal and conservative (Gastil et al., 2006b).As in that research, the results here suggest that liberalism-conservativism is a crude yardstick for the public's stronger, underlying cultural predispositions.It is noteworthy that both individualism and hierarchism yielded overall effects, whereas there was no overall association for liberalism-conservativism, which expresses those worldviews less precisely and conflates them.

Future Research
It is our hope that research advances the work we have begun by developing the coding approach demonstrated herein.Future researchers may use even more diverse and refined coding categories, hopefully stretched across the outcomes of events beyond the Deliberative Poll.In addition, we recommend that researchers doing content analysis of this type consider a novel approach to coding.Particularly in the case of contested political or cultural values, we suggest surveying a pool of potential coders and identifying pairs of coders with contrasting value-emphases, such as one pair of culturally egalitarian coders and another pair hierarchs.A strong test of Is Deliberation Neutral? -22 inter-rater reliability would determine whether codings were consistent not only within these pairs but also across them.

Knowledge and Values in the Practice of Deliberation
Though more definitive findings await future research, theorists and practitioners alike should come to terms with the possibility that democratic public deliberation may be politically neutral while still promoting certain values or orientations.The single clearest finding in our study was that deliberation appears to promote the value of cosmopolitanism on statements of fact.We consider this no tragedy, since one of deliberation's points of emphasis is broadening people's perspectives.What might give theorists more pause is the potential for deliberation to promote particular cultural orientations, ones that stand in opposition to other broadly-held cultural worldviews.Deliberation's roots in democratic theory and its emphasis on the common good led us to expect an association between participation in the Deliberative Poll and attitude shifts toward egalitarianism and collectivism.
That we found modest evidence of such patterns should prompt theorists to ask whether deliberative theory could be framed in more culturally accommodating terms, or whether it has built into it egalitarian and collectivist values.
As for the practice of deliberation, we wonder whether the Deliberative Poll's emphasis on knowledge gains (Luskin et al. 2002;Fishkin 2009) naturally led to our generally stronger bias findings for empirical beliefs (as opposed to personal opinions or policy recommendations).Had the deliberative method been one that stressed values-based dialogue or the exchange of opinions, perhaps we would have seen values driving changes in those dimensions more consistently.On the other hand, if one would find troubling a pattern of change in opinion or policy recommendation that promoted one value over another, then the information-emphasis would appear to be an effective strategy.In either case, those practitioners who deploy before and after surveys should give careful attention to the type of survey items they provide participants, particularly distinguishing among changes in objective knowledge, empirical beliefs (not subject to verification), personal opinion or value statements, and firm policy recommendations.
Finally, we hope that this study encourages practitioners to reflect on how values are presented in their particular set of procedures and practices.
Deliberation asks people to bring their values into the public sphere, where they can honestly present them and apply them to the problem at hand.Given that people will enter the public sphere with conflicting value orientations, there can and should be value conflict.Though we may hope to arrive at a shared conception of the public good, Briand (1999, 13) cautions that -we shouldn't expect people to compromise just for the sake of compromise...We shouldn't ask them to sacrifice beliefs or values...that they honestly consider central to their lives.But,‖ he adds, -we shouldn't let people off the hook, either.‖Every deliberative practice needs to have a means of bringing forward people's values in a way that is welcoming but still critical.

Is Deliberation Neutral? -24
To the extent deliberation happens in groups, as is often the case in structured events like the Deliberative Poll, it is useful to consider here a finding from research on small group communication.Gouran and Hirokawa (1996) Functional Theory of Group Decision Making is an apt description of the problem-analytic element of public deliberation (e.g., Burkhalter et al. 2002), and it stresses the importance of establishing evaluative criteria-or the values we deploy when assessing alternative solutions to a problem.A metaanalysis of research on Functional Theory found that this is one of the critical predictors of effective decision making (Orlitzky and Hirokawa 2001), yet it is often a part of the decision-making process that group members overlook, under-appreciate, or rush through in their deliberation.
Whether we do this because we (often mistakenly) presume a value consensus or because we wish to avoid difficult moral conflict (Pearce and Littlejohn 1997), the present study hopefully refreshes a longstanding concern with the importance of confronting the role of values in the deliberative process.In the end, the clarification and contestation of values is critical to group effective decision making, and it is essential for the robust development of one's own public judgment.
54. Teacher quality is the most important factor in ensuring a successful education.
55. Parental involvement is the most important factor in ensuring a successful education.
56. Class sizes are the most important factor in ensuring a successful education.
57.I approve of the No Child Left Behind Act.
58.The number of Americans without health insurance is an extremely important problem our health care system faces.
59. Medical errors or mistakes are an extremely important problem our health care system faces.
60.The medical malpractice is an extremely important problem our health care system faces.
61.The cost of prescription drugs is among the most important problems our health care system faces.
62. The US should adopt a single-payer system, where a government entity accepts all healthcare fees and pays out all healthcare costs for everyone.
63.The US should require employer coverage of health care, where every employer must provide insurance for all workers.
64.The US should require individual coverage, where individuals must buy minimal coverage, perhaps funded by a tax credit.
65.The US should offer uninsured Americans income tax deductions, tax credits, or other financial assistance to help them purchase private health insurance on their own.
down our 65 Deliberative Poll survey items to distinguish the different types of claim that they represented.In the course of presenting our preliminary findings, it became apparent that the attitude items in this study represent three distinct classes of beliefs-namely, empirical beliefs, personal opinions, and policy recommendations.As we explained to our coders, the first kind of statement concerns perceptions of reality (e.g., -On the whole, more free trade means more jobs, because we can sell more goods abroad‖).The second type of statement gives a personal opinion (e.g., -I approve of the No Child Left Behind Act‖), and the third constitutes an explicit policy recommendation (e.g., -The US should require individual coverage, where individuals must buy minimal coverage, perhaps funded by a tax credit‖).
knowledge gains and attitude changes.The myriad other deliberative events have different durations and procedures and may yield different results.(For a collection of different methods, seeGastil and Levine 2005.)