Democratically Sustainable Local Development? The Outcomes of Mixed Deliberation on a Municipal Merger on Participants’ Social Trust, Political Trust, and Political Efﬁcacy

: Municipal mergers are typically contentious and polarizing issues among both citizens and politicians. In deciding on these, municipal-level referendums are often commissioned by municipal councils. Referendums, though, are also per se polarizing processes that only exacerbate an already polarizing issue. Adding deliberation to referendum processes has been shown in previous studies to be a more democratically sustainable process than mere referendums. In this study, we explore the use of mixed deliberation between citizens and politicians within a municipal merger process in the municipality of Korsholm in Finland, one year before a referendum on the issue occurred. The deliberations were two-hour sessions in February 2018, with local politicians present in each discussion group. Using pre- and post-deliberation surveys, we trace how citizens (n = 117) engaging in deliberation developed their social trust, political trust, and political efﬁcacy during deliberation. Generally, we expected that all of these would be strengthened in deliberation. The results, however, reveal only a few statistically signiﬁcant effects, some of which ran contrary to expectations.


Introduction
When it comes to local administrative development, municipal mergers are arguably one of the broadest reforms (e.g., [1]). This is primarily because municipal mergers have several economic and fiscal effects [2,3] and give rise to redistribution of resources between services or groups of the population (e.g., [4,5]). Furthermore, mergers often alter political representation and power relationships [6]. Even though they are complex, municipal mergers have been one of the most common types of local development reforms around the developed world in recent decades [7]. In the Nordic countries, several waves of large-scale municipal mergers have taken place during this period [8]. In Finland, a new law was adopted in 1990 according to which advisory referendums could be organized also on municipal policies. Since this legal reform, albeit that the municipal councils retain the final decision, many proposed municipal mergers in Finland have included an advisory referendum as a key component in informing the decision whether to merge or not [9]. The high occurrence of municipal mergers in Finland has resulted in the number of municipalities being reduced from just under 600 to around 300 since the 1950s [10]. Furthermore, the salience of municipal mergers can also be seen in the fact that 89% of all municipal referendums held in Finland have been about a proposed municipal merger [11].
While the inclusion of a referendum induces a participatory element in the decisionmaking process and thus creates an opportunity for citizens to have their say on a local development issue that affects them, referendums-and especially concerning municipal mergers-are not without their challenges. Firstly, given the inherent complexity of municipal mergers, the pro or con setup of a referendum greatly simplifies the issue per se [12]. The campaigning efforts by proponents and opponents of a merger likewise focus on rallying supporters and spreading soundbite information (e.g., [13]). Thus, citizens' learning, perspective-taking, and deeper reflection on the merger issue are rare in referendum campaigns [14]. Consequently, referendums are particularly prone not only to polarization on the issue but also to affective polarization whereby prejudices and distrust of the people on the "other side" are common (see for example [1,12,15,16]). Over time, distrust and negative prejudices of the people on the "other side" might further develop into more socially ingrained forms of polarization [17]. Secondly, municipal mergers often induce a strong element of emotionality as citizens tend to see their municipality as symbols for deeply rooted identities [18]. Consequently, while participatory components can generally be argued as being beneficial to citizens' perceived legitimacy of decisions (e.g., [19]), one can raise serious questions regarding whether referendums in conjunction with municipal mergers are de facto democratically sustainable processes or whether they exacerbate tensions between opinion camps on the issue and reduce meaningful reflection on the complexities of a municipal merger.
Considering these potential problems with referendums, scholars have contemplated the use of deliberative elements in conjunction with referendums to achieve a deeper reflection and more perspective-taking on the issue put to a referendum, even across polarizing divides (see [20][21][22][23]). Furthermore, the use of a deliberative Citizens' Jury providing an impartial source of information on a ballot initiative to the broader public, has been tested with promising results [24,25]; see also [14]). Thus, these types of deliberative bodies can increase citizens' capacity for reflective judgment, making them more confident about their ability to make an informed decision about the issue [25] and increase their trust in other citizens and political institutions (e.g., [26]), including those whose opinion differ from their own [27,28]. As these types of public sentiments are considered essential elements for the functioning of democracy (see [29,30]), fostering these is of vital importance for the long-term sustainability of democracy.
As Hammond [31] explains, the rise of the sustainable development paradigm has been followed due to a concern for problems with the current ways of democratic participation. According to The United Nations Agenda 21 on sustainable development, a prerequisite for sustainable development is broad public participation in decision-making [31]. Representative democracies seem to be challenged by phenomena (e.g., lack of political and social trust, lack of citizen engagement, short-term interests) that contribute to an unsustainable political system [31,32], (p. 77). It is argued that a transition to a sustainable society needs new forms of citizen participation and inclusive decision-making processes, otherwise political decisions might be considered illegitimate [33]. Deliberation is seen as a promising instrument to enhance the legitimacy of governance [31] and we regard the democratic intent of deliberative processes to be one way of transitioning society into sustainability as one of the goals of sustainable development is to "ensure responsive, inclusive, participatory and representative decision-making at all levels" [34]. One way of doing this is to explore the potential of mixed deliberation to stimulate attitudes that are pro-social and create participants that hold democratically oriented preferences that are necessary for a sustainability transition [35]. In our view, deliberative processes that have the potential to increase citizens' levels of political efficacy as well as political and social trust can contribute to democratic sustainability.
In this article, we study mixed deliberations-deliberations with both citizens and local politicians at the same table-on a proposed municipal merger in Finland one year before a referendum was held on the issue. Specifically, we focus on whether deliberation affected participants' social trust, political trust, and internal-and external political efficacy. Thus, on a general level, we seek to increase the knowledge about whether deliberative innovations can help create more democratically sustainable processes under testing circumstances such as referendum campaigns on a polarizing issue. Therefore, the article serves to further develop deliberative theory and especially relates to its current focus on bringing deliberation into real-world use (e.g., [36]). Hence, the overarching research question in this study is what kind of effects on political efficacy, political trust, and social trust can be related to participation in mixed deliberation.

Mixed Deliberation
Mixed deliberation refers to deliberative events where citizens and elected representatives meet to discuss political issues in small groups [37,38]. Usually, mixed deliberation entails small group discussion where citizens are in majority and politicians in minority (e.g., [38,39]). Variants of mixed deliberation have been called "mixed-membership deliberation" (e.g., [38]), "directly representative democracy" [40], "mixed assemblies" [39], "mixed discursive spheres" [41] (p. 8) or "mini-publics as collaborative institutions" [42]. Mixed deliberation as a form of hybrid model between citizen and elite deliberation has been tested in Ireland, the UK, Belgium, Finland, and Italy [43][44][45]. Mixed deliberation has thus far received far less attention than other forms of deliberation (see [46]). Criticism against the benefits of deliberation has pointed out that the policy impact of deliberative events has often been weak [47,48]. When deliberation has no connection to actual decision-making, events might not be taken as seriously as when a connection exists [38] (p. 2). Therefore, organizers of deliberation have started inviting politicians to participate in deliberative events to close the gap between deliberative and representative decisionmaking processes (e.g., [38]). The inclusion of politicians in deliberation is thought to strengthen the commitment of politicians to suggestions made as a result of deliberation and should motivate decision-makers to promote these suggestions in formal decisionmaking [43] (p. 28). Citizens' willingness to participate in deliberation is dependent on the assumed political impact of the event, as they might not participate if there is no potential impact [49,50]. Moreover, mixed deliberation can reduce the distance between citizens and their representatives [38,51]. For example, Suiter et al. [52] found that 75% of the citizen members of the Irish Constitutional Convention agreed that their attitude towards politicians improved as a result of the Convention. A study on mixed deliberation in Finland reported mostly positive findings from the viewpoint of politicians as well [43]. However, politicians remain skeptical about the value of deliberative citizen bodies in general [53] and prefer these to have a consultative/advisory role [54].
There is some skepticism regarding the inclusion of politicians in deliberation. There is, for instance, a fear that politicians might dominate the discussions and bring inequality to the deliberation table [39,55]. This, in turn, can further solidify negative sentiments towards politics among citizens instead of producing positive effects on participants, such as higher levels of political efficacy [56,57]. Nevertheless, Farrell et al. [38] found no evidence of politicians dominating discussions in mixed deliberation and concluded that concerns over intellectual domination and imbalances of power were unfounded in the Irish case. Other studies have concluded that citizens tend to gain insight into politics and appreciate the opportunity to meet politicians in mixed deliberation regardless of its impact on actual policy [58] (p. 189) [59] (p. 633). Regardless, Farrell et al. [38] recommend that politicians should, just like lay citizen members, be selected randomly to deliberative forums. In contrast, Flinders et al. [39] found that the presence of politicians had a negative effect on deliberative quality. Moreover, including politicians in deliberation did not have any effects on citizens' attitudes towards politicians and political institutions. Likewise, citizens' levels of internal and external efficacy did not change more in a mixed deliberation setting compared to a citizens-only setting. Attitudes regarding the presence of politicians were mixed, although most participants found it positive. Flinders et al. [39] (p. 42) speculate that the type of politicians participating may have bearing on the process and effects of deliberation in the sense that "more high-profile politicians", compared to local politicians, would yield more positive results since it signals higher importance of the deliberation. Experiments with mixed deliberation featuring US Congressmen, also imply that the profile of the politicians matter when it comes to citizen satisfaction with the process [40]. However, because Gerber and Mueller [60] found gaps between lay citizens and mostly part-time politicians regarding participation and deliberative quality, gaps between ordinary citizens and professional politicians may be even larger. Moreover, the length of deliberative events appears to be important since Flinders et al. [39] (p. 42) found that the citizens' perceptions of the presence of politicians grew more positive over time. Another factor that might have a bearing on the outcome of mixed deliberation is the timing of the event within a decision-making process. Ideally, events should be organized before a political decision has been made, otherwise, citizens might perceive an event as meaningless window dressing by the authorities [37,43] (p. 43).

Political Efficacy and Mixed Deliberation
A commonly cited potential effect that deliberations have on individuals taking part in deliberative discussions is that it can enhance their levels of political efficacy by improving their competence of democratic participation and their perceptions of the political system [61] (pp. [23][24][25][26] [62]. The terms internal and external political efficacy are often used to pinpoint these two aspects. Thus, internal efficacy is the belief that one is capable of effective political action and self-governance while external efficacy is the belief that governing officials listen to the public and that there are legal ways to influence governing decisions [63]. Both are arguably important components of a well-functioning and sustainable political system. Internal political efficacy might increase among participants in deliberation as they 'practice' taking part in a political process themselves [64] (p. 358) [26] (p. 98). This mechanism might also be especially relevant in mixed deliberations since the presence of politicians might make this 'practice' even more real and citizens may directly see that they are equal to politicians in the process. However, Mutz [65] (p. 358) argues that being challenged and confronted with opposing views may cause doubts in one's position and beliefs. Since politicians in mixed deliberations often have a higher deliberative capacity, i.e., prior knowledge and experience in argumentation, than citizens (e.g., [37]), it is plausible that opposing, well argued, views from politicians causes citizens to doubt their own capacity and even reconsider previously held assumptions.
In the context of deliberations on complex issues such as a municipal merger, it is perceivable that as citizens deliberate and encounter more opinions about a given topic, they can also be expected to become better equipped to ponder various aspects of a merger and feel that they have a better grasp of the issue. Gastil [21] (p. 156) also states that deliberation can help voters see beyond simplistic arguments put forward in referendum campaign rhetoric and to ponder various aspects of the issue. Findings also show that citizens become more confident about their ability to make an informed decision about the issue [25]. Thus, their level of internal efficacy would be augmented. In mixed deliberations specifically, there is also strong reason to expect that especially citizens' external efficacy is boosted since they interact in deliberation with politicians and get 'proof,' so-to-speak, of the officials listening to the public. Grönlund et al. [26] (p. 99) remark that external efficacy might increase because participants realize the complexity of political issues and develop an understanding of the actors and processes of representative democracy. Similar observations have been made when citizens' juries have been used in conjunction with referendums (e.g., [66]). Setälä [51] also discusses that mixed deliberation in particular gives citizens an opportunity to engage with politicians, which signals a greater importance of the process and can help to increase their external efficacy ( [51]; see also e.g., [56,67]). Tentatively, this might be due to the simple fact that the politicians can explain things hands on in the deliberations. On the other hand, it is entirely possible that mixed deliberations reveal that politicians do not live up to expectations, which could result in lower external efficacy. Flinders et al. [39], for instance, found no effects on efficacy in their study of mixed deliberation. Nevertheless, Geissel and Hess [56] argue that improvements in the level of political efficacy is more likely to occur in municipalities with an institutional commitment to deliberative procedures.
In light of the discussion in this section, we generally expect participation in mixed deliberation to create positive effects on both forms of political efficacy among the participants. We thus formulate our initial hypotheses accordingly: Hypothesis 1a: Participation in a mixed deliberation contributes to higher levels of internal efficacy among the participants.
Hypothesis 1b: Participation in a mixed deliberation contributes to higher levels of external efficacy among the participants.

Social and Political Trust in Mixed Deliberation
Social trust reflects a belief that in a worst-case scenario others will not willingly do harm and at best even work in one's interests [68] (p. 202) and is derived from an assessment of whether an individual or a group of individuals can be relied upon [69] (p. 651). High levels of social trust are important for the social cohesion in any given political community, as it helps with creating a social environment where democracy can prosper (see [30]). Social trust therefore functions as a form of generalized trust that is expected to link people together and contribute to forming more egalitarian political communities [70] (p. 45). Hence, high levels of social trust are considered essential for a well-functioning political system and democratic stability [71,72]. Following a merger process, a process putting citizens against each other based on preferences on the merger issue, it is not farfetched to expect it to have negative effects on the social trust within the political community. Political trust entails the citizens' confidence that political institutions and procedures function according to positive expectations [29,73]. While representative democracies arguably require a certain level of healthy distrust, or 'warranted trust' [74], the level of political trust can be regarded as one of the key measures of the performance of democracy and the quality of government. Therefore, political trust has come to be perceived as "the glue that keeps the system together, while simultaneously making the political system work" [73] (p. 76). Although the reasons for the decreasing levels of political trust in many liberal democracies remain contested, and there are concerns that growing distrust can undermine the public will for cooperation and support for democratic institutions [73]. The polarizing nature of both municipal mergers and referendums are arguably especially detrimental for political trust. Thus, the dichotomous nature of the referendum in conjunction with the contentious nature of merger issues divides voters into yes and no camps, fostering prejudices and negative perceptions of both actors and institutions on the "other side". Scholars have argued-e.g., [27] (p. 365); see also [75]-that polarizing issues can transform political trust into a partisan form of sentiment, in the sense that citizens only trust political actors and institutions that they perceive as being on their side.
That deliberation increases trust in fellow citizens (e.g., [76]) and trust in certain political actors-see [26] (p. 99)-is well-established by now. The former typically as a result of citizens sitting at the same table engaged in rational, polite, and sincere discussion with each other and the latter often as a spin-off effect of deliberators learning and understanding political processes better. Some (e.g., [77]) have even argued that restoring trust in political institutions should be the primary role of deliberation. In mixed deliberations, it might be argued that both trust-increasing mechanisms could be strengthened. Scholars (e.g., [58,59]) have shown that citizens tend to have positive views on opportunities to meet with politicians in deliberation and one would expect that citizens trust the politicians more as individuals after having engaged in deliberation with them. This would essentially be a form of generalized trust being strengthened through deliberation where some of the persons 'happen to be' politicians. However, mixed deliberation most likely strengthens the learning and understanding path to higher political trust more than the inter-personal path. Thus, when citizen meet politicians face-to-face in deliberations, the politicians have a higher pre-deliberation knowledge and familiarity of the issue at hand [37,78]. They can thus explain facts, reasoning, and various viewpoints on an issue in a detailed and insightful manner and convey a picture of being trustworthy at dealing with the issue. Consequently, a mixed deliberation setting might improve the knowledge about the given topic as well as the image of the other participants, including politicians and fellow citizens.
If we then, finally, consider complex and emotionally laden issues in relation to these trust-enhancing paths in mixed deliberation, some potential caveats are evident. First, the highly salient nature of the issue and its polarizing nature affects the mindsets with which people enter deliberation. Thus, the more a person cares about an issue prior to deliberation, the less willing he/she is to consider other citizens' and politicians' views during deliberation [79][80][81]. This would hamper any potential positive effects on both types of trust. Secondly, as discussed earlier, political trust tends to become partisan in the context of referendums and it is uncertain whether mixed deliberation manages to remove partisan biases in trust. Nevertheless, Warren and Gastil [27] and Lafont [28] have argued that deliberative bodies in conjunction with referendums enhance factual learning as well as the understanding of viewpoints and a variety of arguments, including those contrary to one's own and that this is expected to increase trust in actors with opposite viewpoints.
Based on this overview, we expect both forms of trust to increase following a mixed deliberation. We hence formulate hypotheses accordingly: Hypothesis 2a: Participation in a mixed deliberation contributes to higher levels of social trust among the participants.
Hypothesis 2b: Participation in a mixed deliberation contributes to higher levels of political trust among the participants.

The Context
In 2017, the predominantly rural municipality of Korsholm in Western Finland decided to start negotiating a possible municipal merger with the neighboring urban municipality, the city of Vasa. At the same time, the municipal council in Korsholm also made a decision-in-principle to organize a non-binding referendum on the proposed merger. The consultative referendum in Korsholm was eventually organized in March 2019. In the referendum, a clear majority, 61.3%, voted against the merger, 36.8% were in favor of it, and 1.9% chose the third option on the ballot to not choose either side (turnout 76.4%).
Already from the outset in 2017, the debate on the merger issue was very heated both among the public and even among local politicians, both internally within and between the political parties. Proponents argued that the merger was necessary in order to sustain the vitality of both Korsholm and the region, while the main fear among the opponents was that it would considerably weaken the position of the Swedish language group in the region. Signs of affective polarization were evident whereby opponents of the merger were seen as backwards striving and too emotional by proponents of the merger, whereas the proponents themselves were criticized by opponents for being dreamers and not showing any concrete evidence of positive impacts that the merger would yield for Korsholm's residents.
Opinions on the issue itself were also very divided. A survey sent to all adult citizens in Korsholm (n = 6686), commissioned by the municipality in 2018, measured opinions about the merger on a scale between 0 and 10 (0 being entirely against and 10 entirely for). The mean opinion on that scale was 4.81, but the standard deviation was 4.02, which means that the opinions were heavily distributed towards either extreme of the opinion scale (see [1]). The key aspect of the merger was that it would turn the Swedish-speaking majority in Korsholm into a minority in a merged municipality. The merger was thus at the core of a long-standing sensitive language debate in bilingual Finland. Accordingly, the public views on the merger were highly segregated along language divisions [1]. The above-mentioned survey shows that the mean opinion was only 3.71 among Swedish speakers, whereas it was 7.38 among Finnish speakers. In fact, separate analyses show that the strongest single factor explaining why citizens in Korsholm were in favor of the merger was being a Finnish speaker (see [1] for analyses).
Over time, polarization on the merger issue became even more affective. The survey conducted in 2018 for instance showed that people only believed arguments supporting their own view on the merger and dismissed opposing arguments entirely. Strong affective polarization can be further confirmed by instances of harassment and threat during the campaign [82]. The merger issue even divided the locally dominant Swedish People's Party in Finland (SPP), which is supported by a vast majority of voters in Korsholm. The leading politicians on both sides of the merger campaign represented this party, and in this respect, divisions on the merger issue did not follow partisan alignments. In addition to the issue of linguistic identity, the planned municipal merger would have potentially had significant consequences on public services and local democracy, which made it even more of a salient issue among the public. The issue of municipal merger was furthermore sensitive because a large share of the population was employed by the municipality, and the merger could have posed a threat to their jobs in the long run (in a 2018 survey, 11% of respondents indicated that they were employed by the municipality).

Design of Study
For the purposes of this study, we conducted a series of three deliberative discussions together with the municipality of Korsholm in the winter of 2018, which was one year after the initial decision to negotiate a merger had been made and roughly one year prior to the referendum on the merger. The deliberative discussions were part of the official municipal hearings of citizens prior to negotiations about the merger with Vaasa. The deliberations were conducted as mixed groups containing both citizens and politicians with 225 participants in total. For the purposes of this study, however, we only analyze the citizens taking part (n = 117), since several of our dependent variables measure opinions on politicians and it thus makes no sense to have politicians assess themselves. Due to missing values for some of our dependent variables, the actual number of analyzed citizens is 105. The details of the deliberative discussions are described next.

Procedure and Participants of the Deliberative Discussions
Since the deliberative discussions were part of the municipality's official hearings of the public, we did not enforce any experimental treatments in the deliberative discussions so all of them were thus conducted in small-n groups with facilitation and discussion rules. However, we used a pre-test post-test experimental design in the small-n deliberation groups. Participants thus answered a pre-test questionnaire prior to deliberation (abbreviated to T1 henceforth) and immediately after deliberating (T2 henceforth). Besides basic demographics, these surveys contained items on internal and external efficacy, and political and social trust (see Appendix A for details). By studying pre-test to post-test changes regarding political efficacy and political-and social trust, we can test our hypotheses.
The deliberative discussions were held on 31 January (92 participants divided into 8 discussion groups), 13 February (94 participants, 10 groups) and 20 February (39 participants, 4 groups) in 2018. Since the municipality is geographically vast, these events were held in different parts of the municipality to give all of the municipality's citizens a chance of attending an event nearby. Each event was thus open to all interested; there were no random samples of invited citizens. This resulted in some overlap between the events so that some citizens attended several events and likewise for the politicians. The lack of random sampling also resulted in the sample of citizens deliberating being skewed compared to the population in general (see Table 1 below). Thus, there was an overrepresentation of men, Swedish-speakers, older citizens and those with secondary level education. Furthermore, opinions on the proposed merger were even more negative among the deliberating citizens (average 2.87) than among the public (4.87). The template was exactly the same for each deliberation event: after a brief welcoming speech and information about the process of municipal negotiations of the merger held by the municipality's chief executive officer, participants at each of these three events were randomly allocated to the small-n discussion groups (event one: 8 groups, event two: 10 groups, and event three: 4 groups). We used stratified randomization to ensure that each small-n group contained a mix of politicians and citizens [39,51,83]. The typical ratio between citizens and politicians was 1:3 in each discussion group, as was the ratio of women to men. Each small-n group had an assigned facilitator and participants were initially given time to answer the T1 survey, to read a brief information package about the merger, and to be acquainted with the rules of discussions. Discussions lasted for about two hours, after which the groups summarized the most important aspects brought up in discussion regarding the proposed municipal merger and each participant then answered the T2 survey. The rules and tasks for the facilitators were designed to steer the deliberation processes as close to the normative ideals of deliberation discussion as possible (e.g., [84]). The rules essentially supported the ideals of reasoned justifications, reflection, sincerity, and respect, whereas the facilitator made efforts to ensure reciprocity, inclusion, and equality of discussion. The facilitators were graduate political science students who were trained for the task in two sessions. Some of them also had previous experience of facilitating deliberative discussions.

Measures
All three of the dependent variables, social trust, political efficacy, and political trust, were measured using standard survey items commonly used in survey research on trust and efficacy such as in the European Social Survey, World Values Survey etc. (see also [63,[85][86][87][88]). Social trust was thus measured by asking respondents, on a scale from 0 to 10, with 10 meaning that they feel most people can be trusted, whether they feel that most people can be trusted, or that one can never be too careful when dealing with people. Political efficacy was measured for internal and external efficacy as both summarized efficacy scales and as single survey items. We also chose to study single efficacy items in order to explore in detail which aspects of internal-and external efficacy that taking part in mixed deliberation potentially affected. Political trust was measured by asking respondents to assess their trust in various actors on a scale from 0 to 10, where 10 means that they fully trust the actor in question (see Appendix A for exact questionnaire wording).

Results
The results are organized according to the two sub-aspects of interest: political efficacy followed by social and political trust. All analyses are carried out using paired samples t-tests of pre-deliberation and post-deliberation values for each dependent variable. Significance tests are one-tailed since our hypotheses have directional expectations. Where significant effects are found, we ran post-hoc regressions predicting the post-test value of the dependent variables. This so-called regressor strategy includes the pre-test measure of the dependent variable as covariate in the model in order to correct for potential regression to the mean effects (e.g., [89]). The other variables included in the post-hoc regressions were gender, age, education level, and a measure of how each citizen experienced the deliberative quality of the discussion. Table 2 below displays the findings on how internal and external efficacy developed during deliberation. Table 2. Development of internal and external efficacy during the mixed deliberations (n = 105).

Pre-Test Post-Test Change
Internal efficacy scale (0-3) −0.04 † p < 0.10 one-sided tests. * p < 0.05. ** p < 0.01. Note: The efficacy scales have been recoded so that a higher value corresponds to higher level of efficacy. The single items range between 0 and 3. For the single efficacy items. Higher values for internal and external efficacy mean that the respondent agreed more with the statement.
The findings on political efficacy provide little support for our hypotheses. There were only two rather weak statistically significant findings. Firstly, people, after deliberation, felt less that they understood politics better than most people. This is contrary to H1a, but is potentially tied to the specific topic of municipal mergers. Thus, as citizens engage in deliberation about the merger with politicians, they come to understand the complexity of the issue more and, perhaps, realize that their knowledge of politics is rather limited. The post-hoc regression of this effect, however, suggests that this change is partly also due to regression to the mean. Secondly, the participants increased their agreement with the statement that people can exert influence through voting. This finding lends some support to H1b, albeit it is not easy to decipher why deliberation increases this specific item on external efficacy. Further analyses into the actual content of what was said during deliberation would potentially shed more light on this effect and the causes for it. Again, the post-hoc analysis suggest that regression to the mean is at play here. There was also a statistical trend whereby people agreed less with the statement that politics is too complicated to understand. Here the post-hoc regression showed that, besides regression to the mean, women had higher post-test level of efficacy after deliberation.
We now turn to our findings concerning social-and political trust. Table 3 below displays the findings on how those opinions developed during deliberation. Table 3. Development of social trust and political trust during the mixed deliberations (n = 105).

Pre-Test Post-Test Change
Social trust (0-10) The findings regarding social-and political trust are mostly insignificant and the two statistical trends both run against our hypotheses and the post-hoc analyses suggest that regression to the mean is a factor behind the change as well. Thus, citizens trusted municipal politicians and the municipal council less after deliberation. A potential reason for this could be the fact that politicians' opinions on the merger issue were less critical (average 4.67 on 0 to 10 scale) which resulted in rather heated clashes of opinions in some deliberating groups (based on our annotations from the events). We argued earlier that the fact that people had very strong opinions on the issue before deliberation might hamper how deliberation fosters trust [79][80][81], which most likely plays some part in understanding these findings too.

Discussion
This study set out to explore whether participating in mixed deliberation, in the context of a prospective referendum on a polarizing municipal merger issue, appears to augment citizens' social-and political trust as well as their feelings of internal-and external efficacy. The underlying argument of the paper was that putting contentious issues to deliberation would be a more democratically sustainable process than only putting the issue to a referendum. While democratic sustainability is a multifaceted concept that arguably takes a long time to develop, strengthening trust and belief in the political system and in one's own role in it are arguably important pieces in starting such a development. Generally, however, the results of this study provide little proof of the deliberating participants strengthening any of the aspects in focus here. Most of the sentiments did not change at all during deliberation and the few statistically significant findings were inconsistent with regards to our hypotheses. We presented some potential interpretations to these findings already in the results section, but what do these more-or-less null findings tell us about the democratic sustainability of mixed deliberation in conjunction with a referendum on a polarizing issue?
First and foremost, it should be remarked that although we studied a rather critical case and found few effects of deliberation-which usually is akin to a dead end in terms of generalization-similar findings to ours have been found (e.g., [56]) when studying practices of participatory budgeting in a plethora of different circumstances. While participatory budgeting is not entirely the same as our case of mixed deliberation, we argue that our findings nevertheless appear to fit into a general pattern. Continuing, then, to discussing the key findings of our study, the most important finding is that this study has shown that mixed deliberation is not a universal quick fix for building trust and belief in the political system and its actors. Under circumstances of severe polarization, such as in the case studied here, it might even serve to exacerbate the situation. Nevertheless, as others have demonstrated [39] (p. 42), mixed deliberation matures over time in the sense that citizens' opinions on the politicians present in deliberation gradually become more positive after several deliberation sessions. Since this study focused on only one session of deliberation between citizens and politicians, a session which also was the first time these met in deliberation, there is a possibility that more positive sentiments would have developed among the deliberating citizens if the deliberations had spanned several sessions. Button and Mattson [59] (p. 612) have argued that deliberation serves various roles in various context; provides an arena for venting anger and frustration, creates learning on an issue, strengthens mutual understanding, and (rarely) has a direct legislative impact. None of these, however, usually occur all at once nor at every deliberative event. Besides a direct legislative impact, the deliberations studied here contained elements of all these aspects but, considering our observations at the events, the primary focus tended to be on citizens ventilating anger and frustration and politicians needing to be defensive. Again, if the deliberations had spanned several events and a longer time period, it is perceivable that stronger elements of learning and mutual understanding would have matured among the deliberators. These, admittedly speculative, sentiments of ours point to something that others have reiterated several times (e.g., [41,56,90]), that one-off isolated deliberative events are rarely successful. The key to a sustainable democracy is rather to achieve a system-level culture of deliberation where deliberative spaces are abundant and clearly tied to the political institutions and processes: "If a government aims at conducting a successful deliberative procedure, it should allocate adequate funds and infrastructure. Special staff for citizens' participation and a local participatory plan are crucial for the success. If the local government plans a 'lowbudget' procedure without financial and institutional commitment, the endeavor will most likely not produce any effects" [56] (p. 15).
Besides the systemic perspective on mixed deliberation, the work with gathering data for this study and observing the deliberations in action has highlighted the importance of micro-factors in creating fruitful deliberations. Thus, the importance of the facilitator in keeping the deliberation session on track appears even more crucial in mixed deliberation on contentious issues. As Gerber [91] (p. 114) has pointed out, facilitators need to be subtle but ensure that basic civility and order is maintained in the discussion and that everyone gets the opportunity to speak. Maintaining civility is especially hard when participants enter deliberation with strong emotional investment in the topic of deliberation. The task of ensuring equal opportunities to speak is also especially tricky in mixed deliberation where politicians might tend to dominate the discussion (see [37]). We observed great variation in how 'deliberative' the sessions were and, besides different participants in the sessions, the facilitator was the only thing that varied between the groups. A common situation was that a woman facilitator struggled to maintain order in discussion groups with a lot of 'old angry men' in them. Future research into optimizing facilitation in mixed deliberation is clearly needed and potential effects of the gender of the facilitator on deliberative quality could be explored. Similarly, unraveling the deliberative grand treatment and studying deliberation in each discussion group-since a deliberative event hides a lot of important variation at the group-level (see e.g., [92])-appears to be an important avenue for future research to pursue in order to gain better understanding of micro factors leading to deliberation working or failing. To illustrate, our sessions contained discussion groups that worked similar to textbook examples of successful deliberation but, at the same time, there were several groups that failed and ended in shouting-matches between angry citizens and politicians. Additional important aspects to study further regarding mixed deliberation, are the impact the type of politicians (local/national) might have on the process and outcomes of deliberation [39] (p. 42). Is there so-to-speak a difference if citizens meet local "low-profile" or national "high-profile" politicians in mixed deliberation?
Our study clearly has limitations. The most obvious one is that we have studied one case only, which we have discussed rather thoroughly already in the discussion. This means that generalization to other contexts is by necessity speculative. Nevertheless, it should be pointed out that case studies' main purpose is more about generalizing to a theory-in this case the theory of deliberation in general and mixed deliberation in specific-and not to a broader population [93] (p. 21). Thus, our findings provide important knowledge on the boundaries of deliberation in polarized contexts and on polarized issues and builds on the ever-increasing pool of knowledge on how deliberation as a theory is best adapted to real-world circumstances (see [36]). Other noteworthy limitations are that participants, both citizens and politicians, in the mixed deliberations studied here were selfselected, which resulted in a skewed representation of the population. Ideally, deliberation should feature randomly selected participants to ensure that panels are representative of the population [38]. Nevertheless, non-representativeness of participants is possible in deliberation events using random recruitment as well [33,43]. Allowing participants to selfselect might also introduce partisan bias [38] and result in the participation of citizens and politicians that are particularly engaged in the deliberation issue. However, not allowing participants to self-select can also result in problems. In the case studied here, it is likely that a randomized selection of participants would have been interpreted as a measure by the municipality to shut out and silence critical citizens or politicians. Moreover, this is not a study of the effects of mixed deliberation since we did not have an experimental design testing differences between mixed and non-mixed deliberative events. Therefore, we cannot make causal claims about whether the results would have been different if the deliberative events had been open to citizens only. Another limitation is that since the deliberations studied here took place rather late in the broader merger negotiation process citizens may have viewed the events as mere 'window-dressing' by the municipality to legitimize decisions that had already been made. The potential external policy impact of the deliberations might thus have been unclear to the citizens taking part, which could have lowered their expectations of the outcome of these events. This could possibly partly explain the lack of change in levels of trust and efficacy.

Conclusions
Generally, in science, as Popper taught us, more is often learned from putting something to a critical test rather than creating optimal conditions for it to work. Deliberation as a field of study could earlier be criticized for a clear positivity bias (see [94]) whereby positive effects of deliberation were found in an array of scientific experiments with near perfect circumstances for deliberation to succeed in. In recent years, as deliberation is rapidly moving from the lab into embedded experiment in real-life, knowledge that is much more crucial has arguably been gained. Within this strand of research, this study explored mixed deliberation embedded in a context of a contentious issue (municipal merger) and a polarizing referendum process. A main conclusion of relevance to the broader discussion and theory of deliberative democracy to draw from this study, as we discussed, is that deliberation is no quick solution to complex and contentious societal issues. Potential positive effects of deliberation were few in our results, and whether these would have occurred long-term is uncertain in light of this study. Nevertheless, our conclusion is that, despite being important, effects in terms of trust and efficacy are not everything. The societal climate in Korsholm was more-or-less unsustainable when the deliberations studied here commenced, so much was to gain and little to lose in terms of democratic sustainability. Simply letting out steam was arguably important at this point and the value per se in citizens and politicians sitting down at the same discussion table should not be neglected either [58] (p. 298) [59] (p. 633). Experiences from the Oregon Citizens Initiative Reviews (e.g., [24,25]) give reason for long-term optimism regarding the potential to achieve democratically sustainable combinations of deliberation and referendums. The key, as we discussed, is to achieve a context-wide culture of deliberation so that deliberation becomes part of the ebb and flow of local democracy instead off extraordinary one-off events. For now, one-off events such as the one studied here, are part of the first thriving steps towards a more democratically sustainable system.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to a contract with the municipality of Korsholm.

Conflicts of Interest:
The authors declare no conflict of interest.

Nr
Question Text Question Alternatives