The Civic Benefits of Imperfect Deliberation

Normative theories of deliberative democracy stake claims that deliberation, if done correctly, can enhance citizens’ trust in authorities, foster a greater understanding of political issues and the element of compromise in politics, and increase the legitimacy of the political system overall. Skeptics point out that public deliberations seldom fully satisfy the communicative criteria stipulated in normative theory, raising the question: to what extent may we expect imperfect deliberations to generate the promised civic goods? This article proposes a framework for answering this question and also offers a few preliminary answers. The empirical analyses build on an opinion survey conducted in the wake of a series of public meetings that varied in the extent to which they lived up to normative standards of deliberative democracy. The meetings centered around whether to continue construction on a railway tunnel near the town of Båstad in southwestern Sweden. The findings suggest that even imperfect deliberation may have the potential to generate civic goods, though the analyses also raise doubts regarding the durability of the positive effects as well as the extent to which they will develop uniformly among participants.


Introduction
The call for increased citizen participation in political deliberations has become increasingly prevalent in both academic writing (e.g.Fung 2004;Hunold and Young 1998;Petts 2004;Renn 2006;Shapiro 2003;Warren 1996) as well as in administrative guidelines regarding decision-making (Sharp 2002;Soneryd 2004).As with any idea that gains widespread currency, the call for participation is now characterized by a theoretical pluralism that stimulates innovation but also confusion in both conceptual and institutional realms.One recent theoretical development that seems to have generated some degree of consensus, however, is the need to incorporate the normative ideas from deliberative democracy theory into public participation.
Despite the widespread endorsement, and increasingly also usage, of the norms of deliberative decision making, our current knowledge of the effects of such processes remains scant.Advocates of deliberative democracy argue that if done properly, deliberation can generate civic and political benefits that can both revitalize democracy and make the political process more effective and efficient (see Delli Carpini, Cook and Jacobs 2004;and Ryfe 2005 for reviews).In order for such positive effects to arise, however, a deliberative process must measure up to procedural stipulations which are numerous and challenging. 1 Few deliberative processes meet these requirements completely (Gastil 2006;Mutz 2006), giving rise to questions regarding whether a defective (in the eyes of normative theory) deliberative process may still have any of the benefits suggested in normative theory.In other words, departures from the deliberative ideal may diminish the legitimacy of a process in normative terms, but do they also affect the legitimacy of the process in the eyes of participants?If participants leave a participatory process with negative assessments of the experience, participation may actually have deleterious consequences rather than generating civic goods.Participants may, for example, become less willing to engage in public matters and may feel more skeptical of the legitimacy of the political system more generally.The analyses presented here advance a partial treatment of these questions.The analyses explore citizens' evaluations of decision-making arenas that vary in the degree to which they live up to normative ideals of deliberative participation.
A land use issue in Sweden provides the basis for the empirical analyses.A series of public consultation meetings related to the construction of a railway tunnel took place between 2000 and 2002.In some respects, this process departed markedly from the deliberative ideal but in others it fared quite well.More importantly, the meetings differed from one another in terms of design and how well they measured up to deliberative ideals.In addition, decision making regarding railway issues occurred in a number of other communities during the same period of time but with varying levels of citizen involvement and deliberative opportunities.This contextual variation, coupled with micro-level survey data, allows for an investigation of whether deliberative participation gives rise to beneficial effects suggested by normative theories, even when the process does not fully meet all the procedural requirements.
The analysis begins with an examination of whether local residents' assessments of the decision-making process are more favorable in a community in which authorities arranged deliberative participation than in communities in which other models of decision making were employed.The second set of analyses consider how different designs of the deliberative participation are evaluated by the participating public, and whether departures from the ideal of political equality affect the perceived fairness of the process.The third set of analyses addresses the question of the durability of the positive effects of deliberative participation.

Studying the effects of imperfect deliberation
Successful deliberation promises to yield several civic and political benefits, and recent reviews of existing empirical research suggest that a well-structured deliberation does yield some of the promised benefits (Chambers 2003;Delli Carpini, Cook and Jacobs 2004;Mendelberg 2002;and Ryfe 2005 see also Ackerman and Fishkin 2002;Gastil 2006;Morrell 1999;Sulkin and Simon 2001).As Gastil (2006) and others (Fishkin 1995;Sanders 1997;Young 1996) point out, however, public discussions in the form of public meetings or conversation seldom conform completely to the communicative ideals of deliberative democracy theory.Societal power structures are not likely to be left at the door, people may act strategically, and mutual respect is not a given in a group discussion.Some skeptics of deliberative democracy even reject the deliberative model entirely on the grounds that the social, political and psychological impediments are so large that deliberative decision making will only aggravate existing inequalities rather than rejuvenate democracy (Sanders 1997;Young 1996).
Despite the awareness of the difficulty of achieving what Fishkin (1995, 41) terms complete deliberation, we know little about how departures from the deliberative ideal affect, if at all, the generation of benefits theorized to arise from successful deliberation.If deliberators contribute approximately equally to a discussion but only two-thirds present reasoned arguments and only half are open to others' arguments, what portion will attain a heightened understanding of how their own interests are linked to the collective interest, and how many will leave the deliberation more interested in participating in political affairs?Though the noisiness of human behavior precludes ever conclusively answering such a question, at present, knowledge of the implications of incomplete deliberation is meager indeed.
Full understanding of the implications of deviations from each of the various criteria of complete deliberation for each of the potential beneficial outcomes would require observing a large number of deliberative processes that differ on these characteristics to determine how variations in the process affect each of the theorized beneficial outcomes.The following discussion suggests one means of simplifying the analytical model in order to shed light on several outcomes at the same time.The core of the argument is that in order for a deliberative process or public meeting to generate certain positive effects, those participating must, in general, evaluate the experience favorably.A negative assessment of a deliberative process may plausibly backfire and instead undermine participants' civic virtues.
First, however, a definitional note is in order.Deliberative participation is used here to denote a form of decision making in which citizens engage in discussion with decision makers to weigh the merits and problems of different alternative solutions in a specific matter of public concern.Deliberative participation as used here closely resembles Fung and Wright's (2003) concept of Empowered Participatory Governance (see also Fung 2004).The concept of deliberative participation weds models of public participation that have been debated, used and evaluated for several decades, with the communicative ideals of deliberative democracy (Fung and Wright 2003:15-23). 2The concept therefore differs both from deliberative democracy, which generally does not entail that citizens deliberate together with authorities with decision making power, and from participatory democracy, which generally implies local forums to solve local disputes (Mansbridge 1980), or grassroots mobilization to advance a specific position or agenda (Fung 2004:5;Mutz 2006).As this analysis is primarily concerned with the claims of deliberative democracy theory, the discussion below explores the reciprocal relevance of this study and the theoretical debate on deliberative decision making.
2 Since it is the normative claims of deliberative democracy theory that are the central focus of this analysis, I choose to use the term deliberative participation rather than opting for Fung and Wright's term.The model of decision making examined here does, however, concur with Fung and Wright's model in terms of the properties of the design.The process dealt with an issue localized in issue and geographic space, was intended to link local input to superordinate bodies, and occurred on the cusp of a legislative reform that has since made such processes a requirement.See footnote 8 for a clarification of this point.
Deliberative democracy theorists argue that political deliberation helps citizens and decision makers alike to formulate more well-founded opinions that reflect not only their own needs and experiences but also an understanding of how issues may affect others (Chambers 2003;Gutmann and Thompson 1996;Fishkin 1995).Deliberation, in other words, produces a more enlightened, informed and reasoned public opinion.The infusion of a greater breadth of information and critical commentary into politics will also result in more well-founded and higher quality political decisions (Dryzek 2001;Estlund 1997;Habermas 1984;Lidskog 1997).
Other promised benefits of public political engagement in what has also been called discursive participation (Delli Carpini, Cook and Jacobs 2004) include enhanced public appreciation of the element of compromise inherent in collective decision making, an understanding of the potential of collaboration with others for solving shared problems, and on the whole enhanced legitimacy of the governing apparatus (Cohen 1989;Benhabib 1996;Bohman 1996).In the areas of land use planning and risk management, deliberative public participation may constitute the only means of allaying suspicions that public officials collude with big business at the expense of the local community (Bradbury, Branch and Focht 1999;Fischer 1993;Rabe 1994).
The difficulty of assessing the potential benefits of deliberative participation in real political life lies in the immense potential variation in the design and quality of deliberative processes, but also the considerable number of benefits argued to arise from successful deliberation.In order to attain a full understanding of the potential benefits or risks of deliberative participation, we would need to determine the extent to which each aspect of the deliberative process contributes to fostering or undermining each of these potential benefits.To complicate matters further, there are good arguments supporting the notion that what can be considered proper procedures varies depending on the social and political context, and on the nature of the issue being deliberated (Renn 2006).One means of making the task more manageable is to identify a precondition, or common denominator, of at least some of the promised benefits, and investigate whether different designs of deliberation are associated with such a measure.
The common denominator proposed here is participants' assessments of the fairness and propriety of the deliberative process.Unless those participating in a deliberative process find the process to be procedurally proper and fair, it is unlikely that they will become more accepting of the element of compromise in political matters.Nor are they likely to regard the political system as more legitimate, nor the decision more well-founded and acceptable.Previous empirical research has shown that perceived procedural fairness cultivates confidence in public officials, and especially among those who have been actively involved in an issue (Grimes 2006;Tyler 2006).Examining the effects of different procedural approaches to deliberative participation on citizens' assessments of procedural fairness therefore allows us to shed light on the potential of deliberation to generate other civic goods.

Prerequisite conditions of successful deliberative participation
Many have claimed that successful deliberative participation can enhance the legitimacy and efficiency of policy formation and implementation; few have, in contrast, claimed that arranging successful deliberative participation is a simple enterprise.What distinguishes deliberation from other forms of talk, such as chitchat or negotiation, is that it is characterized by values such as empathy, egalitarianism, open-mindedness, and reason (e.g.Chambers 2003;Delli Carpini, Cook and Jacobs 2004;Gutmann and Thompson 1996).This analysis will focus on six procedural attributes that may contribute to constructing an environment conducive to such dialogue: non-manipulation, open access to agenda-setting, that citizens have a reasonable opportunity to influence the outcome, distribution of knowledge, mutual respect, and political equality.3A brief explanation of each sets the stage for the empirical analyses.
Starting with the most obvious, the process should be free from manipulation, such as for example that the issues on the table in actuality have been decided elsewhere in arenas completely apart from the deliberative process (Petts 2004).Authorities must be forthcoming with information so that people interested in the issue have the opportunity to engage in meaningful dialogue (Bohman 1996).Furthermore, the selection of participants should not be guided by an ambition to avoid conflict or exclude certain points of view.Research has documented instances in which authorities appear to assemble participants known to be amenable to the agenda of the authorities themselves (Sharp 2002, 19), in which case the participatory process only serves to legitimate a decision rather than offer the public a meaningful arena for deliberation and influence.The logistics of the deliberative process in terms of time allotted to discussion, availability of necessary communication aids, and appropriate scheduling and space arrangements must also be designed to allow all interested individuals to participate. 4 second procedural issue that has grown out of attempts to define and delineate the mechanisms of exercising political power relates to agenda-setting.Dahl's early definition of power equated power to the ability of translating one's own preferences into finalized political decisions (Dahl 1956).Bachrach and Baratz (1962) modified that theoretical view in observing that the persons or parties who participate in defining the contours of the political debate also have an influential role in shaping the outcomes of the process.While generally not phrased in terms of open agenda-setting in deliberative democracy theory, arguments have emerged that deliberation should exclude no one, and be free and equal to all (Knight and Johnson 1994).If authorities alone have the power to select the topics of discussion, the deliberation becomes circumscribed and authorities retain a degree of control that also violates the ideal of egalitarianism.
Third, participants must have a reasonable opportunity to influence the decision outcome or at least influence decision makers' conceptualization of what is at stake in the decision.As mentioned above, if authorities pose the deliberative process as a channel of influence while in reality the decision is being made elsewhere, then the deliberative process would be blatantly manipulated.Even barring such a distortion, a process may still vary in terms of the degree of influence promised to, and in the end granted, participants.The core of deliberative democracy is openness to others' arguments; in deliberations between citizens and authorities, authorities must also be willing to incorporate relevant arguments and information.
While the first three procedural stipulations related to the rules of order of the process, the final three procedural stipulations are of a different character as they place demands on all participants involved in the deliberation, authorities and citizens alike.The first of these relates to knowledge and willingness to partake of information provided by other actors involved.Authorities must provide information regarding the project and associated risks, and participants ought ideally to partake of information from various sources in order to understand what the project may entail for themselves and the local environment.In short, potential deliberators should have sufficient knowledge about the issue under consideration to enable them to engage in deliberation if they so desire.
The fifth and perhaps most central procedural attribute of successful deliberative participation deals with the disposition of the authorities and participants themselves.A key to deliberation is that all parties involved must be open to and willing to consider the arguments and experiences of others and participants show one another mutual respect.Individuals involved in deliberation ought to strive to provide reasonable arguments for their own points of view in terms that are justifiable to other participants (Gutmann and Thompson 1996; in which the first language of most local residents was Spanish.Such a decision on the part of those officiating the process must be regarded as a manipulation of the deliberative process.

2000)
. Ideally, deliberative discussion should be characterized by a shared desire to reach an optimal collective solution, rather than merely attempting to 'win' by dominating or cajoling other participants.These points are perhaps the most difficult to translate into rules of order, as they largely depend on the mental disposition of the participants themselves (Ryfe 2005).
The behavior and demeanor of authorities arranging the deliberative participation can play a decisive role in setting a tone of mutual respect.Social psychological research has shown that authorities' demeanor, including whether or not they treat participants with respect, demonstrate receptivity to others' worldviews, and provide explanations and justification for their own ideas, can send signals to participants regarding their own status in the deliberative process (Tyler 2006;Tyler and Lind 1992).Authorities' behavior can also validate or undermine the status of some participants in the eyes of other participants, thereby helping to establish, or erode, respect and egalitarianism in the group.Inappropriate behavior from authorities in this regard can affect the perceived fairness of the process among those who feel slighted and perhaps among participants more generally.
Finally, a stipulation of successful deliberation requires that the process embody the principle of political equality.The weight of a person's voice in a participatory process may depend on mastery of technical knowledge, on ability to convey a point in an expressive and compelling fashion, on self-confidence in public speaking, even on the pitch and resonance of a person's voice.Inequality in a deliberative process may arise from self-censorship stemming from a lack of self-confidence or a feeling of having a minority opinion (i.e.Noelle-Neumann's spiral of silence, 1993).The weight of one's voice in a participatory arena may depend on the social and economic status of the individual.In short, participatory democracy may lead to the reproduction of economic and political power relations more effectively even than a representative model of democracy (Mansbridge 1996).
The empirical analyses of the decision-making processes reveal that on two of these criteria -non-manipulation and distribution of knowledge -the ten meetings examined all fared fairly well.Variation among the meetings existed on three of the criteria -open agenda-setting, opportunity to exert influence, and the degree of mutual respect.The criterion of political equality is instead examined at the individual level.
Before launching into the analyses, a brief summary of the methodological approach used to answer each of the research questions will bridge the preceding discussion with the empirical analyses to follow (see Table 1).The first question asks, to reiterate, whether deliberative participation is perceived as more procedurally fair than other forms of decision making.The survey material includes respondents who did participate in the deliberative meetings while others did not, as well as respondents who reside in communities in which authorities employed other approaches to decision making.A comparison of assessments from these different groups of respondents, controlling for assessments of other relevant attitudes, will address the first question.The second question deals with whether deviations from the normative stipulations of deliberative participation affect public assessments of procedural fairness.As mentioned above, the study followed a set of meetings that differed considerably from one another on key points.Do deviations from the normative criteria affect the perceived fairness of the process?To examine whether deviations from the principle of political equality have a detrimental effect on the perceived fairness of the process, the analyses explore the assessments of individuals who belong to groups generally considered to be disadvantaged in political power structures, or who feel inhibited to participate in a deliberative process.The third of the research questions deals with a question of a somewhat different nature and will be introduced in more detail in later sections.a Two of the process stipulations -non-manipulation and distribution of knowledge -are used to establish that the process in this case met two important conditions of deliberation.

Methods
The analyses build on data collected through two mail surveys, as well as direct observation of the public consultation meetings and communication with Rail Administration officials.One of the surveys was sent in February 2003 to those individuals who own property in the area identified as most likely to feel the effects of the construction of a technically complex railway tunnel.The list of addresses used for administering the survey is the same list the Rail Administration used to distribute notices of upcoming public consultation meetings.Of the 735 who received the survey, 365 (50%) responded.This survey explored local residents' assessments of the ten public consultation meetings and of the tunnel project more generally.
Of the 735 who received the survey, exactly 511 had mailing addresses in the area affected by the tunnel project, and the remaining 224 owned either summer homes or non-residential property in the area but live elsewhere.The respondents were representative in terms of local or non-local residence.In terms of gender, respondents were also representative of the population; men constituted 57 percent of the respondents but also constitute 55 percent of those owning property in the area. 5he relatively low response rate despite the extremely high salience of the issue may be the result of the fact that the survey questionnaires were not marked with identification numbers.Extensive interview-based research with several of the residents of the area revealed an intense distrust toward the Rail Administration and toward public officials in general (Sjölander-Lindqvist 2004).In order that even the more critical and distrusting respondents would feel assured of complete anonymity, questionnaires were not marked in any way.This choice of strategy precluded sending targeted reminders to those who did not respond.
Another, much larger survey, both in scope and sample size, was conducted in September of 2002.The sample consisted of 3,000 individuals, 500 in each of six communities also affected by railway expansion projects.This survey also asked respondents to assess the decision-making processes surrounding the railway construction project in their own community. 6The response rate of the 2002 survey was 69% with respondents distributed evenly among the six communities.

The case
The Hallandsås railway tunnel The tunnel project and the deliberative participation process were the responsibility of the Swedish National Rail Administration (Banverket), and concerned an 8.5 kilometer (5 mile) long tunnel being drilled through a 200 meter high ridge near the town of Båstad in southwest Sweden.Drilling of the tunnel began in 1994 but stopped in 1997 due to technical complications of rather large proportions (Boholm and Löfstedt 1999).The ridge consists primarily of sand, gravel, soil and water, conditions of which the Rail Administration claims it was not fully aware at the outset (though geologists have argued that knowledge of the conditions was available before drilling began, Falkemark 1998).Massive amounts of water drained into the tunnel, largely depleting the aquifer and leaving communities atop the ridge, which rely entirely on well water, without water.A botched effort to stem the flow of water ended in the release of toxic chemicals into local streams.Cows showed signs of severe neural damage and fish and crayfish died in large numbers, at which point the government ordered the tunnel work stopped.
The local negative implications of the tunnel were primarily of two types: the release of toxic chemicals into surface water in the late summer and fall of 1997, and the emptying of groundwater tables leaving residents without access to water.Of the 369 respondents in the tunnel area, approximately every fourth person (98 people total) reported having suffered the consequences of the drainage of the groundwater table.A smaller number, 43 people, reported having been affected by the toxic leakage.
According to the Rail Administration itself, both types of negative effects have been remedied.In the case of the toxic spill, compensation consisted primarily of economic remunerations to cover the cost incurred when contaminated agricultural products had to be destroyed (including even livestock and fish stocks).With respect to the groundwater problems, the complexity and severity of the problem necessitated more intricate solutions.For some households and farms, the Rail Administration attempted to restore access to groundwater tables by drilling new and deeper wells, and in other cases delivered water by tank trucks.Neither of these proved to be viable long-term solutions, and several properties have therefore been incorporated into the municipal water system (a costly venture given the distance from the town of Båstad).
Respondents expressed varying levels of satisfaction with the Rail Administration's handling of these two problems.Of the 98 people who reported suffering the consequences of the drainage of groundwater tables, slightly less than half (45 percent) felt the problem had been resolved satisfactorily.The remaining respondents with water problems report that the Rail Administration has provided an unsatisfactory solution to the problem, or no solution at all.Among those individuals who felt the effects of the toxic spill, about a third (37 percent) felt they had received sufficient compensation; the others expressed dissatisfaction with the Rail Administration's compensation of damages.
The deliberative participation process In May of 1999, the government commissioned the Rail Administration to conduct an Environmental Impact Assessment (EIA) to determine the feasibility and environmental consequences of continuing construction (Sjölander-Lindqvist 2004, 25).The Rail Administration arranged five public consultation meetings between January and September 2000. 7After the completion of the EIA, the Rail Administration continued to arrange approximately three meetings every year.These two sets of meetings, those held in preparation of the EIA and those held subsequent to the EIA, differed markedly from one another in terms of tone and design.
The ten meetings were similar, however, on two of the six procedural criteria delineated above, non-manipulation and distribution of knowledge.An evaluation of the meetings on the criteria on which the meetings differed will be discussed later.Evaluations of the meetings build primarily upon attendance at the meetings and official minutes, and when appropriate also on respondents' own assessments.
All of the meetings included in the analysis satisfied the first criterion fairly well; the meetings were not distorted by blatant manipulation.The time and opportunities for participants to express comments and ask questions was ample throughout.All of the meetings followed the same structure.The Rail Administration used the first hour to present information and address questions raised since the previous meeting.During the second hour (sometimes more), the floor was open for questions and comments.Participants were encouraged to submit written comments and questions if for some reason they were unwilling or unable to express them during the meetings.Roughly a third of survey respondents felt, however, that the time for discussion had not been sufficient.
No apparent manipulation occurred with respect to hand-picking participants in the deliberative process.The meetings were open to the public and were announced in local media.In addition, all individuals owning property within the designated 'affected area' also received notices by mail.Survey respondents overwhelmingly (95 percent) agreed that the Rail Administration had invited all relevant parties to participate.Most respondents (60 percent) stated that they had received clear information regarding upcoming meetings.
The second procedural stipulation for successful deliberative participation, distribution of information and knowledge, can also be said to be met in this case, perhaps more than in most other issues and in most other settings.The circumstance that the meetings followed in the wake of an environmental crisis heightened public awareness and prompted local residents to consult experts and seek information regarding the legal and technical details of the project.The technical intricacy of many comments and questions raised at the meetings indicated a high level of technical and juridical knowledge among participants.Among those who responded to the survey, fewer than ten percent assessed their own awareness of the issue below the midpoint of a 0 to 4 scale from 'uninformed' to 'very well informed'.Public interest in attending meetings was sufficient to sustain a participatory process, though it diminished over the course of the process.Attendance at the first meetings was approximately 150 but dropped to around 15 at the later meetings.Roughly half (52 percent) of the respondents reported that they had not attended any of the ten official public consultation meetings.About 40 percent of the respondents attended one or more of the first set of meetings, while 30 percent attended one or more of the second set of meetings.Some respondents, 23 percent, attended meetings of both the first and second set.

Measuring perceived procedural fairness
The main aim of these analyses is to examine public assessments of a deliberative process compared to other forms of decision making, and also whether assessments of procedural fairness are sensitive to deviations from normative stipulations.As argued above, studying the effect of procedural design on the perceived fairness of the process seems like a plausible approach to studying the promise of deliberative participation, as perceived procedural fairness may be regarded as the sine qua non of many of these potential benefits.
The analyses examine a conceptualization of procedural fairness that previous research has indicated has bearing on the perceived legitimacy of decision-making authorities and of the decision outcome itself, here termed public justification.The concept builds on a battery of questions regarding the authority's behavior in interactions with meeting participants, in particular their openness to concerns and worries, and willingness to give a comprehensible and acceptable account of their own thinking and decisions.
The concept was measured slightly differently in the two surveys used in the analyses.In the 2003 survey sent to the residents most affected by the tunnel project, the public justification construct consisted of the mean of six items.The six questions ask respondents to rate the Rail Administration on the following aspects of the public consultation meetings: receptivity to input from participants, consideration of concerns, quality and volume of information provided, incorporation of participant input, justification of decisions, and whether the Rail Administration brought all relevant issues up for discussion.
Though in some sense analytically distinct, assessments of these various aspects overlap strongly with one another.Correlations among the six components are strong both at the aggregate level but also within various subgroups of respondents, such as those who have been most actively involved (reliability coefficient Crombach's α for the six questions is 0.92 for the whole sample).
In the survey directed at residents of the six other communities, the public justification construct builds on three rather than six survey questions relating to whether the Rail Administration had: listened to citizens' concerns, informed residents about the issue at hand, and, third, shown consideration for the local community (phrased in the negative to avoid an acquiescent response set, Crombach's α = 0.51).

Deliberative participation compared to other forms of decision making
Experimental research has illustrated that it is by no means a foregone conclusion that citizens who participate in a deliberative process will deem public officials to be more legitimate or the decision-making process more fair (Hibbing and Theiss-Morse 2002;Morrell 1999).Morrell (1999) shows that students perceived decision making structured according to the rules of parliamentary procedure to be more fair than a discussion structured to generate complete deliberation.
This first set of analyses compares the process assessments of five different groups of respondents who vary in their participation in the railway issue.The five groups are residents of the ridge area who have been involved in the deliberative participation, 2) ridge residents who have attended informational meetings regarding the tunnel but not the deliberative meetings, 3) ridge residents who have not attended any meetings arranged by the Rail Administration.These groups are also set in contrast to public evaluations of the decision making process in other railway expansion projects, both among 4) residents who have attended meetings of various kinds and 5) residents who have not been active in the railway issue.Although opportunities for dialogue between the Rail Administration and residents existed to varying extents in all of the other communities, nowhere were deliberative opportunities as extensive as in Båstad.
Table 2 shows the estimated marginal means of these five groups' assessments of public justification, controlling for levels of trust for national political institutions.Previous research has observed that individuals who see political institutions as more responsive, are more likely to engage in various forms of political participation (Bäck, Teorell and Westholm 2006, 60;Milbrath 1977, 72).If participating individuals in this case show more positive assessments of the Rail Administration, it may in other words simply be the effect of selfselection rather than an indication that involvement in deliberative participation enhances the perceived fairness of the process.Though the survey material used here did not include measures of efficacy, controlling for trust for national political institutions provides a means of taking the self-selection possibility into account in the analysis.The perceived responsiveness of a political institution has strong bearing on trust for that institution (Grimes 2006).
Does deliberative participation, then, increase the perceived fairness of the decision process?The first set of analyses suggests that indeed it may.The analysis shows that deliberative participation is associated with more positive evaluations of public justification, even when political trust is controlled for (Table 2).Residents who report having attended one or more of the consultation meetings generally evaluate the Rail Administration more charitably on public justification (3.2 on 0 to 6 scale; see note a, Table 2) than residents who received notices but for whatever reason elected not to attend (2.7).However, residents who attended informational meetings but did not attend the consultation meetings were also quite positive (3.2).The difference between attenders and non-attenders is statistically significant.8

Mean of Public justification a
Confidence intervals (95%) Ridge residents who: have attended one or more deliberative meetings 3.2 N=153 3.0 to 3.5 have only attended informational meetings 3.2 N=47 2.8 to 3.7 received notices but did not attend any meetings 2.7 N=107** 2.3 to 2.9 Residents of other communities who: have been active in some way b 3.0 N=401 2.8 to 3.1 have not been active 2.6 N=1476*** 2.6 to 2.7 **p < 0.05, ***p < 0.01 a In order to increase the comparability of the public justification construct between the two surveys, the public justification measure from the Ridge survey only builds on the three questions that were also asked in the survey administered in the other communities.In the remaining analyses, the six item index is used (see Appendix).b Active in other communities means attending an informational meeting, involved in a pressure group, or contacting the Rail Administration directly.
Moreover, those who have attended the consultation or informational meetings in Båstad hold more positive procedural assessments than active residents of other communities (3.2 and 3.0 respectively), where the railway construction did not entail negative ramifications for most residents.Though it is difficult to assess whether this difference is statistically significant as the two values are computed from different samples, a glance at the confidence intervals suggests that the difference may not be the effect of random variation.Residents of other communities who have had direct contact with the Rail Administration (400 individuals, either by attending an informational meeting, having been involved in a pressure group, or contacting the Rail Administration) were also more positive than their inactive neighbors (3.0 and 2.6 respectively, p<0.01), but still not as positive as ridge residents who had attended public consultation meetings (3.2).
The results presented in Table 2 suggest that arranging public consultation meetings can in fact lead to more positive assessments of the decision-making process, provided that the process is not afflicted with blatant manipulation, which in this case it was not.Considering the history of the tunnel construction and its local implications, it would have been reasonable to expect that those who own property in the area would be much more critical of the Rail Administration than any other group of citizens in Sweden.Instead, assessments of the Rail Administration's openness to input and willingness to give an account of its plans and decisions are comparatively positive.A finding which is more difficult to explain, however, is that Båstad residents who attended informational meetings but not consultative meetings were also comparatively positive in terms of the public justification measure.Nonetheless, on the basis of this first analysis, there is reason for cautious optimism.9 The effects of process design on perceived procedural fairness Though the ten meetings arranged by the Rail Administration were similar in some respects, they differed from one another on three of the six procedural criteria discussed above: whether the decision outcome was open for modification, access to agenda setting, and the degree to which the authority had a demeanor that set a tone of mutual respect.Fortuitously (for the purposes of this study), the variation was contained to two fairly distinct sets of meetings: the five held between January and September of 2000 in conjunction with preparation of the environmental impact assessment, and the five held between February 2001 and November 2002.The two sets of meetings were held in different locations, led by different moderators, and occurred under the leadership of different project managers from the Rail Administration, factors which increase the likelihood that survey respondents correctly recall which meetings they attended.As the following analysis reveals, however, the first set of meetings was normatively better in some respects, the second better in others.This circumstance perhaps accounts for the lack of clear answers regarding the effects of such procedural variations.
Decision outcome open to modification.The two sets of meetings differed markedly in terms of the legally stipulated, as well as publicly announced, possibility to influence the future of the tunnel project, with the first set offering a greater scope for influence than the second set.The first five meetings were carried out in conjunction with the preparation of the environmental impact assessment (EIA), and a summary of citizen input from these meetings became one of the grounds upon which the government based its decision regarding whether to permit construction of the tunnel to proceed.Questions and concerns expressed during the first five meetings therefore reached not only the ear of the local project management but also the authority with power to decide the fate of the tunnel project.The second set of meetings in contrast occurred after the government had made the decision to allow construction to continue and dealt with how to contain the negative externalities for the local environment. 10he opportunity for influence (or lack thereof) was also indicated at the beginning of each meeting.In his opening statement at the first of the 10 meetings, the then project manager for the tunnel project explained the aim and purpose of the meetings with the following words: The objective of today's meeting … and with future public consultation, is first and foremost to partake of your views.Your comments and views will be taken into consideration in the upcoming work of preparing the environmental impact assessment.It is my own desire that this first formal public consultation meeting will also be a symbol for a new beginning in a process of building support through a good dialog with all of you that are here and that are affected (Statement by project manager at the first consultative meeting, January 2000.
As the quote indicates, the project manager welcomes input and explains that the meetings held in conjunction with the preparation of the EIA would allow residents the opportunity to weigh in on whether or not the tunnel construction should be allowed to continue.
The second string of meetings, which went under the new name 'Consultation Forum Hallandsås', were expressly intended as an opportunity to "…confer, exchange information, discuss, discuss what has happened, field questions or bring up things that people can be worried or curious about or that people think is wrong" (introduction by moderator, 25 November 2002).The Rail Administration chose not to frame this second set of meetings as an opportunity to exert influence but rather to engage in dialogue.
Agenda-setting.For the first set of meetings, the Rail Administration retained power of agenda-setting with respect to topics covered during the first hour of the meeting (the second hour was reserved for questions and comments from the floor).At the second set of five meetings, however, the Rail Administration welcomed the general public to submit questions or topics before meetings and also invited representatives of several organized stakeholder groups to 'pre-meetings' to discuss which issues would be covered at the main meetings.11This change in meeting design was more apparent than in the matter of influence opportunities; the moderator explained this point of order at the beginning and end of each meeting.
Mutual respect.Perhaps the most striking difference between the first and second sets of five meetings, however, relates to the demeanor of Rail Administration officials.The distance between decision makers and meeting attendees was more pronounced both physically, technically, and socially at the first set of meetings than at the second set.At the first five meetings, the Rail Administration together with its corps of consultants, technical experts and lawyers, sat on an elevated stage and presented detailed technical information.Visual images in the form of graphs and simulations of the tunnel project were in some instances scarcely discernable from the seating area.The spatial arrangement of public officials sitting on a stage about 1.5 m above the auditorium floor, together with the rather impenetrable presentation of technical information, created a power gradient that seemed quite marked to this observer.On a few instances, residents' questions met with disdainful responses from authorities.In the second set of meetings, the Rail Administration employed a different tack.Information was presented more pedagogically and Rail Administration officials and advisors sat among other participants rather than on an elevated stage.
In sum, the first set of meetings complied with the normative ideal of ability to exert influence, but the second set was more deliberative in terms of open agenda setting and mutual respect.Do these differing approaches to deliberative participation affect the perceived procedural fairness of the process?Table 3 shows the comparison of the evaluations of people who attended the two different sets of meetings.The assessments of the fairness of the decision process do vary among respondents who attended the different sets of meetings, though not in the way one might expect given the differences between the two sets of meetings.Respondents who reported having attended only meetings of the first set had a mean assessment of public justification of 3.3 on the 0 to 6 scale, while the mean of those who reported only having attended meetings of the second set was 3.4.The only group that differs significantly (p=0.03)from the others are residents who had attended meetings of both the first and second sets (mean = 2.6).
Before turning to an interpretation of these results, we must first consider the fact that respondents were by no means randomly assigned to attend the different sets of meetings and that these values can therefore not be interpreted at face value.It is plausible, for example, that dissatisfaction with the Rail Administration's handling of the tunnel work prompted local residents to attend the meetings.Self-selection could account for the fact that those who have attended meetings of both the first and second sets also are the most critical.
Table 3 therefore also shows the estimated marginal means of public justification assessments once these historical factors are taken into account.Controlling for negative experiences (for example loss of ground water, exposure to toxic chemicals) does considerably mitigate the difference between those who have attended meetings of both sets and others.The differences among mean assessments are not significant once negative and positive experiences with the tunnel are taken into account.Public justification here builds on six questions as indicated in the text (see Appendix).b The estimated marginal means are computed using a model including the following covariates: ground water problems, satisfied with solution to groundwater problems, affected by toxic spill, satisfied with compensation for toxic spill, expect or have experienced benefits from tunnel construction, expect or have experienced other negative consequences from tunnel construction.
The most striking conclusion of these analyses must, however, be that the procedural design of the meetings does not, in this case, seem to have had a strong and lasting impact on the perceived fairness of the public consultation meetings.
The null result may be a reflection of the fact that the strengths, in terms of normative theory, of the two sets of meetings cancel each other out.The null result may also be a factor of time.At the time of the survey, several years had passed since the first set of meetings.It is not certain that any disapproval of the process that might have existed would persist in respondents' memory three years later.
We cannot dismiss the possibility, however, that deviations from the normative ideal, provided they are not egregious, do not have a strong and averse impact on the assessments of authorities in terms of public justification.Despite the fact that the second set of meetings complied with normative stipulations to a greater degree -respect to a more inclusive approach to agenda-setting and in the more egalitarian, respectful demeanor of authorities -those who only attended meetings of this second set are no more laudatory of the process than those who only attended meetings of the first set.
A deliberative process for all?Equality and procedural assessments One normatively thorny issue relating to deliberative and participatory democracy is how to honor the principle of political equality.The principle implies that citizens or members of a political association should be treated as having equal inherent worth in political decisions, and that no individual or group of individuals should have de jure greater influence in selecting political alternatives or representatives than others.A number of advocates of deliberative democracy have convincingly argued that representative and direct democratic systems have never, in reality, come close to realizing the principle of political equality (e.g.Fishkin 1995, Mansbridge 1996).This critique of representative democracy, however valid, does not diminish the importance of exploring the normative concern with equality in deliberative participation as it has taken form in the real political world.
Empirical research has shown that societal power structures in fact can leave an imprint on small group deliberations.Della Carpini, Cook and Jacobs (2004, 326) cite research indicating that women (at least when the research was conducted, in the 1950s) and black jurors participate less in jury deliberations and express lower levels of satisfaction with the process than men and white jurors (Mendelberg and Oleske 2000).Thus even if deliberative participation were structured so as to ensure equality among participants during the deliberation, it may also have to contend with inequalities that affect participants' behavior during the process.
Cleavages the likes of which may disrupt the deliberative process in juries in the United States do not exist in the setting examined in this analysis, though they certainly cannot be said to be absent in Sweden altogether.The public consultation meetings examined here occurred in setting in which egregious inequalities do not exist, presenting a favorable case for deliberative democracy theory.The analyses below consider whether women and men differ in their assessments of the process, and whether individuals with professional careers differ from individuals with more low-skilled jobs.
In addition, the analysis examines whether individuals who feel reluctant to participate have more critical assessments of the process.The survey included a battery of questions regarding whether respondents had ever felt that they wanted to say something at the consultation meetings but refrained, and if so why.Three of the items related to concepts that have been raised as normative concerns in deliberative processes.Those three reasons for abstaining from participation were: the perception that one's opinion differed from others' present at the meeting, reluctance to speak in front of others, and the perception that one lacked the requisite knowledge of the issue.12Individuals who abstain from speaking in a participatory decision process -whether due to socioeconomic factors such as education, contextual factors such as holding a minority opinion, or psychological factors such as self-confidence -do not, in practice, have the same opportunity to exert influence as others.As many as 76 respondents indicated that they at some point abstained from speaking for one of those three reasons.Are those who feel reluctant to participate actively in participatory processes, less enthusiastic about the process itself?
Factors such as gender and professional status, often cited as bases of inequalities in political life, do not have a strong and systematic bearing on individuals' assessments of the deliberative participation in this case.Women are slightly more critical of the process, but considerable variation exists among both women's and men's assessments.Professional status has approximately the same relationship with assessments of the process, as the variable consists of three categories (low, middle and high).The 'professional' category that distinguished itself markedly from others were retirees, who were much more positive in their procedural assessments (almost a full step on the 0 to 6 scale) than others.The weight of the 'professional' variable disappears, however, once subjectively reported reluctance to speak at the meeting is taken into consideration.Those who felt reluctant to voice opinions and concerns in the process did feel more critical of the process, even when gender and professional status were controlled for, and regardless of whether they belonged to the more critical group that had attended both sets of meetings.The mean of the procedural assessments of these individuals was a half step (-0.45) lower than other respondents'.Interestingly enough, the effect disappears when people's trust for other important political institutions (the government) is taken into account (not shown in table ), which suggests that individuals with lower levels of political trust are also more reluctant to participate in deliberative processes.More work would be needed to sort out the causal relationships among these factors, but it would certainly be troublesome if those already more skeptical of and withdrawn from political life became even more so as deliberative participation becomes a more common feature in the political process.

How durable is the deliberative participation effect?
This final analysis addresses a somewhat different aspect of perceived procedural fairness and deliberative participation.The preceding analyses provide some evidence, albeit modest, that deliberative participation can yield civic benefits.In order for deliberation to have a net positive effect for the political system, it cannot be the case that only individuals whose prior preferences concur with the decision outcome feel the process was fair.If those who favored a different outcome are disenchanted with the process as a whole once the outcome becomes known, then deliberative participation will not enhance to the aggregate perceived legitimacy of the political system in the long term.
This analysis examines whether the apparent positive effects of deliberative participation withstand the announcement of a milestone decision in the tunnel project.In the middle of the survey period, the Environmental Court granted permission to the Rail Administration to release up to 100 liters/second of groundwater during the construction work.Though the decision was not the result of the participatory process itself, it was one of the final remaining administrative hurdles to continued construction work.Almost exactly half (51%) the respondents responded after the announcement of the decision.The model presented in Table 5 also includes several other variables that may affect public justification assessments in order to more accurately assess the impact of this important decision announcement.As the results in Table 5 indicate, the procedural assessments of the deliberative process did indeed react to the announcement of the ruling from the Environmental Court.Respondents who sent in their surveys after the announcement of the decision were considerably more critical (0.4 of a step on the 0 to 6 scale, controlling for other factors) of the deliberative process than those who responded before the announcement of the decision.To the extent that this difference prevails over time, this result is cause for concern.Recall from Table 2 that those respondents who had attended one or more of the deliberative meetings had mean procedural assessments that were 0.5 points higher than those who had not attended meetings.In other words, whatever positive effect the deliberative proceedings seems to have had on the perceived procedural fairness of this decision making process may have been undone by the announcement an unwelcome decision.
Because the Environmental Court had announced well ahead of time that the ruling would be made public on the designated date, it is not impossible that more critical residents waited to hear the ruling before responding to the survey.The multivariate model therefore controls for other factors that affect public justification assessments.Being directly affected by the tunnel, as well as satisfaction with the Rail Administration's reparations of the negative externalities have a strong bearing on the perceived fairness of the public consultation meetings. 13Including these experiences, as well as other factors shown to affect assessments of public justification does not, however, diminish the gap between those who answered before and after the handing down of the ruling.

Conclusion
The verdict on deliberative participation is, as the reader has by now understood, mixed.On the one hand, people who attended the public consultation meetings evaluated the overseeing authority in a more positive light than individuals who had not attended such meetings.The comparatively positive procedural assessments of the process, despite the problem-ridden history of the issue, indicate that deliberative participation may indeed have the potential to make the policy process and perhaps even the political system, more legitimate in the eyes of its citizenry.
Somewhat surprisingly in light of normative theories of deliberative democracy, the extent to which the meetings measured up to normative stipulations, and in particular the extent to which the deliberation was characterized by an atmosphere of mutual respect, did not have a notable impact 13 Though the effect of satisfaction with measures to address the toxic spill problem and the water supply problem are empirically similar, they require different interpretations.The water supply issue has been one of the main items of discussion at the public consultation meetings.Satisfaction with measures to address the water supply problem might, in theory, be the result of satisfaction with the consultation meetings.In contrast, the settlement of the claims of damages resulting from the release of toxic chemicals occurred long before the first consultation meeting, and was largely negotiated with households individually and mediated by lawyers.Dissatisfaction with compensation for damages must therefore have caused more critical assessments of the deliberative process.on participants' evaluations of the deliberative process.The finding that people who attended the more open and deliberative meetings were not more positive to the process than those who attended the more hierarchical meetings does cast doubt on the claims that mutual respect constitutes a central legitimating force in deliberative democracy. 14It may be that the deliberations examined in this analysis were across the board sufficiently complete that the null result reflects a ceiling effect.At the very least, the result draws attention to the need for more research on these issues.
More worrisome for deliberative democracy theory, those residents who had been most diligent in attending meetings tended also to be more critical of the process.Can it be that prolonged deliberation in an issue without a significant degree of power sharing can have negative effects on legitimacy?A more enduring and intense involvement in an issue presents more opportunities to detect inconsistencies.In this case, an in-depth familiarity with the technical, political and legal aspects of the tunnel may have enabled citizens to spot when important items were being withheld from the deliberative forum and instead being decided behind closed doors.Highly involved participants may conceivably even have interpreted the procedural improvements made in the second set of meetings merely as an attempt on the part of the Rail Administration to ingratiate itself with local residents.
The final two sets of analyses presented here offer perhaps the most serious warning signals regarding the promised civic goods of deliberation.First, individuals who felt disinclined to participate actively in the deliberations were considerably less enthusiastic about the process, suggesting that the deliberative format does not have the same legitimating force across the board in a population.Moreover, finding out the result of a key decision regarding the future of the tunnel seems to have had a considerable deleterious effect on the perceived fairness of the deliberative process.Again, these findings, while far from incontrovertible, point to a need for more penetrating investigations of the durability and the universality of promised beneficial effects of deliberative democracy.
Reluctance to speak during meetings (N=76) 1 = individuals who have wanted to speak at the public consultation meetings but abstained for one of the following reasons: perception that one's opinion differed from others', reluctance to speak in front of others, the perception that one lacked the requisite knowledge.
Responded after Environmental Court decision (N=189) 1 = responded after the announcement of a key decision from the Environmental Court regarding a permit to allow a certain volume of water to flow from the tunnel during the construction period.This decision represented the final administrative hurdle for construction to recommence.

Affected by groundwater problem (N=98)
Affected by toxic leakage (N=43) Satisfied with measures to solve water supply problem (N=16) Satisfied with measures to compensate for toxic leakage damages (N=44) Other negative consequences (N=44) 1 = worry about nature, water supply and property values; lost property through expropriation; bothered by noise and road traffic Benefits from the tunnel project (N=40) 1 = receive municipal water; live near the existing tracks which eventually will be moved.

Table 1 .
Overview of empirical analyses of procedural stipulations.

Table 3 .
Assessments of public justification by meeting attended (raw mean scores and estimated marginal means controlling for covariates).

Table 4 .
The effects of departures from the ideal of political equality on perceived procedural fairness.
Note: Dependent variable: Assessments of the fairness of the public consultation meetings.With the exception of professional status, all of the independent variables are dichotomous.Professional status had three categories: low-skilled, high-skilled and professional (see Appendix).**p < 0.05, ***p < 0.01

Table 5 .
Factors potentially affecting the perceived fairness of the deliberative participation (OLS unstandardized coefficients).