Participation, Representation and Expertise: Citizen Preferences for Political Decision-Making Processes

Participation, representation and expertise: Citizen preferences for political decisionmaking processes

Joan Font Institute of Advanced Social Studies (IESA-CSIC)

Magdalena Wojcieszak University of Amsterdam

Clemente Navarro Universidad Pablo de Olavide

This paper was submitted to Political Studies in November 2013. A newer and richer version has been published by Political Studies in February 2015 http://onlinelibrary.wiley.com/doi/10.1111/1467-9248.12191/abstract

Joan Font is the Director of the Institute of Advanced Social Studies (IESA-CSIC) in Cordoba, (Spain). Magdalena Wojcieszak (PhD. Communication, Annenberg School for Communication, University of Pennsylvania) is an Associate Professor at Amsterdam University (Netherlands). Clemente Navarro is Professor at the Centre for Local Political Sociology and Policies at the Universidad Pablo de Olavide (Seville, Spain). Keywords: Democracy, participation, representation, stealth democracy, process preferences 1

PARTICIPATION, REPRESENTATION AND EXPERTISE?: CITIZEN PREFERENCES FOR POLITICAL DECISION-MAKING PROCESSES

Do citizens have preferences regarding how political decisions should be made or do they solely care about the outcomes? If such preferences exist, what are they? Democratic debate has traditionally emphasized representative and participatory governance, analyzing whether citizens prefer to yield their voice to elected representatives or rather want a more direct say in the policy matters of the day. Whatever problems representation and participation had in practice, scholars have generally presumed that citizens desire more participatory opportunities, a presumption that has instigated an impressive body of theory, research and practice on direct and deliberative democracy. This presumption, however, has been questioned. Most notably, in Stealth Democracy (2002), Hibbing and Theiss-Morse have found that the public – at least in the US – would rather passively observe the political system and leave decision-making to impartial experts or effective businessmen, a finding which has popularized a particular version of an expert-based governance model (Lippmann, 1922). Evidence on citizen process preferences more generally, and on stealth democracy model more specifically, has been limited and inconclusive. Some research has found considerable citizen demand for more participatory spaces, thus challenging the stealth democracy thesis. This research, however, has been mostly based on single questions, often tapping the demand for referenda (Bengtsson and Mattila, 2009; Donovan and Karp, 2006), or has focused on citizen willingness to engage in such direct participatory forms as deliberative forums (Neblo et al,

2

2010). This mixed evidence suggests a complex portrayal of citizen preferences for political decision-making processes and underscores the need for further research. Recent political events have also highlighted this complexity. The Occupy movement (or its Indignados Spanish version) or the Iceland grassroots mobilization that led to a new Constitutional project have placed the demand for more direct citizen participation on the political agenda. At the same time, Greece and Italy had Prime Ministers who were not elected by the general public, but rather designated as “technocrats” who would apply efficient technical solutions in their governments. These events also underscore the salience of the under-examined tensions between representative processes and other alternatives that entail greater direct citizen participation and/or expert engagement. Our paper speaks to these issues, adding comparative evidence on citizen preferences for political decision-making processes. We rely on data from a nationally representative survey administered to the Spanish adult population (N = 2450) in February 2011. The questionnaire was developed to test the stealth democracy thesis in Spain and included the original book central items as well as additional measures that tap preferences for decision-making mechanisms representing participatory, representative, and expert-based governance. We use these data to show that citizen preferences for political decision-making processes are multi-dimensional. In examining what dimensions emerge among the public, we extend the discussion on the tensions and potential relationships between direct citizen engagement and governance via elected representatives by revealing a third substantive model, expert-based governance. Before presenting the data, we briefly discuss the debates on representative, participatory and expert based democracy. We then outline the survey and discuss the sociopolitical context in

3

which the study was conducted. After briefly reviewing the preferences for political decisionmaking processes in Spain to set the stage, we rely on Confirmatory Factor Analysis (CFA) to examine the multidimensional character of these preferences. The last section discusses the theoretical, substantive and methodological implications of these results.

Process preferences: representation, participation and stealth democracy Democratic theory has long centered on the debate between representative and participatory governance models, emphasizing the tensions or the dichotomy between these two. On one side of the democratic spectrum stands the model of articulating citizen demands through representation. This model often takes the form of party-based parliamentary rule and functions primarily through elected representatives. (…) At the other end of the spectrum is the model of direct democracy, placing control of government in the hands of the people themselves. This model argues that citizens themselves can make wise decisions on political matters, whether through referenda, town meetings, citizen initiatives, or other direct means (Dalton et al. 2001, p. 142).

Representative system has been the most common and the most established form of governance in Western democracies. In fact, some authors do not see public participation as essential. Politics is complex and governance requires time, knowledge and expertise that cannot be expected from an ordinary citizen. The state would thus be best governed by elected leaders with only periodic engagement by the electorate (Schumpeter, 1943). Another strand of democratic theory has presupposed engaged citizens who monitor the leaders and take political matters in their own hands, via assemblies, community organizing or the famous town hall meetings described by de Tocqueville (2000 [1851]; see also, Fung and Wright, 2001). Such direct participation would benefit the public as well as the state (Roberts, 2004).

4

In fact, some studies have found demand for more participatory spaces, with citizens in Western democracies supporting direct engagement in policy-making through referenda (Anderson and Goodyear-Grant, 2010, Bengtsson and Mattila, 2009; Bowler, Donovan and Karp, 2007; Dalton et al., 2001; Donovan and Karp, 2006) and through group deliberations on salient policy issues (Neblo et al., 2010; Jacobs, Lomax and Delli Carpini, 2009). Many governmental and non-governmental organizations have followed suit, funding and organizing various initiatives aimed at increasing direct citizen engagement in political decision-making (Gastil, 2000; Smith, 2009; Papadopoulos and Warin, 2007). In contrast, Hibbing and Theiss-Morse (2002) have shown that most citizens do not seem to desire such an engagement. Americans are dissatisfied with debates and compromises typically present in governing bodies, perceive politicians as corrupt and self-interested, distrust their fellow citizens, and do not want to take part in activities that entail conflict. As a result, many Americans would rather passively observe the political process, and support stealth democracy, where the goal is for decisions to be made “efficiently, objectively and without commotion and disagreement,” and where citizens do not routinely play a role in making decisions, providing input or monitoring decision makers (Hibbing and Theiss-Morse 2002, p. 143). This bleak finding has popularized the debate on expert-based governance. The fact that most Americans want rule by “empathetic, non-self-interested decision makers” (p. 216) and actually prefer the rule by unaccountable experts over other forms of decision-making, echoes what Lippmann (1922) proposed almost a century ago. He recognized that most citizens lack the time, interest, expertise, or willingness to engage meaningfully in the complex political process, and suggested that pragmatic decisions on specialized affairs would be best and most efficiently

5

made by knowledgeable and skilled experts. Because “wicked problems” (Fischer 1993; Roberts 2001) and complex issues, such as budget cuts, climate change, or the use of biotechnologies, are increasingly present on the political agenda and are beyond the expertise of ordinary citizens and of most elected politicians, some scholars have suggested that these matters can be best resolved via expert governance (see Brooks, Stasiak and Zyro, 2012).

The Dimensionality of Political Process Preferences We argue that in order to advance this debate it is necessary to acknowledge that citizen process preferences have different and independent dimensions. In order to begin to address this question and test the complexity of citizen preferences it is necessary to capture attitudes towards various processes (representative, participatory and expert-based governance) and to examine their underlying structure. Research on preferences for representative or participatory processes typically has studied a limited spectrum of political mechanisms, mostly focusing on referenda (Budge, 1996; Bowler et al., 2007; Dalton et al., 2001; Donovan and Karp, 2006), or less often on deliberative forums, as studied in the US (Neblo et al., 2010; Jacobs et al., 2009). Some scholars have argued that these two participatory proposals should be distinguished because the most typical mechanisms of direct democracy, referenda, clearly differ from intense deliberative practices like assemblies, town meetings or participatory budgeting (Webb, Bale and Taggart, 2010; Webb, 2013). However, because most surveys have tested support for referenda or gauged citizen attitudes towards deliberative forums, the distinction between these low and high-intensity proposals has not been tested within a single study. Hence, the complexity of citizen preferences for a wider range of political decision-making mechanisms remains relatively unexamined.

6

In turn, studies on stealth or expert-based processes have not been analyzed systematically. Hibbing and Theiss-Morse (2002) present a strong case for support of this model, but their quantitative evidence is not undisputable. For example, in the first part of the book they rely on a bipolar process scale that captures the opposition between participatory and representative ideas and shows considerable support for more participatory practices, whereas in the second part their argument is based on the stealth democracy index that averages four distinct itemsi. To complicate matters further, Hibbing and Theiss-Morse (2002: 139) find that the process scale and the stealth democracy items are unrelated. The authors conclude that “far from being diametrically opposed, the apparent desire to empower people often cohabits with the desire to empower entities virtually unconnected to the people” (Hibbing and Theiss-Morse, 2002: 140). Evidence from other countries suggests that certain support for a stealth democracy model goes clearly beyond the US case (Gangl, 2007; Hilde and Michels, 2012; Webb, 2013). These results suggest that citizen preferences for political decision-making processes are complex and multidimensional and must be analyzed as such. Addressing these complexities we ask what are the relevant dimensions of citizens’ process preferences. Can these preferences be reduced to an opposition between participation and representation or do they also need to incorporate preferences for expert governance as an independent dimension? In asking these questions, we thus capture the classical debate around participatory and representative governance as well as another important yet less explored debate on the role that experts, technical arguments or efficiency criteria should play in political decision-making.

7

The relationship between the different dimensions In addition to examining the multidimensional nature of citizen process preferences, we aim to shed some light on how the debates on participatory, representative, and expert-based governance relate with each other. We suggest three potential relationships between them (see Figures 1a-1c). These alternatives are differentiated based on the most important factor that may guide citizen preferences: opposition to conventional politics, prioritizing values versus efficiency in policy-making, and the extent of citizen engagement in the policy process. [Figures 1a-1c about here] First, as argued by Hibbing and Theiss-Morse (2002: 140), it is possible that citizens are dissatisfied with representative governance and would favor any alternative to it as long as it reduces the power of politicians. Inasmuch as both participatory and expert-based processes present alternatives for those dissatisfied with the existing representation, these two forms of decision-making would be empirically close to each other and distinct from the representative process (Figure 1a). Previous research, using different operationalization of expert-based governance and drawing on data from two different countries, would suggest such a relationship. For example, Bengtsson and Mattila (2009) show that those Finish citizens who support the stealth democracy model and those who support referenda-based democracy share a common critical position towards representative institutions (see also Webb, 2013 on the British case). If, however, citizens are mostly concerned with the criteria that should guide political decision-making, whether political values or efficiency and expertise, then representative and participatory processes would be close to each other and opposed to expert-based governance (Figure 1b). Qualitative research from Spain and Italy suggests such an association, with many citizens favoring the incorporation of some expertise into political decision-making as long as

8

this expert advice is constrained and the final decision is made by a political body composed of citizens or elected representatives (Ganuza et al, 2013). Finally, it is possible that the extent of citizen involvement would be the most important consideration for the public, guiding its process preferences. In this case, expert-based and representative governance, which entail limited citizen engagement, would be close to each other and opposing the participatory process (Figure 1c). Movements like the Italian Movimento cinque stelle, which criticizes both the technocratic approach of the Monti government as well as the traditional party representation, could represent this position. All in all, we argue that citizen preferences for political decision-making processes are multidimensional. We test whether these preferences reflect the distinction between participatory, representative, and expert-based governance. Our discussion also examines which of the three alternative relationships between the three tested processes (i.e., participatory, representative, and expert), as represented in Figure 1, best reflects citizen preferences.

Method and context We rely on data from a national survey administered face-to-face on a representative sample of the Spanish population (N = 2450). The survey was sponsored and conducted by Centro de Investigaciones Sociológicas (CIS, Survey Research Center), the main public opinion survey institute in Spain, in February 2011.ii The sample was stratified to region and size of municipality. Municipalities were selected randomly as primary sampling units, and from those municipalities, sections were selected with probability proportional to size. In the final stage, a person was selected by performing a random walk through the section with age and sex quotas.iii

9

The survey included the most important Stealth Democracy (2002) questions and additional items that tap alternative decision-making processes (see the Appendix for question wording).iv Spain is a relevant context in which to test citizen process preferences. Spain has a shorter and different democratic history than the other countries where these ideas have been tested (Netherlands, Finland, UK, US), representing a Southern European citizenship model with low political interest (Martin and Van Deth, 2007), low organizational membership (Morales, 2009), and low political participation, except for protest activities (Van Deth, Montero and Westholm, 2007). At the same time, the demand for more participatory opportunities has recently re-emerged on the political agenda through the Indignados movement, which shook the Spanish political scene in 2011 (Hughes, 2011) and was largely supported by the general public.v The specific context of early 2011 was one where the economic crisis had already hit Spain, with unemployment rising to over 20% and with 82% of the public mentioning it as one of the three most important problems facing the country. The political situation was also rated negatively, with austerity policies, cuts in social services as well as salary reductions for public employees already implemented, even if the most dramatic cuts were introduced later in 2012vi.

Measures Political Decision-Making Processes. To comprehensively portray citizen preferences, we included items that tap participatory, representative, and expert models. We also differentiated between the two participatory proposals suggested above, referenda and assemblies. For each model, we asked two batteries of questions. Respondents firs evaluated - on a scale from 0 (“It’s the worst form of decision making”) to 10 (“It’s the best form of decision making”) the following general proposals: frequently consulting citizens about their opinions (M

10

= 7.20, SD = 2.51), facilitating it for people to debate political decisions (M = 7.24, SD = 2.31), leaving the decisions to experts in a subject (M = 7.00, SD = 2.21) and electing politicians who make decisions (M = 7.19, SD = 2.44). The second battery asked the respondents to evaluate – on the same scales – the mechanisms through which these models could be implemented: referenda (M = 6.20, SD = 2.62), assemblies and public meetings for collective decision-making (M = 6.57, SD = 2.43), expert consultancy in decision-making (M = 6.71, SD = 2.22), and leaving decisions to the politicians (M = 4.89, SD = 2.57). Stealth Democracy Items. Our survey also included the original SD index to further tap expert-based governance. On a scale from 1 (“Strongly Disagree”) to 4 (“Strongly Disagree”), respondents expressed their agreement with four statements: (1) Elected officials would help the country more if they would stop talking and just take action on important problems (M = 3.64, SD = 1.19), (2) What people call “compromise” in politics is just selling out on one’s principles (M = 2.51, SD =.93), (3) Our government would run better if decisions were left up to successful business people (M = 2,33, SD = .99), and (4) Our government would run better if decisions were left up to non-elected, independent experts rather than politicians or the people (M = 2.93, SD = .84). Process Scale. Respondents were asked to place themselves between two poles with 0 representing a situation where, in Spain, it is the citizens who would take all decisions and 10 indicating a situation where elected politicians would take all decisions (M = 4.45, SD = 2.47).

11

Results Before turning to Confirmatory Factor Analysis (CFA), we examine the preferences for political decision-making processes among the Spanish population. When it comes to participatory and representative governance, as assessed by the process scale, most Spaniards choose central positions, with few more preferring the participatory to the representative pole. When it comes to the additional items that tap the three decision-making processes, the means presented above indicate higher support for the general principles (e.g., consultations with citizens) than for the specific mechanisms (e.g., referenda).vii The respondents also seem to favor the participatory principles (i.e., citizen consultation and facilitating debates) slightly more than the representative and expert-based processes. The uniformly high means, however, suggests that the differentiation between the three tested models may not be clear-cut for the public. When it comes to the specific mechanisms, however, respondents see expert consultancy as a better form of decision-making than the participatory and – especially – the representative mechanisms. Lastly, with regard to the SD index, Spain emerges as the most pro-stealth democracy country (compared to the US and Finland, the two other countries where the index has been tested), with only 1% of the population disagreeing with all the four items and 40% expressing some agreement with at least three (compared to 26-27% in the other two countries). As Figure 2 illustrates, this high rating is due the staggering 95% of the respondents agreeing that politicians should stop talking and start acting, an item which does not directly tap any of the three tested decision-making processes. Also, and consistent with the pattern described above, majority of the sample supports expert involvement in decision making, and this percentage is higher than in the other two countries (62% versus 31% in Finland and 30% in the US). [Figure 2 about here]

12

As the first step towards testing the multi-dimensionality of citizen process preferences, we explore the correlations between the items. The analyses confirm the aforementioned conceptual argument regarding the internal consistency of the SD index: the Cronbach’s alpha for the overall index is unsatisfactory.viii Also, the items measuring lack of trust (elected officials should stop talking and that political compromise means selling out) are insignificantly or weakly related with the remaining two SD items and the other variables.ix Exploratory factor analysis further shows that these two items form an independent factor, different from all the other variables included in the analysis. These two items were thus excluded from the CFA analyses. These preliminary analyses also find that the remaining SD items (support for government run by experts and by businessmen) are significantly related to the other indicators of expert-based governance. Also, the process scale is meaningfully associated with all the items that reflect support for the representative or the participatory models, offering a test of its face validity. In all cases the measures that tap preferences for the general decision-making principles are significantly correlated with the items measuring the specific mechanisms to implement these principles, with those respondents who support consulting citizens also supporting referenda, or those wishing that political decisions be left to experts also favoring expert consultancy, for example. [Table 1 about here]

Dimensionality of political decision-making processes We rely on CFA to explore whether the three proposed dimensions of political decisionmaking processes - participatory, representative, and expert – emerge as distinct and accurately

13

represent citizen preferences. To validate measurement models using CFA and address our research question, we test three different configurations of citizen preferences (Jöreskog, 1993). The first one assumes a one factor model with two poles: participation vs. representation. This option implies that the debate about experts’ role in politics is not an independent one and is captured by this single opposition. The second model presumes a two factor solution, with one factor capturing support for participatory processes and the second factor capturing support for representative as well as expert-based governance. Lastly, the third model implies three distinct factors, resembling the three proposed types of political processes (participative, representative and expert-based). In order to contrast the goodness-of-fit of these models, besides Chi2 test, we use other fit measures (Bentler, 1990; Hu and Bentler, 1999). To avoid potential problems due to nonmultinormality of variables, we relied on robust measures of CIF and RMSEA (Satorra and Bentley, 1990).x The models were tested including covariances among all the factors (a saturated model), as an appropriate structure to test measurement models (Bagozzi and Yi, 1988) and to offer information about the relationship between the factors.xi As seen in Table 2, fit measures show that only the third model fits the data and results in a good representation (measurement model) of process preferences among Spanish citizenry (Chi2=248.808, df=28; p < 0.001; CIF=0.945; SRMR= 0.047; RMSA=0.070). This implies that Spanish citizens distinguish between the three decision making processes. As expected, beside the classical differentiation between representative and participative poles, Spaniards also identify a separate factor in the political process, one in which the main principle is expertise. [Table 2 about here]

14

The one-dimensional model shows poor fit indexes, and insignificant factor loadings (Table 3). The three-dimensional model, which identifies the participatory, representative and expert-based governance, has the best good-of-fit measures and in this case all factors loadings are significant (p < .001). The correlations between factors provide some preliminary empirical evidence to address our second question, about how these factors relate with each other. Model 3 shows that there is a clear opposition between participatory and representative processes (covariance equal to -.64). In turn, the factor representing expert-based governance has a (small) positive correlation with both, and its relationship is slightly stronger with the factor representing representative processes (.16) than with the factor indicating participatory processes of political decision-making (.09). In the discussion, we address the meaning and implications of these results. [Table 3 about here]

Discussion Traditionally, political scholars have focused on tensions between empowering citizens by giving them greater say in the political process and representative governance where citizens are activated only periodically and primarily to monitor the democratic process through electoral engagement. More recently, the publication of Stealth Democracy (2002) has extended the debate, popularizing at the public opinion level, the idea of support to expert governance. The present study offers several noteworthy findings that add to the existing evidence on this debate. First, preferences for political decision-making processes are not one-dimensional and cannot be reduced to the opposition between direct citizen engagement and government via elected representatives. On the one hand, as indicated by the strong negative correlation between

15

the representative and participatory factors, the core tension does emerge between participative and representative processes. Yet, these two models are not seen as simple opposites, in that they appear as two separate factors, not as two poles in a single factor. In many cases these two approaches to political decision-making do clash, as it is either the public or elected representatives who have the last word in the process. However, some citizens may support representation while also favoring some direct citizen engagement (or vice versa), especially if these two processes are compared to alternative forms of governance (e.g., expert-based, authoritarian or others). This result is not a mere statistical artifact as it also emerges in alternative CFA models with different items, and with the process scale excluded from the analysis.xii In fact, a similar result emerged in the Finish case (Bengtsson, 2012) further validating the comparative substantive meaning of our findings. The results from both the correlational and the CFA analyses also suggest that the public does not differentiate between low and high intensity participatory processes (Webb et al., 2010), at least when these are presented vis-à-vis different alternatives like representation or expertise. Referenda and assembly based models really represent quite different forms of decision making implying quite different costs and advantages, but they are too little well known by most citizens who simply identify them as two possible ways to put into practice their participatory goals. Second, citizen process preferences are complex, with Spaniards clearly distinguishing between participatory, representative and expert-based governance. This third distinct dimension may capture citizen desire that governments are run by experts and technocrats (i.e., preferences based on the actor who makes the decisions) and/or that political decision-making processes prioritize non-political criteria of efficiency and expertise (i.e., preferences based on the values that guide the process). In fact, the relationships between the three identified factors suggest that

16

who makes political decisions may matter more than the values underlying the process. That is, the reality of citizen process preferences in Spain reflects the idea shown by Figure 1c, in that there is a strong negative correlation between participation and representation and a (weaker) positive correlation between expert-based government and representation. These associations do not point toward a value-centered interpretation reflected in Figure 1.b, as the expert-based dimension is positively correlated with both the participatory and the representative dimensions. In any case, because direct citizen involvement and expert decision making are not negatively correlated, the results may also indicate that citizen preferences are based on the contrast between the existing representative processes and other alternatives, as reflected in Figure 1a, i.e., probably due to general dissatisfaction with the existing representative model (Webb, 2013). These patterns indicate a real debate between those citizens who do not want increased citizen involvement in political decision-making and those who strongly support such involvement. Because our questionnaire was not designed specifically to answer the question as to what moves citizen preferences, these insights must be further validated. Future research that designs survey items to disentangle the role of the different considerations that guide individual support for participatory, representative and expert-based governance would shed light on what citizens like and dislike in these distinct processes. It may be the case that the sociopolitical context in Spain in the early 2011, a few months before the emergence of the Indignados movement and with rising public disaffection and discontent, influenced the results. While we cannot discard potential context effects (as is the case with any investigation), public demands for more participatory processes were not visible at the time the survey was conducted. Also, other studies at distinct time points find similar results (e.g., the European Values Study from 2008 also showed high support for expert governance in

17

Spain). In any case, the idiosyncratic nature of the sociopolitical climate would matter more if our sole goal was to portray citizen preferences for political decision-making processes; yet there are no substantive reasons to expect that this climate would change the relationships between the dimensions of democratic governance, which we aimed to reveal. Also, our results have two methodological implications. Inasmuch as the debate about how policy decisions should be made is increasingly salient, more surveys will likely include measures of citizen preferences for political decision-making processes. Our study suggests that the process scale (Hibbing and Theiss-Morse, 2002) is a parsimonious and versatile instrument that, at least in the Spanish case, accurately captures the central opposition between participation and representation. Inasmuch as this opposition is still important for voters, including the scale in different surveys would allow exploring the attitudinal and behavioral correlates of citizen process preferences. Yet, because this opposition does not fully capture these preferences, items tapping public support for other possible political decision-making processes should also be included, tested, and validated comparatively. Our findings also invite political scientists to extend our results and closely examine citizen process preferences. Using bipolar scales that force respondents to choose between the virtues and limits of representative, participatory and expert-based models as well as designing questions that shed light on what lies behind these choices (the actors involved and/or the values used in decision-making, among others) would offer important information on the complexity of citizen preferences towards political governance. Which processes, actors, and values do citizens want? Is a non-partisan institution like the European Central Bank seen as a successful example of an expert-run body? Who, according to the public, are the efficient experts? For example,

18

would a successful businessman like Silvio Berlusconi be favored over a politician with technical competence and limited partisan attachment like Mario Monti? Although our study naturally cannot address all these questions, it shows that, at the very least, three major dimensions of political decision making processes emerge. For some citizens, the representative model characterized by the US Congress or the German Bundestag seems to be the best possible option. Others would prefer citizen assemblies implemented by the Indignados or the Occupy movements, or the constant multilevel community referenda in Switzerland. Still, for other citizens, the technical and non-partisan governments tested in Italy and Greece would be the preferred option.xiii The three models that have emerged among the Spanish public are present in the real political debates and concrete sociopolitical situations of different European countries. What is still missing are the details on how much support each model receives in various countries and what are, according to the public, the specific advantages and disadvantages of each. Our research has highlighted the multidimensional character of citizen preferences with a dataset that contained more detailed measures than prior surveys on these issues. Further research is needed to fully explore the remaining questions, which are rendered increasingly important by the ongoing sociopolitical events internationally and the role that certain citizen demands for more participatory opportunities and moves towards more technocratic forms of decision-making are playing in them.

19

Appendix: wording of questions measuring support to the three models of public decisionmaking Questions measuring support for citizen participation: 1. We would like you to rate the following ways of making political decisions. Please, use a 0-10 scale where 0 means “is the worst form to make decisions” and 10 means “is the best way to make decisions”: Consult frequently citizens about their opinions 2. We would like you to rate the following ways of making political decisions. Please, use a 0-10 scale where 0 means “is the worst form to make decisions” and 10 means “is the best way to make decisions”: Make easy that people participate and discuss the most important political decisions 3. We would like you to rate the following ways of making political decisions. Please, use a 0-10 scale where 0 means “is the worst form to make decisions” and 10 means “is the best way to make decisions”: Organise referenda frequently 4. We would like you to rate the following ways of making political decisions. Please, use a 0-10 scale where 0 means “is the worst form to make decisions” and 10 means “is the best way to make decisions”: Organise assemblies and meetings so that people can take decisions by themselves Questions measuring support to representation: 1. We would like you to rate the following ways of making political decisions. Please, use a 0-10 scale where 0 means “is the worst form to make decisions” and 10 means “is the best way to make decisions”: Elect the politicians that will take political decisions 2. We would like you to rate the following ways of making political decisions. Please, use a 0-10 scale where 0 means “is the worst form to make decisions” and 10 means “is the best way to make decisions”: Let the government take the decisions 3. We want to know how you would like that political decisions were taken in Spain. In a scale from 0 to 10 where "0" means citizens making all decisions on their own, and "10" means politicians making all the decisions on their own, where would you place yourself? Questions measuring support to expert government: 1. We would like you to rate the following ways of making political decisions. Please, use a 0-10 scale where 0 means “is the worst form to make decisions” and 10 means “is the best way to make decisions”: Allow experts in each topic to make decisions 2. Please, tell me if you agree strongly, agree, disagree or strongly disagree with the following sentences: Our government would run better if decisions were left up to successful business people 3. Please, tell me if you agree strongly, agree, disagree or strongly disagree with the following sentences: Our government would run better if decisions were left up to non-elected, independent experts rather than politicians or the people 4. We would like you to rate the following ways of making political decisions. Please, use a 0-10 scale where 0 means “is the worst form to make decisions” and 10 means “is the best way to make decisions”: Take political decisions through consultations to experts

20

References Anderson, Cameron and Goodyear-Grant. Elizabeth (2010). “Why are highly informed citizens sceptical of referenda?” Electoral Studies. 29(2): 227-238. Bagozzi, Richard.P. and Yi, Youjae. (1988). ‘On the Evaluation of Structural Equation Models’, Journal of the Academy of Marketing Science, 16(1): 74-94. Bengtsson, Asa. (2012). “Citizens perceptions of political processes. A critical evaluation of preference consistency and survey items”, Revista Internacional de Sociología, 70 (5). Bengtsson, Asa and Mattila, Mikko. (2009). “Direct democracy and its critics: Support for direct democracy and “stealth” democracy in Finland”. West European Politics, 32 (5), 10311048. Bentler, Peter. (1990). ‘Comparative fix indexes in structural models’. Psychological Bulletin, 107, 238-246. Bowler, Shaun; Donovan, Todd and Karp, Jeffery. (2007). “Enraged or Engaged? Preferences for Direct Citizen Participation in Affluent Democracies”, Political Research Quarterly 60; 351 Brooks, Stephen; Stasiak, Dorota and Zyro, Tomasz (eds) (2012): Policy expertise in contemporary democracies, Surrey: Ashgate. Budge, Ian (1996): The new challenge of direct democracy. London: Polity Press. Dalton, Russell, Bürklin, Wilhelm and Drummond Andrew (2001). “Public opinion and direct democracy”, Journal of Democracy, 12, 4:141-153. Díaz de Rada, Vidal and Nuñez, Adoración (2008): Estudio de las incidencias en la investigación por encuesta, Madrid: CIS.

21

Donovan, Todd and. Karp, Jeffrey A. (2006). “Popular Support for Direct Democracy”, Party Politics, 12(5): 671-688. Fischer, Frank (1993). “Citizen participation and the democratization of policy expertise: from theoretical inquiry to practical cases”. Policy Sciences 26 (1): 163-187. Fishkin, James (2009): When the people speak, Oxford: Oxford University Press. Fishman, Robert (2004): Democracy’s voices. Social ties and the quality of public life in Spain, Ithaca: Cornell University Press. Font, Joan, Navarro, Clemente, Wojcieszak, Magdalena; Alarcón, Pau (2012): “¿Democracia sigilosa en España? Preferencias de la ciudadanía española sobre las formas de cisión política y sus factores explicativos, Opiniones y Actitudes, 71, Madrid: CIS. Fung, Archon and Wright, Erik Olin (2001): "Deepening democracy: innovations in empowered participatory governance", Politics & Society, 29 (1): 5-41 Gangl, Amy (2007): “Examining citizens’ beliefs that government should run like business”, Public Opinion Quarterly, 71 (4), 661-670. Ganuza, Ernesto; García Espin, Patricia and de Marco, Stefano (2013): “The Political Turn of Citizens: What Does Disaffection Mean in Spain?” Paper presented at the ECPR Joint Sessions, Mainz. Gastil , John . (2000) By Popular Demand: Revitalizing Representative Democracy through Deliberative Elections. Berkeley: University of California Press Hibbing, John and Theiss-Morse, Elisabeth (2002). Stealth Democracy. Cambridge: Cambridge University Press. Hilde, Coffée and Michels, Ank (2012): “Education and support for representative, direct and stealth democracy”, Paper prepared to be presented at the APSA Meeting.

22

Hu, Li-Tze. and Bentler, Peter (1999): ‘Cut of Criteria for Fit Indexes in Covariance Structure Analysis: Conventional Criteria versus New Alternatives’, Structural Equations Modeling, 6(1): 1-55. Hughes, Neil (2011): ‘Young People Took to the Streets and all of a Sudden all of the Political Parties Got Old’: The 15M Movement in Spain”, Social Movement Studies: Journal of Social, Cultural and Political Protest, 10, 4. Jacobs, Lawrence R; Lomax Cook, Fay and Delli Carpini, Michael (2009): Talking together. Public deliberation and political participation in America, Chicago: University of Chicago Press. Joreskog, Karl (1993). ‘Testing structural equations models’, in Bollen, Keneth. and Long, Scot J. (eds.): Testing structural equation models, Newbury Park (CA): Sage, 294-316. Lippmann, Walter (1922). Public Opinion. New York: Harcourt Brace. Martin, Irene and Van Deth, Jan. 2007. “Political interest”, in Van Deth, J; Montero, J.R. and Westholm, A (eds). Citizenship and involvement in European democracies. London: Routledge Morales, Laura. (2009) Joining Political Organisations. Colchester: ECPR Press. Muñoz, Jordi; Rico, Guillem and Anduiza, Eva (2012): “Austerity policies and political involvement: evidence from a panel survey in Spain (2010-11)”, Paper presented at the Workshop "Change in political attitudes: Panels and experiments", UAB. Neblo, Michael et al (2010). “Who Wants To Deliberate—And Why?”, American Political Science Review, 104 (3), 566-583. Papadopoulos, Yannis and Warin, Philippe (2007): “Major findings and paths for research: A concluding note”, European Journal of Political Research, 46: 591-605

23

Roberts, Nancy (2004). “Public Deliberation in an Age of Direct Citizen Participation”. American Review of Public Administration, 34, 315-353 Roberts, Nancy (2001). “Coping with Wicked Problems”. In L. Jones, J. Guthrie & P. Steane (Eds.), International Public Management Reform: Lessons From Experience. London: Elsevier. Satorra, Albert and Bentley, Peter (1990): ‘Model conditions for asymptotic robustness in the analysis of linear relations’. Computational Statistics and Data Analysis, 10: 235-249. Schumpeter, Joseph (1943). Capitalism, socialism and democracy. London: Allen & Unwin. Smith, Graham (2009): Democratic innovations. Cambridge: Cambridge University Press. Tocqueville, Alexis De (2000 [1851]). Democracy in America. Chicago: University of Chicago Press. Van Deth, Jan; Montero, José Ramón and Westholm, Anders (eds). 2007. Citizenship and involvement in European democracies. London: Routledge. Webb, Paul (2013): “Who is willing to participate? Dissatisfied democrats, stealth democrats and populists in the United Kingdom”, European Journal of Political Research, 52: 747-772. Webb, Paul; Bale, Tim and Taggart, Paul (2010): “Deliberative versus parliamentary democracy in the UK: a experimental study”, SEI Working paper 118.

24

Figure 1a. Three models of process preferences: Dissatisfaction with representation Citizens

Representative

Non-conventional by

Experts Figure 1b. Three models of process preferences: experts or values? Politicians

Experts (efficiency)

Politics by

Citizens Figure 1c. Three models of process preferences: which role for citizens? Politicians

Participative by: citizens

Representative by

Experts

25

Figure 2. Support to Stealth Democracy traits in Spain, Finland and the U.S.

Sources: CIS survey 2860, Hibbing and Theiss-Morse (2002) and Bengtsson and Mattila (2009)

26

Table 1. Correlation between process preferences variables Model consult citizens

Model experts

Model participation

Model politicians

Mechanism referenda

Mechanism experts

Mechanism assemblies

Mechanism politicians

Decisions by successful businessmen

Decisions by experts

-0,282 **

0,081 **

-0,294 **

0,194 **

-0,270 **

0,042

-0,278 **

0,426 **

-0,019

0,054 *

1

0,181 **

0,644 **

0,062 **

0,440 **

0,088 **

0,380 **

-0,231 **

0,021

0,044 *

0,181 **

1

0,227 **

0,340 **

-0,002

0,449 **

0,094 **

0,111 **

-0,196 **

-0,278 **

0,644 **

0,227 **

1

0,138 **

0,362 **

0,135 **

0,451 **

-0,259 **

0,020

-0,017

0,062 **

0,340 **

0,138 **

1

-0,057 **

0,228 **

0,017

0,248 **

-0,056 *

-0,069 **

0,440 **

-0,002

0,362 **

-0,057 **

1

0,245 **

0,503 **

-0,229 **

-0,022

0,029

0,088 **

0,449 **

0,135 **

0,228 **

0,245 **

1

0,285 **

0,081 **

-0,170 **

-0,298 **

0,380 **

0,094 **

0,451 **

0,017

0,503 **

0,285 **

1

-0,267 **

-0,030

-0,023

-0,231 **

0,111 **

-0,259 **

0,248 **

-0,229 **

0,081 **

-0,267 **

1

-0,031

0,018

-0,019

0,021

-0,196 **

0,020

-0,056 *

-0,022

-0,170 **

-0,030

-0,031

1

0,343 **

Decisions by experts

0,054 *

0,044 *

-0,278 **

-0,017

-0,069 **

0,029

-0,298 **

-0,023

0,018

0,343 **

1

Less talk, decisions

0,0 73 **

-0,121 **

-0,078 **

-0,1 40 **

-0,043 **

-0,070 **

-0,063 **

-0,119 **

-0,133 **

0,074 **

0,088 **

Principles or compromise

0,057 *

-0,006

-0,0 58**

0,0 02

0,0 16

-0,0 45

-0,016

-0,026

0,132 **

0, 132 **

0,080 **

Process scale

Process scale Model consult citizens Model experts

1

-0,282 **

0,081 **

Model participation

-0,294 **

Model politicians

0,194 **

Mechanism referenda

-0,270 **

Mechanism experts

0,042

Mechanism assemblies

-0,278 **

Mechanism politicians

0,426 **

Decisions by successful businessmen

** p < .01; * p < .05. Source: CIS 2860.

27

Less talk, decisions

Principle or compromi

0,0 73 **

0,057 *

-0,121 **

-0,006

-0,078 **

-0,0 5 8**

-0,140 **

0,0 02

-0,043 **

0,0 16

-0,070 **

-0,0 45

-0,063 **

-0,016

-0,119 **

-0,026

-0,133 **

0,063 *

0, 088 **

0,132 **

0, 142 **

0,080 **

1

0,074**

0,074 **

1

Table 2. Citizen preferences for political decisions processes. Confirmatory factor analysis: fit indexes Fit indexes Chi2 (1) Degrees of freedom Significance AIC CAIC FIT (>.90) CFI (>.90) SRMR (