Building Faith in Democracy: Deliberative Events,Political Trust and Efficacy

Abstract

Governments have turned to public deliberation as a way to engage citizens in governance with the goal of rebuilding faith in government institutions and authority as well as to provide quality inputs into governance. This article offers a systematic analysis of the literature on the effects of deliberative events on participants’ political efficacy and trust. The systematic review contextualizes the results from a 6-day deliberative event. This case study is distinctive in highlighting the long-term impacts on participants’ political trust and efficacy as key outcomes of the deliberative process unfold, that is, City Council receives then responds to the participants’ recommendations report. Using four-wave panel data spanning 2.5 years and three public opinion polls (control groups), the study demonstrates that participants in deliberative events are more efficacious and trusting prior to and after the deliberative event. Despite the case study’s evidence and the systematic review of existing literature, questions remain about whether enhanced opportunities for citizen engagement in governance can ameliorate low levels of political trust and efficacy observed in Western democracies.

Keywords

deliberative democracy political trust political efficacy political participation

Introduction

Democratic institutions are faced with a crisis of legitimacy. This crisis affects government’s ability to create and enforce policies and laws to govern the lives of citizens (Dalton, 2004; Levi and Stoker, 2000; Marien and Hooghe, 2011; Norris, 2011). This crisis is most evident in growing levels of distrust of government institutions and authorities (Fournier et al., 2011; Holbrook, 2004; Levi and Stoker, 2000; Pew Research Center, 2017). While some dispute these trends (Thomassen and Van Ham, 2017), they acknowledge a gap between citizens’ expectations for democratic governance and their satisfaction with the performance of democratic systems (Norris, 2011). In this context, some governments have turned to public deliberation as a way to engage citizens in governance with the goal of rebuilding faith in government institutions and authority (Marien and Kern, 2017; Spada and Ryan, 2017) as well as to provide quality inputs into the decision-making process (French and Laver, 2009). Deliberative events are thought to be one way to deepen citizens’ involvement in governance (Wright, 2010).

Deliberative events are not a panacea for the flaws in representative democracy, and given the small scale of these initiatives, their impacts on the broader political system may be limited (Lafont, 2015; Warren, 2008). One study suggests that hearing about deliberative events increased political efficacy and trust among respondents to a random digit dialling survey (Boulianne, 2018a). Including citizens in the decision-making process may increase the participants’ perceived legitimacy of the democratic system (Grönlund et al., 2010; Halvorsen, 2003; Strandberg and Grönlund, 2012; Warren, 1992). However, this process may also have detrimental effects on legitimacy, if the input is perceived to have no influence on decision-makers (Halvorsen, 2003). In sum, in the short term, participation in a deliberative event could positively impact faith in government institutions, but the long-term impact may depend on decision-makers’ responses to this input.

This article explores the role of political trust and efficacy in a public deliberation held in Edmonton, Canada, in Fall 2012 (n = 55). The distinctiveness of this event and its methodological design are illustrated through a systematic review of existing research on the effects of deliberative events on political trust and efficacy. Using four-wave longitudinal survey data of participants and three cross-sectional surveys (control groups), I examine the following questions: How do political trust and efficacy impact the decision to participate in a deliberative event and how does the deliberative event impact political trust and efficacy? This study is distinctive in highlighting the long-term impacts on participants’ political trust and efficacy as key outcomes of the deliberative process unfold. Over 2.5 years, political trust and efficacy are tracked as the deliberative event concludes, as a recommendations report is presented to City Council (April 2013), and when City Council issues a formal policy response related to the panellists’ recommendations (April 2015). The case study is distinctive in its linkage to the policy-making process and the intensity of the deliberative process (six Saturdays). These distinctive characteristics are critical to understanding how deliberative events impact faith in democratic systems in the short and long term.

Literature Review

Political Trust

I see trust as an issue of whether particular political actors or institutions are expected to act in citizens’ best interests (Holbrook, 2004). Trust is based on expectations about future behaviour, which relate to, but are distinct from, assessments of past behaviour, which are assessed with survey questions about performance, efficiency, and satisfaction (Holbrook, 2004). Political trust is an expectation that government will do ‘what’s right’ (American National Election Studies, 2015; Munno and Nabatchi, 2014; Pew Research Center, 2017). Political trust is a multi-dimensional concept. Individuals may differ in their trust of political actors (e.g. civil servants versus local politicians versus party leaders) compared to institutions. Trust levels may differ based on levels of government as well as the branches of government (Cole et al., 2004; Gallup, 2013; Hetherington and Husser, 2012; Norris, 2011; Pew Research Center, 2013; Pickup et al., 2004). However, these nuances are rarely examined in studies about trust and political participation. In the context of deliberation, I believe that a nuanced measure is appropriate because these events tend to focus on single policy areas, which fall under a specific level and branch of government.

Various studies document a decline in levels of political participation in well-established democracies, particularly around voter turnout (Blais et al., 2004; Norris, 2011; Putnam, 2000). There has also been a simultaneous decline in trust in government (Citrin and Luks, 2001; Levi and Stoker, 2000). As such, scholars have examined whether the two trends are related (see literature summary in Levi and Stoker, 2000).

Some scholars argue that a degree of scepticism, that is, distrust or mistrust, might encourage meaningful public engagement (Lenard, 2008; Parkins and McFarlane, 2015; Warren, 1999). If citizens uncritically trust institutions, they may forgo opportunities to get involved because they assume that the institutions share their interests and will act accordingly (Warren, 1999, 2009) Alternatively, high levels of trust mixed with some scepticism may motivate people to get involved in decision-making processes (Parkins and McFarlane, 2015). Warren (2009) suggests that democratic institutions should organize meaningful public engagement in domains where there is contention and possibly mistrust, rather than try to engage citizens in every decision. This structure would allow citizens to more strategically allocate their time and energy around certain issues, while remaining passive and trusting around other issues (Warren, 2009).

Participation in a public deliberation is expected to increase levels of trust in political institutions and leaders (Gastil et al., 2012; Grönlund et al., 2010; Halvorsen, 2003; Luskin and Fishkin, 2002; Nabatchi, 2012; Strandberg and Grönlund, 2012). Through deliberation, citizens experience firsthand the need for compromises and trade-offs in order to reach a democratic consensus. Simply learning about policy-making processes as part of a deliberative exercise might increase overall trust in institutions (Grönlund et al., 2010). Participants might also learn more about public officials and come to identify with them on the basis of shared personal characteristics (Ulbig, 2007). This exchange might also increase levels of trust.

The deliberative process might also increase faith in policy decisions. By learning about arguments supporting standpoints different from their own, participants might view decisions, which they do not personally favour, to be more legitimate than they otherwise would (Grönlund et al., 2010; Warren, 1992). Additionally, deliberative events might foster a feeling among participants that their perspectives are being taken seriously by decision-makers and that their interests are being fairly represented within the political process (Halvorsen, 2003). All of these processes might contribute to increased confidence in the performance of government as well as an overall perception that government is acting in ways that benefit citizens. In other words, these processes contribute to building trust in government.

While the deliberative event, itself, may have positive effects on trust, these positive effects may not endure over the long term. This intense form of participation requires citizens to invest a considerable amount of time and energy. If the decision-makers react favourably to the deliberative outcomes, for example, reports or deliberative polls, then the impact will continue to be positive. However, if decision-makers respond negatively or are unresponsive to the input, the participants could become critical about their time investment and the deliberative process, causing them to report lower levels of trust and political efficacy (Fournier et al., 2011). For example, Dyck (2009) suggests that direct democracy initiatives or referendums could diminish trust if politicians or bureaucrats fail to fully implement widely supported initiatives. Indeed, he argues that those citizens who are not exposed to direct democracy may be more trusting of government, than those who are exposed but are unsatisfied with the implementation (Dyck, 2009).

Existing research on the role of trust in deliberative events tends to show positive effects, but not consistently (Table 1 exception Tomkins et al., 2010). The studies largely rely on surveys of panellists before and after the deliberative event. The studies treat political trust as an outcome, rather than a motivation for participation, which contradicts how trust is modelled in relation to other forms of engagement (Quintelier and Van Deth, 2014). As a consequence, the studies cannot address whether participants who are more trusting self-select to engage in deliberative events or whether a degree of mistrust is motivating participation. The studies also cannot determine whether the deliberative event is the cause of increased trust or whether the change is caused by some external stimuli. This analysis requires a survey of a control group to coincide with the post-deliberation survey of participants.

Table 1.

Summary of Literature on Political Trust in Deliberative Events.

Author	Duration, mode, topic	Measure	Pre–post design	Exp-control design	Significance at 0.05 level
Luskin and Fishkin (2002)	One weekend, face to face, Australian Referendum	Trust government	No	Yes	Yes
Grönlund et al. (2010)	1 day, face to face, Nuclear power in Finland	How much do you trust the following Finnish institutions? Scale 1–4 (1 = just a little, 4 = a great deal) – The parliament	Yes t1 vs t4	No	Yes
			Yes t1 vs t5	No	Yes
Munno and Nabatchi (2014)	3 days, face to face, Reclaim November Ohio	How much of the time can you trust government to do what is right? Jury 1	Yes	No	Yes
Munno and Nabatchi (2014)	3–4 days, face to face, Reclaim November Ohio	How much of the time can you trust government to do what is right? Jury 2	Yes	No	Yes
Strandberg and Grönlund (2012)	2 hours, online, Energy policy in Finland	How much do you trust the following Finnish institutions? Scale 1–4 (1 = just a little, 4 = a great deal) – The parliament	Yes	No	No
Tomkins et al. (2010)	1 day, face to face, Lincoln’s budgeting priorities	Lincoln City government can usually be trusted to make decisions that are right for the residents as a whole.	Yes	No	Yes

Literature summary is restricted to studies examining institutional trust, which relate to measures used in the current study. For a full literature summary of political trust and deliberation, contact the author. Studies are excluded if they did not report statistical tests.

Additionally, existing research on deliberative events focuses on broad or general measures of institutional trust (e.g., trust in Parliament), whereas scholarship on political trust points to the multiple dimensions of political trust. Deliberative events may only impact political trust in the policy domain under study as well as the level of government involved in the deliberative event. Finally, studies have not examined the long-term impacts of the deliberative process, that is, tracking changes in participants’ views as their recommendations are discussed and acted upon by decision-makers. The single study exploring long-term effects on political trust examines a 3-month time span for a 1-day deliberative event that was not explicitly tied to the policy-making process (Grönlund et al., 2010). As such, this new study makes a key contribution to the study of political trust in the context of deliberative processes and their outcomes. The hypotheses about (a) general trust and (b) policy domain-specific trust are as follows:

H1a (Selection effect): Compared to non-participating citizens (control group), participants report higher levels of general political trust, on average, prior to the deliberative event, using contemporaneous survey data (September 2012).

H2b (Positive immediate effect): Participants will report higher average levels of domain-specific political trust immediately after the deliberative event (December 2012), compared to before the deliberative event (September 2012).

H3a (Positive long-term effect): Participants will report higher average levels of general political trust at 6 months and 2.5 years after the event, compared to before the deliberative event (September 2012).

H3b (Positive long-term effect): Participants will report higher average levels of domain-specific political trust at 6 months and 2.5 years after the event, compared to before the deliberative event (September 2012).

H4a (Deflating effect in the post-deliberation period): Participants’ average levels of general political trust will decrease in the post-deliberative period (6 months versus 2.5 years).

H4b (Deflating effect in the post-deliberation period): Participants’ average levels of domain-specific political trust will decrease in the post-deliberative period (December 2012, June 2013 and June 2015).

H5b (Enduring differences): Compared to non-participating citizens (control group), participants report higher levels of domain-specific political trust, on average, after the deliberative event, using contemporaneous survey data (June 2013).

Political Efficacy

Political efficacy refers to citizens’ perceptions that they can influence government. Efficacy can be measured in terms of individuals’ perception of their own ability to influence political decisions (internal efficacy) as well as their assessment about whether there is opportunity to influence government (external efficacy). While political trust relates to political actors and institutions’ actions in the best interests of citizens, political efficacy refers to citizens’ capacity to influence these actors and institutions. While political efficacy and political engagement are correlated (e.g. Gastil and Xenos, 2010; Smets and Van Ham, 2013), the causal direction remains unclear (Fournier et al., 2011; Gastil and Xenos, 2010). I expect that people who choose to participate in a deliberative exercise will report higher levels of efficacy, on average, than the general public. The opportunity to influence policy decisions is a prime motivation for their involvement in a deliberative event, particularly an event that involves spending six Saturdays discussing climate-change policy. Furthermore, those citizens with high political efficacy may be particularly attracted to this specific deliberative event because of its clear linkages to City government and City Council’s decision-making process.

Only a handful of studies have examined political efficacy as a motivation for participation in deliberative events. Four studies found no differences in political efficacy between participants and control groups (Andersen and Hansen, 2007; Fishkin and Luskin, 1999; French and Laver, 2009; Nabatchi, 2010), while another study found that participants’ political efficacy was higher compared to a control group (Tomkins et al., 2010). However, these deliberative events all involved considerably less of a time commitment than the event in this current case study (6 days). The greater length of this event, as well as its linkage to the policy-making process, might bias the sample in favour of those who feel efficacious.

Deliberative events are expected to increase political efficacy (Carpini et al., 2004; Mutz, 2008; Nabatchi, 2010, 2012). However, whether or not a deliberative exercise will achieve this result depends on contextual issues related to the deliberation, that is, whether it is well-organized, whether it has a clear link to a policy-making process, and the participants’ overall satisfaction with the deliberative process (Gastil et al., 2008; Grönlund et al., 2010; Morrell, 2005). More generally, the effectiveness of these types of events likely depends on perceptions of procedural justice (Marien and Kern, 2017). The effect of participation in deliberative exercises on political efficacy has been tested in a variety of contexts (Table 2). More than half of the effects are positive and significant (Table 2). As observed with political trust, the vast majority of studies track participants over time, rather than comparing participants and control groups. This approach makes it difficult to attribute the changes in political efficacy to the deliberative event and its outcomes, which is particularly problematic when trying to assess the enduring and long-term effects. A handful of studies test for long-term effects (Table 2). These studies do not use a post-deliberation control group to determine whether the public’s level of political efficacy may have changed as well, reflecting the effects of external factors on both the participants and non-participants.

Table 2.

Summary of Literature on Political Efficacy in Deliberative Events.

Author	Duration, Mode, Topic	Measure	Pre–post design	Exp-control design	Significant at 0.05 level
Andersen and Hansen (2007)	One weekend, face to face, Euro	Citizens like you have no say on decisions made by the government and parliament	No	Yes	No
		Citizens like you have no say on decisions made by the government and parliament	Yes t1 vs t2	No	Yes
		Citizens like you have no say on decisions made by the government and parliament	Yes t2 vs t3	No	Yes
		Citizens like you have no say on decisions made by the EU	No	Yes	No
		Citizens like you have no say on decisions made by the EU	Yes t1 vs t2	No	No
		Citizens like you have no say on decisions made by the EU	Yes t2 vs t3	No	No
Fishkin and Luskin (1999)	Weekend, face to face, National Issues Convention	I have no say in what government does (reverse coded)	Yes	No	Yes
		Officials care about what I think	Yes	No	Yes
Fournier et al. (2011)	20 days, face to face, Electoral reform in the Netherlands	MPs do not care about opinions of people like me	No	Yes	No
		MPs do not care about opinions of people like me	Yes	No	Yes
	Three events (British Columbia 26 days; the Netherlands, 20 days; Ontario 24 days), face to face, Electoral Reform	Government does not care what people like me think	No	Yes	Yes
		Government does not care what people like me think	Yes	No	Yes
Grönlund et al. (2010)	1 day, face to face, Nuclear power in Finland	An ordinary citizen cannot influence politics	Yes, t1 vs t2	No	No
			Yes, t1 vs t3	No	No
Luskin and Fishkin (2002)	3 days, face to face, British monarchy	People like me have no say in government	Yes	No	No
Luskin and Fishkin (2002)	3 days, face to face, Australian referendum	People like me have no say in government	Yes	No	Yes
Luskin and Fishkin (2002)	3 days, face to face, National Issues Convention	People like me have no say in government	Yes, t1 vs t2	No	Yes
		People like me have no say in government	Yes, t1 vs t3	No	No
		People like me have no say in government	No	Yes	Yes
		People like me have no say in government	Yes, t1 vs t2	No	Yes
		Public officials care what people like me think	Yes, t1 vs t3	No	Yes
		Public officials care what people like me think	No	Yes	Yes
Munno and Nabatchi (2014)	3 days, Face to Face, Reclaim November Ohio	People like me don’t have any say in government	Yes	No	Yes
		Elected officials don’t care about what people like me think	Yes	No	Yes
Munno and Nabatchi (2014)	3–4 days, face to face, Reclaim November Ohio	People like me don’t have any say in government	Yes	No	No
		Elected officials don’t care about what people like me think	Yes	No	Yes
Nabatchi (2010)	1 day, face to face, Children	Four-item efficacy index including elected officials don’t care what people like me think and people like me don’t have any say about what the government does	Yes, t1 vs t2	No	Yes
			Yes, t2 vs t3	No	No
			Yes, t1 vs t3	No	No
Strandberg and Grönlund (2012)	2 hours, online, Energy policy	An ordinary citizen cannot influence politics	Yes	No	No
Tomkins et al. (2010)	1 day, face to face, Lincoln’s budgeting priorities	Public officials in Lincoln City government care about what people like me think.	Yes	No	Yes
		Residents have a great say in important Lincoln City government decisions.	Yes	No	Yes

EU: European Union; MP: member of parliament.

Literature summary is restricted to studies examining measures of efficacy similar to those used in the current study. For a full literature summary of political efficacy and deliberation, contact the author. Studies are excluded if they did not report statistical tests.

An exception is the Danish Deliberative Poll on the Euro, which involved two control groups and multiple surveys of the panellists, including a survey 3 months after the conclusion of the deliberative event (Andersen and Hansen, 2007). Comparing the control groups during the deliberative poll and panellists after the deliberative poll, the researchers found no difference in the levels of political efficacy. Comparing participants at the time of recruitment, immediately before the deliberative poll, at the end of the deliberative poll and 3 months later, the authors found mixed results about changes in political efficacy (Andersen and Hansen, 2007).

As described with political trust, I expect that if participants’ input is considered by decision-makers, then their sense of political efficacy may be increased (or sustained if already at elevated levels) in the period following the deliberative event. However, if their input is perceived to be ignored or dismissed, then the deliberative process could have negative impacts on political efficacy in the long term. Halvorsen (2003: 539) writes that ‘high-quality citizen involvement could backfire if the public learns over time that their participation is meaningless. A history of participation with no visible impact on agency decisions can be worse than no participation at all’ (also see Dyck, 2009). This article highlights the short- and long-term effects of a deliberative event on participants’ political trust and efficacy. The case study is unique in the intensity of the deliberative event (6 days) as well as its clear linkage to the policy-making process. The hypotheses for (a) external efficacy and (b) internal efficacy are as follows:

H6a (Selection effect): Compared to non-participating citizens (control group), participants report higher levels of external political efficacy, on average, prior to the deliberative event, using contemporaneous survey data (September 2012).

H7a (Positive immediate effect): Participants will report higher average levels of external efficacy immediately after the deliberative event (December 2012), compared to before the deliberative event (September 2012).

H7b (Positive immediate effect): Participants will report higher average levels of internal efficacy immediately after the deliberative event (December 2012), compared to before the deliberative event (September 2012).

H8a (Positive long-term effect): Participants will report higher average levels of external efficacy at 6 months and 2.5 years after the event, compared to before the deliberative event (September 2012).

H8b (Positive long-term effect): Participants will report higher average levels of internal efficacy at 6 months and 2.5 years after the event, compared to before the deliberative event (September 2012).

H9a (Deflating effect in the post-deliberation period): Participants’ average levels of external efficacy will decrease in the post-deliberative period (December 2012, June 2013 and June 2015).

H9a (Deflating effect in the post-deliberation period): Participants’ average levels of internal efficacy will decrease in the post-deliberative period (December 2012, June 2013 and June 2015).

H10a (Enduring differences): Compared to non-participating citizens (control group), participants report higher levels of external efficacy, on average, after the deliberative event, using contemporaneous survey data (June 2013).

H10b (Enduring differences): Compared to non-participating citizens (control group), participants report higher levels of internal efficacy, on average, after the deliberative event, using contemporaneous survey data (June 2013 and June 2015).

Case Study: Citizens’ Panel on Edmonton’s Energy and Climate Challenges

In 2006, the City of Edmonton adopted a Public Involvement Framework, setting forth an agenda for integrating citizen engagement more systematically into the City’s decision-making process (City of Edmonton, 2005). In 2009, the Centre for Public Involvement was formed as a partnership between the City of Edmonton and the University of Alberta. The Centre for Public Involvement has organized several citizen panels to gather input on the topics, such as Internet voting, urban food and agricultural policy, as well as energy efficiency and climate change. In this article, I highlight the work of the Citizens’ Panel on Edmonton’s Energy and Climate Challenges. This panel was formed through a partnership between the Centre for Public Involvement, the City’s Office of Environment and Alberta Climate Dialogue, a university-community research organization. The panel was designed to engage a representative sample of Edmontonians in a deliberation about the City’s policies to address climate change.

The deliberative event meets the definition of a citizen jury offered by French and Laver (2009: 42). A citizen jury involves a group of citizens who are provided with briefing materials, meet face to face for at least 1 day and discuss different policy options while interacting with experts and political leaders. This deliberative event involved meeting for six Saturdays from October to December 2012 at the University of Alberta. The participants were tasked with writing a report to City Council about how to address climate and energy issues within the City of Edmonton. This deliberative event is distinctive in the involvement of City administrators and in its formal connection to the policy-making process. While some scholars are concerned that these deliberative events may be public relations exercises (French and Laver, 2009), in this case, this event had formal ties to the policy-making process, that is, the report would be tabled to City Council (April 2013) and City Council would consider the report in developing its policy plan (April 2015).

This case study is distinctive in the intensity of the deliberation and its connection to the decision-making process. Because of these distinctive characteristics, the case study is ideally suited to assess short- and long-term effects on political trust and efficacy. The research tracks participants immediately after the deliberative event (December 2012), when the recommendations report is tabled to City Council (April 2013) and, finally, when City Council formally responds to the recommendations report (April 2015). In this case study, the possibility of a negative effect could be explained by City Council’s reaction to the report. If participants feel that City Council did not give the report full consideration, they may feel disillusioned with democracy and consequently, express lower levels of their own efficacy and the system’s openness to citizens’ input. In addition, they may distrust government, if they viewed government as failing to act on their recommendations and in their best interests.

Like the Dublin Citizen Jury (French and Laver, 2009), participants were recruited by way of a random digit dialling recruitment strategy in September 2012. If they expressed an interest in participating, they were sent additional information. Once their participation was confirmed, they were offered a CDN$400 honorarium. While panellists were recruited using random digital dialling, quotas were established around age, education, gender, ethnicity, ward, income and employment in energy sector (Boulianne, 2018b). The quotas ensured that the panel characteristics matched the Census profile for the community. The recruitment process also sought to include those who believe in climate change as well as those who do not (Boulianne, 2018b; Centre for Public Involvement et al., 2013).

Approximately 66 participants were recruited to participate in the panel and 55 continued their involvement through all six Saturdays of the event. The number may seem small compared to other major deliberative projects such as the Citizens Assemblies on Electoral Reform in Ontario and British Columbia (Fournier et al., 2011) or Fishkin and Luskin’s (1999; Luskin and Fishkin, 2002) deliberative polls. However, our population size is much smaller in scale compared to those projects. As such, our sampling ratio is better. Our deliberative group is comparable in size to other key deliberative projects (Fournier et al., 2011 (Electoral Reform in the Netherlands); French and Laver, 2009) and much larger than other deliberative events studying political trust and efficacy (Munno and Nabatchi, 2014; Tomkins et al., 2010).

Methods

I use several sources of data to examine my research hypotheses. I use multiple random sample surveys of the general population, and I use multiple surveys of panellists before and after the deliberative event. The random sample surveys help us understand how panellists differed from the broader body of citizens. This comparison helps us understand the degree to which trust and efficacy may have motivated participation. In addition, I use the random sample surveys to understand how panellists have changed as a result of participating in the deliberative event. In particular, I compare random samples’ reports of trust and efficacy 6 months (June 2013) and 2.5 years (June 2015) after the deliberative event to panellists’ responses during the same time period to highlight the enduring effects of the deliberative process. In this sense, the random sample survey respondents serve as a control group, whereas the deliberating panellists are the experimental group (French and Laver, 2009). This comparison helps us understand how the deliberative event transformed these panellists, making them different from their non-participating counterparts in the general population.

I have three random sample surveys of the general population. The first survey was conducted as an interactive voice response (IVR) telephone survey by Probit, a market research firm. The survey was conducted in September 2012. A random sample of 1271 Edmontonians participated in this survey. The American Association of Public Opinion Research (AAPOR) Response Rate #1, which includes only fully completed surveys, is approximately 1%. The second and third random sample surveys were interviewer-led telephone surveys. The surveys were conducted in June to July 2013 and again in 2015, as part of the Alberta Survey series, which are free to download from the University of Alberta’s repository: https://dataverse.library.ualberta.ca/dvn/. The AAPOR Response Rate #2, which includes partially completed interviews, is approximately 11% for both surveys (Edmonton-based samples). These telephone surveys were conducted by the University of Alberta’s Population Research Lab. Approximately 400 Edmontonians completed this survey. However, because of a split ballot experiment embedded within the survey in 2013, the results presented are based on half of the total respondents (see Boulianne, 2018a). However, in 2015, the results are based on all 400 respondents residing in or around the City of Edmonton. The public opinion surveys coincide with the surveys of the panellists.

For those recruited to participate in the deliberative event (n = 66), I compared responses before and after the deliberative event to assess changes in political trust and efficacy over time. All panellists completed a short telephone survey (recruitment survey), and then were invited to complete a self-administered (Self-Assessment Questionnaire (SAQ)) web survey (pretest data). Of the 66 panellists, 44 completed the pretest survey. The short timeline between recruitment and the start of the deliberative event made it difficult to administer the pretest survey as a web survey, reducing the response rate.

As mentioned, panel attrition reduced the number of participants in the six-Saturday deliberative event (66 to 55). Among the 55 remaining panellists, 49 completed the survey on the last day of the deliberative event (Session 6 posttest). I distributed self-administered surveys (web or print depending on panellists’ preferences) 6 months and 2.5 years after the deliberative event to determine whether the changes were enduring. Surveying deliberative participants at multiple points in the deliberative process is common. What is unique in my design is that participants were surveyed 6 months (June 2013) and 2.5 years (June 2015) after the deliberative event to assess the long-term effects of the deliberative event in the context of their report being tabled to City Council (6 months) and City Council issuing a policy statement, which relates to the panellists’ recommendations (2.5 years).

The first question is about trust in the municipal government in general. This question is similar to questions posed by others (see Table 1). Following the literature’s findings about the nuances of political trust, I also asked about trust in a particular policy domain. I asked about trust in government around the policy issue (domain) under discussion at the deliberative event, namely, climate change. The two measures provide a holistic picture about how trust in government may shape the motivation to participate in a deliberative event as well as how participation in a deliberative event shapes trust. There were concerns about survey fatigue, given the number of surveys administered in the 4 months from recruitment to the end of the deliberative event. To minimize the burden of survey response, the two trust questions were asked at different points in time, rather than being repeated at each point in time. In the 6-month and 2.5-year follow-up surveys, both trust questions were asked.

The political efficacy questions were adapted from the American National Election Studies (2015). However, instead of asking a negatively framed ‘don’t care’, we asked a positively framed question and asked about the portion of politicians who care about what people think. This measure is used as an indication of external efficacy, following Luskin and Fishkin (2002). The second survey question was adapted from ‘People like me don’t have any say about what the government does’ to ‘How much can people like you affect what the government does?’ with response options of not at all, a little, a moderate amount, a lot and a great deal. In the literature, this measure is used in both measures of internal efficacy (Luskin and Fishkin, 2002) and external efficacy (Gastil and Xenos, 2010). To distinguish this measure from the other measure, I call this measure internal efficacy, again, following Luskin and Fishkin (2002). The existing literature uses these two versions of political efficacy equally (see Table 2).

The analysis focuses on aggregate means and t-test of group means. The reason for focusing on aggregate means, instead of individual-level changes, is that a core component of the analysis is comparing panellists’ responses to random sample surveys of the general population. This comparison establishes whether panellists reported higher levels of trust prior to the deliberative event, compared to non-participants, and whether panellists reported higher levels of trust after the deliberative event, compared to non-participants. In addition, respondents had the option of submitting their surveys anonymously. We offered this option to address concerns about social desirability and privacy. This strategy limits the analysis of individual-level changes, but is expected to increase the validity of the survey responses.

Results

The first set of analyses compares levels of general trust for the general public and panellists, as per Hypothesis 1a (selection effect). Panellists reported higher, on average, levels of political trust in municipal government compared to the general population (Table 3, Comparison 1). The difference is approximately 0.32 of a point on a five-point scale for the two surveys conducted simultaneously (t = 3.75, p < 0.001).

Table 3.

General Trust in Government.

	General population, IVR, September 2012, n = 1271	Recruitment, phone, September 2012, n = 65	Six-month survey, SAQ, June 2013, n = 46	2.5-year survey, SAQ, June 2015, n = 45	Difference, t-test, p-value
1	M = 3.16, SD = 1.11	M = 3.48, SD = 0.64			+0.32, 3.75, p < 0.001	H1a
2	M = 3.16, SD = 1.11		M = 3.15, SD = 0.89		−0.01, 0.07, p = 0.941
3	M = 3.16, SD = 1.11			M = 3.38, SD = 1.03	+0.22, 1.40, p = 0.162
4		M = 3.48, SD = 0.64	M = 3.15, SD = 0.89		−0.33, 2.15, p = 0.034	H3a
5		M = 3.48, SD = 0.64		M = 3.38, SD = 1.03	−0.10, 0.58, p = 0.564	H3a
6			M = 3.15, SD = 0.89	M = 3.38, SD = 1.03	0.23, 1.14, p = 0.258	H4a

IVR: interactive voice response; SAQ: Self-Assessment Questionnaire; SD: standard deviation.

The last column denotes the relevant hypotheses being tested. If there is a blank, the comparison does not relate to a specific hypothesis.

The reported t-test is the absolute value based on a two-tailed test.

Figure 1 tracks political trust before the deliberative event (September 2012), 6 months after the deliberative event when the report was presented to City Council (June 2013) and, finally, 2.5 years (June 2015) after the deliberation when City Council issued its new Energy Transition Plan, which incorporated the panel’s recommendations. Political trust dropped from its elevated level during the pretest to 6 months after the event, before rebounding 2.5 years after the event (Figure 1). In sum, while panellists reported higher levels of trust at the recruitment stage, compared to their non-participating counterparts (Hypothesis 1a), the deliberative event (Hypothesis 3a) and outcomes of the deliberative process did not further increase general political trust (Hypothesis 4a).

Figure 1.

General Trust in Government.

In terms of trust around climate-change decisions, panellists reported higher levels of trust after the deliberative event, compared to before the deliberative event. Figure 2 illustrates the trend through the four points of data collection. Average levels of trust around climate-change decisions started at 2.55 and ended at 3.05 on a four-point scale.

Figure 2.

Trust in Government related to Climate-Change Decisions (Domain-Specific).

The increase was observed at the posttest during the last of the six meetings (Table 4, Comparison 1) and 2.5 years later (Comparison 3). Comparing the pretest and immediate posttest, the increase was 0.45 on a four-point scale (t = 2.77, p = 0.007), in line with Hypothesis 2b (positive immediate effect). Comparing the pretest and posttest conducted 2.5 years after the deliberative event, when City Council issued a policy statement connected to the panel’s report, the increase was 0.50 on a four-point scale (t = 2.81, p = 0.006). This finding is in line with Hypothesis 3b (positive long-term effect).

Table 4.

Trust in Government Related to Climate-Change Decisions (Domain-Specific).

	Pretest survey, SAQ, October 2012, n = 38	Posttest survey, SAQ, December 2012, n = 45	Six-month survey, SAQ, June 2013, n = 47	General population, phone, June 2013, n = 210	2.5-year survey, SAQ, June 2015, n = 43	Difference, t-test, p-value
1	M = 2.55, SD = 0.76	M = 3.00, SD = 0.71				+0.45, 2.77, p = 0.007	H2b
2	M = 2.55, SD = 0.76		M = 2.83, SD = 0.82			+0.28, 1.63, p = 0.107	H3b
3	M = 2.55, SD = 0.76				M = 3.05, SD = 0.84	+0.50, 2.81, p = 0.006	H3b
4		M = 3.00, SD = 0.71	M = 2.83, SD = 0.82			−0.17, 1.06, p = 0.292	H4b
5		M = 3.00, SD = 0.71			M = 3.05, SD = 0.84	+0.05, 0.30, p = 0.765	H4b
6			M = 2.83, SD = 0.82		M = 3.05, SD = 0.84	−0.22, 1.26, p = 0.211	H4b
7			M = 2.83, SD = 0.82	M = 2.44, SD = 0.83		+0.39, 2.94, p = 0.004	H5b
8		M = 3.00, SD = 0.71		M = 2.44, SD = 0.83		−0.56, 4.65, p < 0.001
9	M = 2.55, SD = 0.76			M = 2.44, SD = 0.83		−0.11, 0.81, p = 0.419
10				M = 2.44, SD = 0.83	M = 3.05, SD = 0.84	+0.61, 4.35, p < 0.001

SAQ: Self-Assessment Questionnaire; SD: standard deviation.

The last column denotes the relevant hypotheses being tested. If there is a blank, the comparison does not relate to a specific hypothesis.

The reported t-test is the absolute value based on a two-tailed test.

Looking at the simultaneous surveys of the general population and panellists 6 months after the deliberation (June 2013), the increased level of trust in climate-change decisions is clear (Table 4, Comparison 7). The difference was 0.39 on a four-point scale (t = 2.94, p = 0.004) in line with Hypothesis 5b (enduring differences). In sum, the general measure of trust demonstrated already high levels of trust before the deliberative event and panellists’ levels of trust did not further increase. The nuanced measure of trust in climate-change decisions shows an increase in trust immediately after the deliberative event, which continued to remain elevated 6 months after the event and 2.5 years later.

In terms of government officials caring about citizens’ opinions, panellists reported higher levels of external efficacy before the deliberative process began, compared to the general public (Table 5, Comparison 1). The difference was 0.57 on a five-point scale (t = 4.64, p < 0.001). As observed with general trust, panellists began the process with elevated levels of external efficacy, offering support for Hypothesis 6a (selection effect).

Table 5.

Government Officials Care (External Efficacy).

	General Population, IVR, September 2012, n = 1273	Recruitment, phone, September 2012, n = 66	Posttest survey, SAQ, December 2012, n = 49	Six-month survey, SAQ, June 2013, n = 46	General population, phone, June 2013, n = 210	2.5-year survey, SAQ, June 2015, n = 45	Difference, t-test, p-value
1	M = 2.79, SD = 1.19	M = 3.36, SD = 0.96					+0.57, 4.64, p < 0.001	H6a
2	M = 2.79, SD = 1.19		M = 3.41, SD = 0.91				+0.62, 4.62, p < 0.001
3	M = 2.79, SD = 1.19			M = 3.09, SD = 1.05			+0.30, 1.89, p = 0.059
4	M = 2.79, SD = 1.19					M = 2.89, SD = 1.13	+0.10, 0.58, p = 0.562
5		M = 3.36, SD = 0.96	M = 3.41, SD = 0.91				+0.05, 0.28, p = 0.780	H7a
6		M = 3.36, SD = 0.96		M = 3.09, SD = 1.05			−0.27, 1.38, p = 0.170	H8a
7		M = 3.36, SD = 0.96				M = 2.89, SD = 1.13	−0.47, 2.28, p = 0.025	H8a
8				M = 3.09, SD = 1.05		M = 2.89, SD = 1.13	−0.20, 0.87, p = 0.387	H9a
9	M = 2.79, SD = 1.19				M = 2.58, SD = 1.12		+0.21, 2.49, p = 0.013
10		M = 3.36, SD = 0.96			M = 2.58, SD = 1.12		−0.78, 5.52, p < 0.001
12			M = 3.41, SD = 0.91	M = 3.09, SD = 1.05			−0.32, 1.58, p = 0.118	H9a
13			M = 3.41, SD = 0.91		M = 2.58, SD = 1.12		+0.83, 5.49, p < 0.001
14			M = 3.41, SD = 0.91			M = 2.89, SD = 1.13	−0.52, 2.44, p = 0.017	H9a
15				M = 3.09, SD = 1.05	M = 2.58, SD = 1.12		+0.51, 2.95, p = 0.003	H10a
16					M = 2.58, SD = 1.12	M = 2.89, SD = 1.13	+0.31, 1.67, p = 0.096

IVR: interactive voice response; SAQ: Self-Assessment Questionnaire; SD: standard deviation.

The last column denotes the relevant hypotheses being tested. If there is a blank, the comparison does not relate to a specific hypothesis.

The reported t-test is the absolute value based on a two-tailed test.

However, in the period following the conclusion of the deliberation, panellists’ efficacy decreased, as per Hypothesis 9a (deflating effect in the post-deliberation period). From the pretest to the 2.5-year posttest (Hypothesis 8a), panellists’ average rating of external efficacy decreased by 0.47 on a five-point scale (t = 2.28, p = 0.025). This same pattern of decline is evident among the general public (Table 5, Comparison 9). Despite the co-occurring decline, panellists’ efficacy remained higher than the general public’s external efficacy as measured by simultaneous surveys (Table 5, Comparison 15), in line with Hypothesis 10a (enduring differences). Six months after the conclusion of the deliberation, panellists reported, on average, 3.09, whereas the general public reported, on average, 2.58 on the question of whether government officials care (t = 2.95, p = 0.003). In sum, panellists’ views about politicians’ caring were elevated prior to the deliberative event, and in the post-deliberation period, external efficacy subsequently decreased at 2.5 years after the event (Figure 3). The decrease followed a trend in the general population’s views about politicians’ caring. However, panellists’ views about government caring were more favourable than their non-participating counterparts in the general public.

Figure 3.

Government Officials Care (External Efficacy).

In terms of perceived ability to influence government (internal efficacy), panellists reported higher levels immediately after and 6 months after the deliberative event compared to pretest levels (Figure 4), as per Hypothesis 7b (positive immediate effect) and 8b (positive long-term effect). Internal efficacy increased from 2.89 to 3.38 to 3.43 from the pretest to the last of the six meetings to 6 months after the deliberative event when the report was presented to City Council. However, 2.5 years after the event concluded, when City Council issued its new Energy Transition Plan, the average level of internal efficacy dropped to 2.83, as per Hypothesis 9b (deflating effect in the post-deliberation period). The increased efficacy reported after the deliberation did not endure, as levels of efficacy returned to pretest levels.

Figure 4.

Ability to Affect Government (Internal Efficacy).

The increases in internal political efficacy described above were significant (Table 6, Comparison 1), as per Hypotheses 7b. Comparing two simultaneous surveys of the panellists and the general population demonstrates that panellists report being more efficacious than the general population (Table 6, Comparison 8), as per Hypothesis 10b. Six months after the deliberative event, panellists reported, on average, higher levels of internal efficacy, than the general population surveyed at the same time. The difference was 0.84 on a five-point scale (t = 5.18, p < 0.001). As mentioned, average levels of efficacy decreased 2.5 years after the deliberation when City Council issued its new Energy Transition Plan, which incorporated the panel’s recommendation, compared to responses gathered after the deliberative event (Table 6, Comparisons 9 and 10), as per Hypothesis 9b. The decrease returned panellists to their original views (Comparison 5), which also mimicked average levels observed in the general public at the exact same time period (Comparison 15). In sum, the deliberative event and the period of 6 months after the deliberative event demonstrated an increase in participants’ perceived ability to influence government. However, this increase did not persist 2.5 years after the deliberative event. For both measures of political efficacy, the deliberative event clearly increased levels of efficacy 6 months after the deliberation.

Table 6.

Ability to Affect Government (Internal Efficacy).

	Pretest survey, SAQ, October 2012, n = 44	Posttest survey, SAQ, December 2012, n = 48	Six-month survey, SAQ, June 2013, n = 44	General population, phone, June 2013, n = 210	2.5-year survey, SAQ, June 2015, n = 42	General population, phone, June 2015, n = 400	Difference, t-test, p-value
1	M = 2.89, SD = 0.87	M = 3.38, SD = 0.94					+0.49, 2.60, p = 0.011	H7b
2	M = 2.89, SD = 0.87		M = 3.43, SD = 1.00				+0.54, 2.70, p = 0.008	H8b
3	M = 2.89, SD = 0.87			M = 2.59, SD = 1.07			−0.30, 1.99, p = 0.048
4	M = 2.89, SD = 0.87				M = 2.83, SD = 0.94		−0.06, 0.31, p = 0.757	H8b
5	M = 2.89, SD = 0.87					M = 2.97, SD = 1.13	−0.08, 0.56, p = 0.576
6		M = 3.38, SD = 0.94	M = 3.43, SD = 1.00				+0.05, 0.25, p = 0.803	H9b
7		M = 3.38, SD = 0.94		M = 2.59, SD = 1.07			+0.79, 5.00, p < 0.001
8			M = 3.43, SD = 1.00	M = 2.59, SD = 1.07			+0.84, 5.18, p < 0.001	H10b
9		M = 3.38, SD = 0.94			M = 2.83, SD = 0.94		−0.55, 2.77, p = 0.007	H9b
10			M = 3.43, SD = 1.00		M = 2.83, SD = 0.94		−0.60, 2.87, p = 0.005	H9b
11				M = 2.59, SD = 1.07	M = 2.83, SD = 0.94		+0.24, 1.47, p = 0.143
12		M = 3.38, SD = 0.94				M = 2.97, SD = 1.13	+0.41, 2.79, p = 0.005
13			M = 3.43, SD = 1.00			M = 2.97, SD = 1.13	+0.46, 2.86, p = 0.004
14				M = 2.59, SD = 1.07		M = 2.97, SD = 1.13	−0.38, 4.09, p < 0.001
15					M = 2.83, SD = 0.94	M = 2.97, SD = 1.13	−0.14, 0.90, p = 0.369	H10b

SAQ: Self-Assessment Questionnaire; SD: standard deviation.

The last column denotes the relevant hypotheses being tested. If there is a blank, the comparison does not relate to a specific hypothesis.

The reported t-test is the absolute value based on a two-tailed test.

Table 7 summarizes the results of hypothesis-testing. Some hypotheses are tested multiple times (e.g. multiple tests about changes in efficacy and trust in the post-deliberation period). The Bonferroni correction is intended to account for the problems of multiple tests and committing Type 1 errors. The adjustment involves dividing the alpha/level of significance by the number of tests of a hypothesis. The alpha/level of significance is set at 0.05 as per existing research (Tables 1 and 2) and as per guidelines in top political science journals (e.g. American Journal of Political Science). For three tests of a hypothesis, the threshold for each individual test would be calculated as 0.05/3 = 0.017. Each individual test would need to meet this threshold. There are three tests of the hypothesis about a deflating effect in the post-deliberative period for domain-specific trust, external efficacy, and internal efficacy. The findings suggest that both measures of trust remain constant in the post-deliberation period; internal efficacy deflates in the post-deliberation period as two of the three tests meet the adjusted level of significance.

Table 7.

Summary of Hypothesis-Testing.

	General political trust (Table 3, Figure 1)	Domain-specific trust (Table 4, Figure 2)	External efficacy, (Table 5, Figure 3)	Internal efficacy, (Table 6, Figure 4)
Hypothesis: Selection effect	Hypothesis 1a		Hypothesis 6a	Hypothesis 6b
September 2012	p < 0.001	No data	p < 0.001	No data
Hypothesis: Positive immediate effect		Hypothesis 2b	Hypothesis 7a	Hypothesis 7b
September 2012 (pretest) and December 2012 (posttest)	No data	p = 0.007	p = 0.780	p = 0.011
Hypothesis: Positive long-term effect	Hypothesis 3a	Hypothesis 3b	Hypothesis 8a	Hypothesis 8b
September 2012 (pretest) and June 2013 (6 months)	p = 0.034	p = 0.107	p = 0.170	p = 0.008
September 2012 (pretest) and June 2015 (2.5 years)	p = 0.564	p = 0.006	p = 0.025	p = 0.757
Hypothesis: Deflating effect in the post-deliberation period	Hypothesis 4a	Hypothesis 4b	Hypothesis 9a	Hypothesis 9b
December 2012 (posttest) and June 2013 (6 months)	p = 0.258	p = 0.292	p = 0.118	p = 0.803
December 2012 (posttest) and June 2015 (2.5 years)		p = 0.765	p = 0.017	p = 0.007
June 2013 (6 months) and June 2015 (2.5 years)		p = 0.211	p = 0.387	p = 0.005
Hypothesis: Enduring differences		Hypothesis 5b	Hypothesis 10a	Hypothesis 10b
June 2013	No data	p = 0.004	p = 0.003	p < 0.001
June 2015				p = 0.369

As for two tests of a hypothesis, the threshold for each individual test would be 0.05/2 = 0.025 (Bonferroni correction). For the two tests of positive long-term effects of deliberative events on panellists, all but the general trust meet this new adjusted threshold for significance (Table 7). In sum, there is mixed support for the hypothesis about positive, enduring effects of deliberative events on political trust and efficacy.

Finally, there are two tests of internal political efficacy that compare panellists and a control group (random sample of non-participants). One test was completed at 6 months (June 2013) and the other at 2.5 years post-deliberation (June 2015). Six months after the deliberation, panellists reported higher, on average, levels of internal efficacy compared to their non-participating counterparts (p < 0.001). These differences do not continue at 2.5 years for measures of internal efficacy. At 2.5 years, panellists did not differ from their non-participating counterparts in terms of internal efficacy (Table 7).

The remaining hypotheses rely on single tests. For domain-specific trust and external efficacy, panellists report higher levels than their non-participating counterparts, supporting the hypotheses about enduring differences. For domain-specific trust and internal efficacy, panellists report higher levels immediately after the deliberation, compared to before the deliberation, supporting the hypothesis about positive, immediate effects. As for the selection hypothesis, the findings are consistent: panellists began the deliberative process with higher levels of trust and efficacy, compared to their non-participating counterparts.

Discussion

In sum, the findings demonstrate that when comparing panellists to the general population, panellists are distinctive in having higher levels of trust and efficacy before the deliberative event (hypothesis about selection effects). The findings suggest a self-selection bias where trust and efficacy are factors influencing the decision to participate. Furthermore, when comparing panellists 6 months after the deliberation to simultaneously collected polls of the general population, panellists remain distinctive in their high levels of trust and efficacy (hypothesis about enduring differences). This finding is a key contribution since most studies rely on participant surveys and few use control groups composed of a representative sample of the public (Tables 1 and 2). When studies do use such designs (Tables 1 and 2), they do not find significant differences. The only study to find significant differences between panellists and a randomly selected group is Luskin and Fishkin (2002). This finding has not been replicated until now. The current study shows significant differences between panellists and a control group. The current study is distinctive in demonstrating that these differences endure in the long term (Tables 4 and 5).

As to how the deliberative event and the period after this event affect panellists, the findings depend on the specific measure. Like Gastil et al. (2008), I examine a more nuanced measure of political trust. Gastil et al. (2008) look at changes in trust in juries and judges when assessing the effects of deliberative talk among jurors. Likewise, I examine changes in trust around municipal policy-making on climate change among panellists who participated in a deliberative event about climate change policies. This approach is in contrast to other studies that examine general trust in government institutions, such as Parliament (see Table 1). We believe that this more nuanced approach establishes a stronger theoretical connection between changes in trust and participation in a deliberative event. This stronger theoretical connection is matched with stronger empirical evidence (Table 4 versus Table 3). As per the hypothesis about positive immediate effects, participation in a deliberative event increases trust in government decision-making in the policy domain under discussion.

Our findings offer mixed evidence for the hypotheses about long-term effects. I found that 6 months after the deliberative event, participants still had high levels of trust in the municipal government’s policies around climate change when compared to the general population (Table 4). Both measures of efficacy show a similar pattern, with panellists reporting higher levels compared to the general population at the 6-month follow-up, which is when the panellists’ report was presented to City Council (Tables 5 and 6), as per the hypothesis about enduring differences. These findings are a critical contribution to research in this area, which rarely examines enduring effects, and when enduring effects are examined, they are not matched to government officials’ responses to participants’ recommendations. As mentioned, there is a possibility that government’s response could deflate or sustain political trust and efficacy depending on their response to the recommendations report. There is some fluctuation in efficacy and trust in the post-deliberation period, but the hypothesis about deflating effects is only supported in relation to internal efficacy.

Can deliberative events help address legitimacy in government? My findings suggest that deliberative events generate a context-specific trust in government and changes in political efficacy. These findings suggest that deliberative events can affect perceived legitimacy in government. This article presents systematic evidence of deliberative events’ impact on participants’ political trust and efficacy (Tables 1 and 2). However, a full-scale, systematic review of the outcomes of deliberative events is overdue. This review will need to account for the nuances of the deliberative event, such as perceptions of procedural justice and other characteristics of the deliberative event, which might explain the outcomes on trust and efficacy. Furthermore, published studies tend to be biased towards significant findings (Lipsey and Wilson, 2001). In the field of deliberative democracy, we should also attend to a positivity bias as Spada and Ryan (2017) point out that few published studies highlight failures or negative impacts of deliberative events. In understanding which events have positive and which have negative effects, the characteristics of the deliberative events as well as participants’ perceptions of procedural justice will likely factor into the observed effects. In sum, the degree to which deliberative events can address low levels of faith in democratic systems remains an open question.

Footnotes

Funding

This work was supported by a Community-University Research Alliance grant from the Social Sciences and Humanities Research Council to the Alberta Climate Dialogue group, #833-2009-4011. The author would like to thank David Kahane, Kristjana Loptson, and the Alberta Climate Dialogue Steering Committee for supporting this research.

Author Biography

Shelley Boulianne is an associate professor of sociology at MacEwan University. She completed her PhD in sociology at the University of Wisconsin–Madison. Her research focuses on media use, civic and political engagement, as well as survey research methodology.

References

American National Election Studies (2015) User’s Guide and Codebook for the ANES 2012 Time Series Study. Available at: http://www.electionstudies.org/studypages/anes_timeseries_2012/anes_timeseries_2012_userguidecodebook.pdf#page=1&zoom=auto,-42,798 (accessed 7 October 2015).

Andersen

Hansen

(2007) How Deliberation Makes Better Citizens: The Danish Deliberative Poll on the Euro. European Journal of Political Research 46 (4): 531–556.

Blais

Gidengil

Nevitte

et al . (2004) Where Does Turnout Decline Come from? European Journal of Political Research 43 (2): 221–236.

Boulianne

(2018a) Mini-Publics and Public Opinion: Two Survey-Based Experiments. Political Studies 66 (1): 119–136.

Boulianne

(2018b) Beyond the Usual Suspects: Representation in Alberta Climate Dialogue’s Deliberative Exercises. In: Hanson

(ed.) Public Deliberation on Climate Change: Lessons from Alberta Climate Dialogue. Edmonton, AB, Canada: Athabasca University Press. Available at: http://www.aupress.ca/books/120271/ebook/05_Hanson_2018-Public_Deliberation_on_Climate_Change.pdf (accessed 20 February 2018).

Carpini

MXD

Cook

Jacobs

(2004) Public Deliberation, Discursive Participation, and Citizen Engagement: A Review of the Empirical Literature. Annual Review of Political Science 7: 315–344.

Centre for Public Involvement, Alberta Climate Dialogue, City of Edmonton, et al. (2013) Citizens’ Panel on Edmonton’s Energy and Climate Challenges. Available at: https://www.edmonton.ca/city_government/documents/PDF/CitizensPanel-EnergyClimateChallenge.pdf (accessed 5 February 2018).

Citrin

Luks

(2001) Political Trust Revisited: Déjà vu All Over Again? In: Hibbing

Theiss-Morse

(eds) What Is It about Government That Americans Dislike. New York: Cambridge University Press, pp.9–27.

City of Edmonton (2005) City Policy C513. Available at: https://www.edmonton.ca/city_government/documents/PDF/city_procedure.pdf (accessed 13 February 2018).

10.

Cole

Kincaid

Rodriguez

(2004) Public Opinion on Federalism and Federal Political Culture in Canada, Mexico, and the United States, 2004. The Journal of Federalism 34 (3): 189–201.

11.

Dalton

(2004) Democratic Challenges, Democratic Choices: The Erosion of Political Support in Advanced Industrial Democracies. Oxford: Oxford University Press.

12.

Dyck

(2009) Initiated Distrust: Direct Democracy and Trust in Government. American Political Research 37 (4): 539–168.

13.

Fishkin

Luskin

(1999) Bringing Deliberation to the Democratic Dialogue. In: McCombs

Reynolds

(eds) The Poll with a Human Face: The National Issues Convention Experiment in Political Communication. Mahwah, NJ: Lawrence Erlbaum, pp.3–38.

14.

Fournier

van der Kolk

Carty

et al . (2011) When Citizens Decide: Lessons from Citizen Assemblies on Electoral Reform. Oxford: Oxford University Press.

15.

French

Laver

(2009) Participation Bias, Durable Opinion Shifts and Sabotage through Withdrawal in Citizens’ Juries. Political Studies 57 (2): 422–450.

16.

Gallup (2013) Trust in Government. Available at: http://www.gallup.com/poll/5392/trust-government.aspx (accessed 7 October 2015).

17.

Gastil

Xenos

(2010) Of Attitudes and Engagement: Clarifying the Reciprocal Relationship between Civic Attitudes and Political Participation. Journal of Communication 60 (2): 318–343.

18.

Gastil

Black

Deess

et al . (2008) From Group Member to Democratic Citizen: How Deliberating with Fellow Jurors Reshapes Civic Attitudes. Human Communication Research 34 (1): 137–169.

19.

Gastil

Knobloch

Kelly

(2012) Evaluating Deliberative Public Events and Projects. In: Nabatchi

Gastil

Weiksner

et al . (eds) Democracy in Motion: Evaluating the Practice and Impact of Deliberative Civic Engagement. Oxford: Oxford University Press, pp.205–230.

20.

Grönlund

Setälä

Herne

(2010) Deliberation and Civic Virtue: Lessons from a Citizen Deliberation Experiment. European Political Science Review 2 (1): 95–117.

21.

Halvorsen

(2003) Assessing the Effects of Public Participation. Public Administration Review 63 (5): 535–543.

22.

Hetherington

Husser

(2012) How Trust Matters: The Changing Political Relevance of Political Trust. American Journal of Political Science 56 (2): 312–325

23.

Holbrook

(2004) Trust in Government. In: Geer

(ed.) Public Opinion and Polling around the World: A Historical Encyclopedia. Santa Barbara, CA: ABC-CLIO. Available at: http://ezproxy.macewan.ca/login?url=http://search.credoreference.com/content/entry/abcopinion/trust_in_government/0?institutionId=2632 (accessed 20 February 2018).

24.

Lafont

(2015) Deliberation, Participation, and Democratic Legitimacy: Should Deliberative Mini-Publics Shape Public Policy? The Journal of Political Philosophy 23 (1): 40–63.

25.

Lenard

(2008) Trust Your Compatriots, But Count Your Change: The Roles of Trust, Mistrust and Distrust in Democracy. Political Studies 56 (2): 312–332.

26.

Levi

Stoker

(2000) Political Trust and Trustworthiness. Annual Review of Political Science 3: 475–507.

27.

Lipsey

Wilson

(2001) Practical Meta-Analysis. Thousand Oaks, CA: SAGE.

28.

Luskin

Fishkin

(2002) Deliberation and ‘Better Citizens’. In: Joint Sessions of Workshops of the European Consortium for Political Research, Turin, 22–27 March. Available at: http://cdd.stanford.edu/research/papers/2002/bettercitizens.pdf (accessed 13 February 2015).

29.

Marien

Hooghe

(2011) Does Political Trust Matter? An Empirical Investigation into the Relation between Political Trust and Support for Law Compliance. European Journal of Political Research 50 (2): 267–291.

30.

Marien

Kern

(2017) The Winner Takes It All: Revisiting the Effect of Direct Democracy on Citizens’ Political Support. Political Behavior. Epub ahead of print 6 September. DOI: 10.1007/s11109-017-9427-3.

31.

Morrell

(2005) Deliberation, Democratic Decision-Making and Internal Political Efficacy. Political Behavior 27 (1): 49–69.

32.

Munno

Nabatchi

(2014) Public Deliberation and Co-Production in the Political and Electoral Arena: A Citizens’ Jury Approach. Journal of Public Deliberation 10: 1–29.

33.

Mutz

(2008) Is Deliberative Democracy a Falsifiable Theory? Annual Review of Political Science 11: 521–538.

34.

Nabatchi

(2010) Deliberative Democracy and Citizenship: In Search of the Efficacy Effect. Journal of Public Deliberation 6: 1–47.

35.

Nabatchi

(2012) An Introduction to Deliberative Civic Engagement. In: Nabatchi

Gastil Weiksner

et al . (eds) Democracy in Motion: Evaluating the Practice and Impact of Deliberative Civic Engagement. Oxford: Oxford University Press, pp.3–17.

36.

Norris

(2011) Democratic Deficit: Critical Citizens Revisited. Cambridge: Cambridge University Press.

37.

Parkins

McFarlane

(2015) Trust and Skepticism in Dynamic Tension: Concepts and Empirical Refinements from Research on the Mountain Pine Beetle Outbreak in Alberta, Canada. Human Ecology Review 21 (1): 133–153.

38.

Pew Research Center (2013) Trust in Government Nears Record Low, But Most Federal Agencies Are Viewed Favorably. Available at: http://www.people-press.org/files/legacy-pdf/10-18-13%20Trust%20in%20Govt%20Update.pdf (accessed 13 February 2015).

39.

Pew Research Center (2017) Public Trust in Government Remains Near Historic Lows as Partisan Attitudes Shift. Available at: http://assets.pewresearch.org/wp-content/uploads/sites/5/2017/05/03145544/05-03-17-Trust-release.pdf (accessed 18 October 2017).

40.

Pickup

Sayers

Knopff

et al . (2004) Social Capital and Civic Community in Alberta. Canadian Journal of Political Science / Revue Canadienne de Science Politique 37 (3): 617–645.

41.

Putnam

(2000) Bowling Alone: The Collapse and Revival of American Community. New York: Touchstone.

42.

Quintelier

Van Deth

(2014) Supporting Democracy: Political Participation and Political Attitudes: Exploring Causality Using Panel Data. Political Studies 62 (S1): 153–171.

43.

Smets

Van Ham

(2013) The Embarrassment of Riches? A Meta-Analysis of Individual-Level Research on Voter Turnout. Electoral Studies 32 (2): 344–359.

44.

Spada

Ryan

(2017) The Failure to Examine Failures in Democratic Innovation. PS: Political Science & Politics 50: 772–778.

45.

Strandberg

Grönlund

(2012) Online Deliberation and its Outcome – Evidence from the Virtual Polity Experiment. Journal of Information Technology & Politics 9 (2): 167–184.

46.

Thomassen

Van Ham

(2017) A Legitimacy Crisis of Democracy? In: Van Ham

Thomassen Aarts

et al . (eds) Myth and Reality of the Legitimacy Crisis: Explaining Trends and Cross-National Differences in Established Democracies. Oxford: Oxford University Press, pp.3–16.

47.

Tomkins

Pytlik Zillig

Herian

et al . (2010) Public Input for Municipal Policymaking: Engagement Methods and Their Impact on Trust and Confidence. In: 11th Annual International Conference on Digital Government Research, 3–6 September 2009. Available at: https://pdfs.semanticscholar.org/f5bd/b6a7d3170e620c69fae89f397f5f34ce1ed6.pdf (accessed 13 February 2018).

48.

Ulbig

(2007) Gendering Municipal Government: Female Descriptive Representation and Feelings of Political Trust. Social Science Quarterly 88 (5): 1106–1123.

49.

Warren

(1992) Democratic Theory and Self-Transformation. American Political Science Review 86 (1): 8–23.

50.

Warren

(1999) Introduction. In: Warren

(ed.) Democracy and Trust. Cambridge: Cambridge University Press, pp.1–21.

51.

Warren

(2008) Citizen Representatives. In: Warren

Pearse

(eds) Designing Deliberative Democracy: The British Columbia Citizens’ Assembly. Cambridge: Cambridge University Press, pp.50–69.

52.

Warren

(2009) Two Trust-Based Uses of Minipublics in Democracy. Paper Presented at the American Political Science Association Annual Meeting, Toronto, ON, Canada.

53.

Wright

(2010) Envisioning Real Utopias. New York: Verso.