Representatives Attitudes towards Citizen Protests: The Impact of Ideology, Parliamentary Position, and Experiences

Transcription

1 Representatives Attitudes towards Citizen Protests: The Impact of Ideology, Parliamentary Position, and Experiences Mikael Gilljam, Mikael Persson and David Karlsson CONTACT CORRESPONDING AUTHOR: Mikael Persson Department of Political Science University of Gothenburg Box Gothenburg Sweden Telephone: BIOGRAPHICAL NOTE: Mikael Gilljam, PhD, professor at the Department of Political Science, University of Gothenburg, Sweden. Mikael Persson, PhD candidate at the Department of Political Science, University of Gothenburg, Sweden. David Karlsson, PhD, senior lecturer at the School of Public Administration, University of Gothenburg, Sweden.

2 What affects political representatives attitudes towards citizen protests? We test the impact of the political representatives left-right ideology, parliamentary position and experiences of citizen protests. Using data from a pioneering survey covering all local political representatives in Sweden (n=9101, response rate 70 percent), we examine attitudes towards controversial non-institutionalized forms of citizen protests. The results show that representatives farthest to the right show considerably lower protest acceptance than those farthest to the left. Representatives in office show significantly lower levels of acceptance than those in opposition. Finally, the more protest experiences, the higher protest acceptance. The paper concludes by discussing how these results affect our general views about leader responsiveness in off-election democracy. Keywords: Citizen protests, Off-election democracy, Political representatives, Leader responsiveness, Protest acceptance. 2

3 INTRODUCTION Enhancing democracy by increasing citizen participation is of primary concern among politicians as well as political scientists. When distrust in democracy is growing and representatives tend to lose contact with citizens, increasing political participation in noninstitutionalized forms is pointed out as a remedy (e.g. Zukin et al. 2006; Klingemann and Fuchs 1995; Stolle et al. 2005). But do the representatives really accept these kinds of non-institutionalized forms of citizen protests? In this paper we look closer at a prerequisite for what Hanna Pitkin (1967) calls responsiveness of political leaders towards citizens attempts to influence politics: Do the political representatives accept the chosen mode of citizen participation. If they do not even do that they are not very likely to act responsively towards the citizens opinions. Our aim is to contribute to the research on processes and mechanisms in what we refer to as off-election democracy. We define off-election democracy as communication, or attempts to communication, between citizens and political representatives through channels other than elections (cf. Brothén and Gilljam 2006:193-5). More precisely, we focus on a subcategory of political participation in off-election democracy citizens protest activities and a particular precondition of leader responsiveness the political representatives attitudes towards these protests. The few existing studies on leader responsiveness towards citizen participation have mainly dealt with the impact of different strategies and protest characteristics (e.g. Uba 2005) and the impact of the participants social composition on leader responsiveness (e.g. Verba and Nie 1972; Schumaker 1975; Schumaker and Getter 1977; Wohlgemuth 2006). Differing from these studies we shift focus from the citizens to the representatives 3

4 and investigate whether three fundamental factors normally steering human behavior values, self-interest, and habits and experiences affect the representatives attitudes towards citizen protests. More precisely we test the impact of the representatives leftright ideology, parliamentary position and experiences of citizen protests on their protest acceptance. Research on political participation has mainly dealt with the input side of participation by focusing on questions such as the causes of political participation (e.g. Verba, Schlozman and Brady 1995; Wolfinger and Rosenstone 1980) and the changes in modes of participation over time (Zukin et al. 2006, Dalton 2007) and in different contexts (van Deth et al. 2007). However, the output side of participation i.e. the effects of participation has not got enough attention among scholars in the field (cf. Verba and Nie 1972 ch. 17; Teorell 2006). This study is an effort to cure this imbalance by focusing on a more or less blind spot in research on the output side of political participation: How different characteristics of the representatives affect their attitudes towards citizen protests. Whether representatives values, self-interest, and experiences affect their attitudes towards citizen protests is of central importance for understanding the relationship between representatives and citizens in off-election democracy. If some representatives are more responsive than other ones it implies that the success of citizen protests, even if they are of the same kind, may differ depending on whom the targeted representatives are. This, in turn, means that some citizens may influence public matters more than other ones, which conflicts with the democratic principle of political equality. Due to the 4

5 varying level of responsiveness among representatives it is unclear whether citizens in off-election democracy influence public matters equally or not. In the following, three hypotheses on what affects representatives attitudes towards citizen protests are tested: We examine the impact of political ideology, parliamentary position and experiences of citizen protests. We use data from a pioneering internet-based survey covering all local representatives in Sweden. The results show that representatives with a leftist ideology are more positive towards citizen protests than representatives with a rightist ideology. Furthermore, representatives belonging to the parliamentary majority show significantly lower levels of acceptance than opposition representatives. Moreover, we find a positive effect of the representatives own experiences of citizen protest. In addition, we find that the effect of experiences is conditional on ideology and parliamentary position. Positive effects of protest experiences are only found among opposition and leftist representatives. Next, we present the three hypotheses on what affects political representatives attitudes towards citizen protests. Then we present our data, analytical strategies and empirical analysis. In the concluding section we discuss the implications of the findings for our understanding of protests acceptance and leader responsiveness in off-election democracy. THEORY AND HYPOTHESES Political participation in general and citizen protests in particular may be perceived as a dilemma for political representatives. On the one hand, the representatives should act 5

6 responsively towards opinions among the citizens; citizen participation and protests articulate opinions about issues considered important by the citizens. On the other hand, the messages put forward by citizen participation and protests may conflict with the political alignment and the election mandate of the representatives and may thus complicate the implementation of their political program. Taking these options into account we could either expect that representatives are positive and accept protests, since it enables them to act responsively towards the citizens opinions, or that the representatives do not accept citizens protests since that would interfere with their decision-making. In all likelihood, however, representatives vary in their acceptance of citizen protests, and this variation may be explained by three fundamental factors normally steering human behavior: values, self-interest, and habits and experiences. Therefore, in this particular context, we use left-right ideology, parliamentary position and experiences of citizen protests as explanatory factors. 1 THE IDEOLOGY HYPOTHESIS It is reasonable to expect that political representatives normative ideological views on politics and democracy influence their attitudes towards citizen protests. In Europe, political parties at the left and the centre right have historically different traditions concerning their modes of participation. The working class movement has a long tradition of bringing out their political messages by means of, for example, public mass demonstrations. Centre right political parties lack this tradition and have relatively few mass protests in their history. The explanation for this is of course ideological. On the one hand, citizens with a leftist ideology have historically tried to bring about social change and have thus had strong reasons to protest. The centre right parties, on the other hand, have tried to preserve the existing structures and have thus had fewer reasons to protest 6

7 (e.g. Lewin 1988). Furthermore, recent comparative research on protest behavior in a large number of countries shows that protesting citizens mostly have a leftist ideology (Dalton, Sickle and Weldon 2009). Thus: (H1) The further to the left the representatives are, the more likely they are to accept citizen protests. If this hypothesis is proven valid it indicates, interestingly, that citizen protests may receive different responsiveness depending on the left-right composition of the local parliament and on the ideological positions of the representatives targeted. THE PARLIAMENTARY POSITION HYPOTHESIS The second main factor that we believe affects attitudes towards citizen protests is whether political representatives are in office or in opposition. Obviously, for reasons of self-interest, it seems easier to accept protesting citizens when you are in opposition without responsibility for decisions made. Moreover, citizen protests against decisions taken by the majority are likely to correspond to the views of the opposition. Furthermore, citizen protests interfere with the possibilities for representatives in office to make and implement decisions. While there are, to our knowledge, no studies on how attitudes towards citizen protests are affected by representatives parliamentary positions, there is some evidence that representatives self-interest affect their views on politics and democracy. First, according to John May s law, party top leaders hold less extreme positions than nonleaders (May 1973). Top leaders need to take full political responsibility, and, faced with 7

8 brute reality, cannot afford to make too extreme decisions. Since they need to care both about being re-elected and about political stability, they tend towards the ideological centre. Second, Bowler, Donovan and Karp (2006; see also Bowler, Donovan and Karp 2002) show that election winners show less support for constitutional reforms than election losers. Especially interesting is their finding that self-interest has a stronger impact than ideology and values. We apply this argument to our case by testing whether representatives in office are less likely to accept citizen activities that hinders and complicates the democratic process than those in opposition. Representatives in office simply want the democratic process to run smoothly. Those in opposition, on the contrary, like to see their competitors activities interrupted and therefore appreciate citizens expressing political views that conflict with the majority view. The reason for the opposition s protest acceptance may also be just malicious pleasure. Hence: (H2) Political representatives in majority position show lower levels of protest acceptance than opposition representatives. If this hypothesis is correct the situation for protesting citizens is, at least in the short run, somewhat disappointing: When protesting in non-institutionalized forms citizens get most responsiveness from the less powerful representatives in opposition and least from those in the ruling majority. In addition, to test the generalizability of the effect of parliamentary position, we will examine whether this effect is dependent on the representatives ideological position: Is 8

9 the effect of parliamentary position equally strong among representatives regardless of their ideology? THE EXPERIENCE HYPOTHESIS The third main factor that can affect attitudes towards citizen protests is the representatives experiences of citizen protests in their own municipality. Experiences may however either increase or decrease protest acceptance. On the one hand, after having experienced citizen protests the representatives may consider them a legitimate mode of participation. More contacts between citizens and representatives, even in the form of protest activities, could make the representatives more appreciative about citizen communication. Also, public protests may, in the eyes of the representatives, constitute an effective way of getting informed about citizen preferences on important issues. On the other hand, however, experiences of citizen protests that interfere with the formal political process can make the representatives less accepting. Still, the character of the protests is relevant. Extreme, violent riots, for instance, are not likely to make the representatives more accepting. Such activities are, however, rare exceptions in the Swedish context, and thus we hypothesize that protest experiences in general increase protest acceptance. In their seminal study on leader responsiveness, Sidney Verba and Norman Nie (1972) showed a general positive curve linear relationship between levels of citizen participation and leader responsiveness. They found that the higher the level of citizen participation, the more the leaders preferences concurred with citizens preferences. Moreover, in an attempt to replicate the study of Verba and Nie in the Swedish context, Daniel 9

10 Wohlgemuth found that citizen participation in non-electoral channels had a positive effect on policy agreement between citizens and representatives (Wohlgemuth 2006). It is, however, not clear if these findings are extendable to protest acceptance, in particular since we are not meausuring preference congruence but merely attitudes towards the modes of protests. Nevertheless, drawing on Verba and Nie (1972) as well as Wohlgemuth (2006), we find it reasonable to assume that the higher the amount of citizen protests in their own municipality, the more accepting the representatives will be. Looking more broadly at political socialization research we find a more sophisticated argument in support of a positive experience effect. According to the theory of adaptive preferences, individuals tend to accept things as they are. This is the sour grape phenomenon: Individuals who are content with what they have show higher well-being than individuals who yearn after things they cannot get anyway. If one cannot get the sweet grapes, one imagines that they are sour anyway: Sour grapes, said the fox, when he could not reach them (e.g. Elster 1983). This psychological mechanism has been applied to explain phenomena as diverse as Indian women complying with their subordinate position (Nussbaum 2001; Sen 1995) and the shaping of consumption preferences (Halleröd 2006). We find it reasonable to assume that also political representatives attitudes are formed in this way. When confronted with citizen protests, political representatives begin to accept them as legitimate rather than sticking to their disappointment. Thus, profiting from past research on leader responsiveness as well as the theory of adaptive preferences, we propose that representatives experiencing citizen protests accept them as a more legitimate form of citizen participation. Hence: 10

11 (H3) The higher the level of protest experience, the higher the level of protest acceptance. Furthermore, by explicitly modeling the interaction between experience and ideology as well as between experience and parliamentary position, we try to find the degree to which the experience hypothesis may be generalized. More exactly we test whether the effect of protest experiences is the same for left- and right-wing representatives, and whether the effects of protest experiences are the same for representatives with different parliamentary positions (majority or opposition). DATA AND MODELLING STRATEGY We use data from a pioneering survey of all local representatives in Sweden (9101 responses, response rate 70 percent). 2 Sweden consists of 290 municipalities that are the entities for local government in Sweden. During the general elections that take place every fourth year elections are simultaneously held at the national level (riksdag), the county level (landsting/region) and the municipality level (kommun). Our sample covers representatives at the local parliaments in municipalities. Swedish municipalities are responsible for, among other things, childcare and education (from pre-school to upper secondary schools), social service and elderly care. In fact, the municipalities carry out most of Sweden s extensive social welfare services. Hence, most of the political decisions that directly affect peoples daily life are made at the municipal level. For that reason citizen protests in Sweden are mostly taking place at the municipal level. Consequently, local representatives are the most frequent targets of citizen protests and they are therefore those most interesting to study in this context. 11

12 The dependent variable in this study attitudes towards citizen protests is measured with the following question: If you imagine a situation in which there is a political proposal on whether to close a public school, how much acceptance do you have for the following forms of protests?. Respondents were asked to rate their level of acceptance for a number of different forms of protests on a scale from 0 (no acceptance whatsoever) to 10 (full acceptance). It is important to note that the dependent variable does not measure the influence of protests or the representatives responsiveness. Rather, it measures which forms of citizen protests the representatives find acceptable. All forms of protests acceptance items deal with citizens activities in off-election democracy contexts and include unauthorized public demonstrations, parents refusing to send their children to school, occupation of the school building, illegal strikes among teachers, citizens approaching political representatives in their homes, citizens who speak out without permission at the municipal council, and citizens who un-contradicted attack political representatives in local media. All of these modes of protest are noninstitutionalized, controversial, and include violations of norms, civilian contracts and laws. The reason for using these modes of protests is that we want to capture the variation of attitudes among representatives. If we had just asked about uncontroversial forms of protests such as collecting names or wearing a badge we would have risked getting only positive responses from all representatives. Probably all representatives accept wellbehaved citizens who communicate politely, but by introducing controversial forms of protests, interesting variations in representatives attitudes towards citizen protests can hopefully be found. Table 1 shows the Pearson correlations between the items and table 2 shows their means and standard deviations. We use an additive index of the seven protest 12

13 items as the dependent variable, rescaled to vary between 0 (low acceptance) and 10 (high acceptance), with a Cronbach alpha value of ************** TABLE 1, 2 AND 3 ABOUT HERE ************** Now let us turn to the independent variables (their means and standard deviations are shown in table 3). First, for left-right ideological position, respondents were asked to place their position on a scale ranging from 0 (the left extreme) via 5 (neither left nor right) to 10 (the right extreme). 4 Respondents position in parliament is measured by a dummy variable taking the value 1 if the respondent represents a party holding the majority of the seats in the local parliament or is part of a majority coalition and 0 if the respondent represents a party in opposition. Experience of citizen protests is measured by a question on how respondents perceive the amount of protests in their own municipality compared to other municipalities. The item has a five-point scale from 1 (low level of protest) to 5 (high level of protest). 5 In the analyses presented all independent variables are recoded to vary between 0 and 1. 6 Since the data set has a nested structure in which respondents are clustered in 290 different local governments, a multilevel regression model is used in order to secure accurate estimates. For that reason, we begin by estimating a random intercept only model (null model) to investigate whether the variance at the municipal level is significant and multilevel modeling therefore needed. 13

14 Model 0: Protest acceptance ij = α + u j + e ij Results from model 0 in table 4 show that the variance at the municipal level is small but statistically significant. Thus, if we should disregard this dependency the results would be biased (cf. Goldstein 1995; Hox 2002; Snijders and Bosker 1999). Hence, multilevel regression techniques will be used in the following analyses. RESULTS We start by evaluating the hypothesis on the impact of ideology by estimating the effects of left-right ideology on protest acceptance (H1). We also include a random coefficient for ideology to allow the effect to vary across municipalities: Model 1: Protest acceptance ij = α + β 1 Left-right ideology ij + u 0j + u 1j Left-right ideology ij + e ij **************** TABLE 4 ABOUT HERE **************** As expected, results from model 1 in table 4 show that left-right ideology significantly affects protest acceptance. We see that representatives to the right have about one and a half unit less of protest acceptance than representatives to the left. So, indeed, there is a significantly higher level of acceptance among the political representatives to the left. Hence, H1 is confirmed. 7 14

15 Next, we turn to evaluate the main effect of parliamentary position (H2): Model 2: Protest acceptance ij = α + β 1 Parliamentary position ij + u 0j + u 1j Parliamentary position ij + e ij Results from model 2 in table 4 show that H2 is supported. Being a representative in the majority of the local parliament decreases protest acceptance with 0.7 units on the protest acceptance scale. Being in office apparently changes political representatives views on protesting citizens. 8 We evaluate this finding further by investigating whether the effects of ideology reduces the effect of being in office, taking both independent variables into account: Model 2a: Protest acceptance ij = α + β 1 Left-right ideology ij + β 2 Parliamentary position ij + u 0j + u 1j Parliamentary position ij + e ij Interestingly, results from model 2a in table 4 show that the significant effect of parliamentary position remains, but is slightly reduced, when controlling for left-right ideology. In fact, under control for each other, both ideology and parliamentary position prove to have significant impact on protest acceptance. The effect of parliamentary position is an interesting finding. It tells us that politicians in majority and opposition have somewhat different self-interests. The obvious follow up question is whether this difference is a general phenomenon or if the effect of parliamentary position is unequally strong among representatives with different 15

16 ideologies. To answer this question we model the interaction between left-right ideology and parliamentary position: Model 2b: Protest acceptance ij = α + β 1 Left-right ideology 1ij + β 2 Parliamentary position 2ij + β 3 Left-right ideology 1ij X Parliamentary position 2ij + u 0j + u 1j Parliamentary position 2ij + e ij According to results from this test, the effect of parliamentary position seems to be a general phenomenon. The interaction term between ideology and parliamentary position (model 2b in table 4) is not significant (p=.12). Apparently, being in a majority position has about the same negative effect on protest acceptance for leftist and rightist representatives. Representatives in opposition show higher levels of protest acceptance regardless of their position on the left right-scale. We now move forward to evaluate whether experiences of protests increase protest acceptance (H3): Model 3: Protest acceptance ij = α + β 1 Experience ij + u 0j + u 1j Experience 1ij + e ij **************** TABLE 5 ABOUT HERE **************** Results from model 3 in table 5 confirm H3 by a considerable, positive coefficient for the experience variable. As hypothesized protest experience increases protest acceptance: 16

17 The more experience, the more acceptance. 9 However, we need to control for the other two independent variables to evaluate the strength of this effect. We therefore estimate a full model including all three independent variables: Model 3a: Protest acceptance ij = α + β 1 Parliamentary position 1ij + β 2 Experience 2ij + β 3 Left-right ideology 3ij + u 0j + u 1j Parliamentary position 1ij + u 2j Experience 2ij + e ij The results from the full model (model 3a in table 5) show that the significant effect of experiencing protests is somewhat reduced when controlling for ideology and parliamentary position. It is interesting, though, that the effects of ideology and parliamentary position in model 3a remain almost unchanged when adding experience. Thus, ideology, parliamentary position, as well as protest experience have significant effects on protest acceptance when controlled for each other. There is a significant random slope variance of protest experience signifying that the effect of experience varies between municipalities, which, in turn, indicates that the variation in the effect of experience may be due to interaction with other explanatory variables. Therefore, we test the generalizability of the effect of experience by modeling the interactive relationship between experience and the two other independent variables respectively. First, in order to evaluate whether the experience effect is equally strong across groups with different ideologies, we test the interaction between experience and ideology. Second, in order to estimate whether representatives in opposition and majority react in the same way when experiencing citizen protests, we test the interaction between experience and parliamentary position. We begin by testing the interaction between experience and ideology: 17

18 Model 3b: Protest acceptance ij = α + β 1 Experience 1ij + β 2 Left-right ideology 2ij + β 3 Experience 1ij X Left-right ideology 2ij + u 0j + u 1j Experience 1ij + e ij The results in model 3b in table 5 show that the interaction term is significant at the 95 per cent level. 10 Interestingly, the experiences of protests affect rightist and leftist representatives differently. Among the leftists those with more experiences show higher levels of protest acceptance, but among the rightists we see no such difference. Hence, the protest experience effect on protest acceptance is significant only among representatives with a leftist ideology. We proceed by testing whether the effect of protest experience differs between representatives in different parliamentary positions: Model 3c: Protest acceptance ij = α + β 1 Experience 1ij + β 2 Parliamentary Position 2ij + β 3 Experience 1ij X Parliamentary Position 2ij + u 0j + u 1j Experience 1ij + e ij Results from model 3c in table 5 indicate that there is a significant interaction between protest experiences and parliamentary position at the 95 percent confidence level. The effect of protest experience is stronger for opposition representatives. While representatives in office are relatively unaffected by their experiences, representatives in opposition show much more acceptance when they have experienced high levels of protests. 18

19 CONCLUSIONS AND DISCUSSION This article evaluates three explanations for political representatives attitudes towards citizen protest. We find distinct and significant effects of left-right ideology, parliamentary position and experiences of citizen protests: First, representatives to the far most right show about one and a half unit less protest acceptance than do the representatives farthest to the left. Second, representatives in office show significantly lower levels of acceptance than do representatives in opposition. Third, representatives protest experience in their own municipality has a significant effect on protest acceptance. However, the effect of protest experiences is conditional on both ideology and parliamentary position: Representatives with a leftist ideology and representatives in opposition are to a considerable extent positively affected by their own protest experience, whereas those with rightist ideology and those in majority are unaffected. Which, then, are the wider consequences of these results for our ideas about leader responsiveness and communication between citizens and political representatives in offelection democracy? Our results are essential since they reveal that the principle of equal political influence is honored quite differently in democratic elections and in off-election democracy. In elections, political equality is strictly regulated by the principle one person one vote. In off-election democracy, equivalent citizen protests might receive different levels of responsiveness from the representatives depending on who the representatives are. Furthermore, our results indisputably prove that representatives with a leftist ideology accept protests more than representatives to the right. Does this imply that leftist representatives are more responsive towards citizens than representatives with a rightist 19

20 ideology? Not necessarily. At least two interpretations are possible. On the one hand, representatives with different ideologies may, while being equally responsive towards citizens, still communicate with them differently. While leftist representatives show responsiveness to the protesting citizens, representatives with a rightist ideology may be responsive in relation to citizens who use other means of communication. On the other hand, representatives with a rightist ideology may believe that communication in offelection should be limited and therefore, in fact, be less responsive towards citizens. Moreover, our results show that parliamentary position is a key factor in explaining representatives attitudes towards citizen protests. This finding echoes the results presented by Bowler, Donovan and Karp (2006) showing that self-interest is a central factor in explaining representatives attitudes towards institutional change. Yet, the exact nature of the causal mechanism behind the parliamentary position effect in our case is still unclear. Is the effect mainly caused by the mere self-interest of the majority representatives, or is it mainly caused by the opposition s malicious pleasure seeking? In order to answer this question we would need panel data following political representatives going from majority to opposition, and vice versa. It is, nevertheless, striking that citizens, when using the most controversial forms of political participation, get most sympathy from the powerless representatives, and least from the powerful ones. Actually, the effect of parliamentary position on representatives protest acceptance raises a similar question as the effect of ideology: Are majority representatives less responsive towards citizens in general, or do they just communicate differently with them in offelection democracy? 20

21 Finally, the effect of protest experiences on attitudes towards citizen protests also has important implications for our ideas about the representatives responsiveness. Will higher levels of acceptance among political representatives who have experienced protests encourage more citizens to more non-institutionalized protests? And will that lead to a rising spiral of this kind of citizen protests? For the moment, however, the questions on how democracy functions that our results raise are open and can only be empirically settled. Further studies on the effects of citizen protests in off-election democracy could give us the answers. NOTES 1 Self-interest, values and experiences/habits can be seen as the three central factors explaining human behaviour in social science. In political science research on attitudes and behaviour, the impact of self-interest is stressed by the rational choice approach (e.g. Downs 1957); the importance of values and ideologies are manifested in numerous studies demonstrating the effects of e.g. the left-right and liberal-conservative dimensions (e.g. Klingemann and Inglehart 1976; van der Eijk et al. 2005); the impact of experiences and habits are most notably expressed by research on party identification (e.g. Campbell et al. 1954; 1960). The same three factors are also central in other subfields of political science. In international relations, realists focus on the self-interest of actors (e.g. Machiavelli 1988), while liberalists stress the importance of ideological preferences (e.g. Doyle 1986), and constructivists focus on collective identities shaped by actors experiences (e.g. Wendt 1992). In public administration research, one important field focuses on the independent role of ideas (e.g. Odell 1982; Goldstein 1993), while 21

22 pluralists like Bentley (1967) stress the impact of self-interest and power, and institutionalists focus on how habits and experiences within institutions shape behaviour (e.g. March and Olsen 1989). 2 The response rate, 69.8 percent, equals responses out of the non-vacant seats in the Swedish municipalities The non-responses are equally distributed geographically. The distribution of representatives from different parties in the survey closely resembles the real distribution: there is a slight overrepresentation of Social Democrats, 38.3 percent in the survey compared to the true value of 37.0 percent; for all other parties the difference between the survey and the true values are less than one percentage point. As regards gender there are 42.0 percent women and 58.0 percent men in the survey, which is insignificantly different from the actual distribution of 42.3 percent women and 57.7 percent men. However, the survey is slightly less representative when it comes to age; representatives over fifty years are somewhat overrepresented (64.0 percent in the sample compared to the true value of 59.7 percent). 3 Using an index based on factor-scores from a principal component analysis as the dependent variable does not significantly alter the results. 4 Question wording: Sometimes it is said that political opinions can be placed on a leftright scale. Where do you place yourself on such a left-right scale? Scale: 0 (left) to 10 (right). 5 Question wording: What is your estimation regarding how much citizen protests that occurs in your municipality? 1: My municipality is one of the municipalities in Sweden in which it occurs most protests. 22

23 2: My municipality has a rate of protest above the mean in Sweden, but do not belong to those that have the highest level of protest. 3: My municipality has a rate of protest around mean in Sweden. 4: My municipality has a rate of protest below the mean in Sweden, but do not belong to those that have the lowest level of protest. 5: My municipality is one of the municipalities in Sweden in which it occurs lest protests. In the analyses the scale is reversed so that higher values are equal to higher levels of protests. 6 There are, of course, other characteristics that might affect protest acceptance such as age and gender: Young and female representatives could perhaps be expected to have a more open-minded attitude towards non-institutionalized form of political participation. Additional analyses show that age has a very small but significant effect: older representatives are more negative towards protests than younger. However, when adding age as a control it does not significantly alter the effects of the other independent variables in the models. Gender does not have any significant effect on protest acceptance whatsoever. 7 The random slope variance for ideology is not significant, and will therefore not be included in the forthcoming models. 8 The random slope variance for parliamentary position is significant, and will therefore be included in the forthcoming models. 9 One might argue that an aggregated measure of protest experience at the municipal level is a better measure of protest experience. However, such a measure is not necessarily closer to the true amount of protests in the real world since it also draws on measures of 23

24 perceptions. Naturally, such a measure has a narrower spread and, for that reason probably, no significant effect on protest acceptance, neither in a bivariate model nor in a model with control for ideology and parliamentary position. However, when interacting the aggregate measure of experience with ideology and parliamentary position the interaction terms are significant and have the same signs as when using the individual level measure of experience. 10 The interaction term is significant under control for parliamentary position as well. 24

25 REFERENCES Barber, B. (1984). Strong Democracy. Berkeley: University of California Press. Bentley, A.F. (1967) The Process of Government Cambridge: Belknap Press. Bowler, S., Donovan T. & Karp J. (2006). Why Politicians Like Electoral Institutions: Self-Interest, Values, or Ideology? Journal of Politics 68: Bowler, S., Donovan T. & Karp J. (2002). When Might Institutions Change? Elite Support for Direct Democracy in Three Nations. Political Research Quarterly 55: Bäck, H. & Gilljam, M. (eds.) (2006). Valets mekanismer (Mechanisms of Elections). Malmö: Liber. Brothén, M. & Gilljam, M. (2006). Mellanvalsdemokrati (Off-Election Democracy). In Bäck, H. & Gilljam, M. (eds.) Valets mekanismer (Mechanisms of Elections). Malmö: Liber. Campbell, A., Convere, P., Miller, W.E., & Stokes, D. (1960) The American Voter New York: Wiley. Campbell, A., Gurin, G. & Miller, W. E. (1954) The Voter Decides Evanston: Row, Peterson & Company.

26 Cronin, T. (1989). Direct Democracy. The Politics of Initiative, Referendum, and Recall. Cambridge: Harvard University Press. van Deth, J., Montero, J. R. & Westholm, A. (eds.) (2007). Citizenship and Involvement in European Democracies. A Comparative Analysis. New York: Routledge. Dalton, R., van Sickle, A. & Weldon, S. (2009). The Individual-Institutional Nexus of Protest Behaviour. British Journal of Political Science 40: Dalton, R. (2008). Citizenship Norms and the Expansion of Political Participation. Political Studies 56: Doyle, M.W. (1986) Liberalism and World Politics The American Political Science Review 80: Downs, A. (1957) An Economic Theory of Democracy New York: Harper & Row. Eijk, C. van der, Schmitt, H. & Binder, T. (2005) Left-Right Orientations and Party Choice, in J. Thomassen (ed), The European Voter: A Comparatice Study of Modern Democracies Oxford: Oxford University Press, Esaiasson, P. & Holmberg, S. (1996). Representation from Above: Members of Parliament and Representative Democracy in Sweden. Aldershot: Dartmouth. 26

27 Eulau, H., Wahlke, J., Buchanan, W. & Ferguson, L. (1959). The Role of the Representative. Some Empirical Observations on the Theory of Edmund Burke. American Political Science Review 53: Goldstein, J. (1993) Ideas, Interests, and American Trade Policy. Ithaca: Cornell University Press. Halleröd, B. (2006). Sour Grapes: Relative Deprivation, Adaptive Preferences and the Measurement of Poverty. Journal of Social Policy 35: Klingemann, H. D. and Inglehart, R. (1976) Party Identification, Ideological Preference and the Left-Right Dimension Among Mass Public, in Budge, I. Crewe, I., & Farlie, D. (eds), Party Identification and Beyond Lewin, L. (1988). Ideology and Strategy. A Century of Swedish Politics. Cambridge: Cambridge University Press. Machiavelli, N. (1988) The Prince Cambridge: Cambridge University Press. March, J.G. & Olsen, J.P. (1989) Rediscovering institutions: the organizational basis of politics New York: The Free Press. May, J. (1973). Opinion Structure of Political Parties: The Special Law of Curvilinear Disparity. Political Studies 21:

28 Nussbaum, M. C. (2001). Symposium on Amartya Sen s Philosophy: Adaptive Preferences and Women s Options. Economics and Philosophy 17: Odell, J. S. (1982) U.S. International Monetary Policy Princeton: Princeton University Press. Pateman, C. (1970). Participation and Democratic Theory. London: Cambridge University Press. Pitkin, H. (1967). The Concept of Representation. Berkeley: University of California Press. Przeworski, A. (1999). Minimalist Conception of Democracy: A Defense. In Shapiro, I. & Hacker-Gordon, C. (eds.) Democracy s Value. Cambridge: Cambridge University Press. Sen, A. K. (1995). Gender Inequality and Theories of Justice. In M. C. Nussbaum & J. Glover (eds.) Women, Culture, and Development. Oxford: Clarendon Press. Snijders, T. & Bosker, R. (1999). Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modeling. London: Sage. Schumaker, P. (1975). Policy Responsiveness to Protest-Group Demands. The Journal of Politics 37:

29 Schumaker P. & Russell G. (1977). Responsiveness Bias in 51 American Communities. American Journal of Political Science 21: Schumpeter, J. A. (1994) [1942]. Capitalism, Socialism and Democracy. London: Routledge. Stolle, D., Hooghe, M. & Micheletti, M. (2005) Politics in the Supermarket: Political Consumerism as a Form of Political Participation. International Political Science Review 26: Teorell, J. (2006). Political Participation and Three Theories of Democracy: A Research Inventory and Agenda. European Journal of Political Research 45: Uba, K. (2005). Political Protest and Policy Change: The Direct Impacts of Indian Anti- Privatization Mobilisation, Mobilization. An International Journal 10: Wahlke, J., Eulau, H., Buchanan, W. & Ferguson, L. (1962). The Legislative System. Explorations in Legislative Behavior. New York: Wiley & Sons. Wendt, A. (1992) Anarchy is what States Make of it: The Social Construction of Power Politics, International Organization 46: 2,

30 Verba, S. & Norman, N. (1972). Participation in America: Political Democracy and Social Equality. New York: Harper & Row. Wolfinger, R. & Rosenstone, S. (1980). Who Votes? Yale: University Press. Wohlgemuth, D. (2006). Den responsiva demokratin? Effekter av medborgarnas delaktighet i den lokala demokratin. University of Uppsala: Department of Government. Zukin, C., Keeter, S., Andolina, M., Jenkins, K. & Delli Carpini, M. X. (2006). A New Engagement? Political Participation, Civic Life, and the Changing American Citizen. New York: Oxford University Press. 30

31 Table 1. Correlations between items in the protest acceptance index, Pearson Correlations. Speaking Refuse sending children to Occupation of Strike among Visit politicians without permission at Demonstration school the school teachers in their homes public meetings Refuse sending children to school.50 - Occupation of the school Strike among teachers Visit politicians in their homes Speaking without permission at public meetings Protest in mass media Table 2. Summary of protest acceptance items. Mean Standard error Min Max n Demonstration Refuse sending children to school Occupation of the school Strike among teachers Visit politicians in their homes Speaking without permission at public meetings Protest in mass media Table 3. Summary of independent variables. Mean Standard error Min Max n Left-right ideology Parliamentary position Protest experience Comment: The table show the original values from the raw data, in the models the variables are standardized to vary between 0 and 1. 31

32 Table 4. Multilevel models of effects on protest acceptance (two levels), maximum likelihood estimation Model 0: Model 1: Varying intercept Effects of only model Ideology Model 2: Effects of parliamentary position Model 2a: Effects of parliamentary position and ideology Model 2b: Interaction between parliamentary position and Left- Right Ideology FIXED PART: Left-Right Ideology (0= left, 1= right) Parliamentary position (0= opposition, 1=majority) Ideology X Parliamentary Position *** *** *** -.673*** -.566*** -.682***.249 Intercept 2.86*** 3.55*** 3.234*** 3.784*** 3.840*** RANDOM PART: Municipality level variance Individual level variance Random slope variance: Left-Right Ideology Random slope variance: Parliamentary position.037**.029*.032**.032**.035** 4.180*** 3.968*** 4.027*** 3.891*** 3.887*** *** Number of municipalities Number of individuals Bayesian Information Criteria Comment: * p<.10 ** p<.05 *** p<.01 32

33 Table 5. Multilevel models of effects on protest acceptance (two levels), maximum likelihood estimation Model 3: Effects of experience Model 3a: Effects of ideology, parliamentary position and experience Model 3b: Interaction between Experience and Left- Right Ideology Model 3c: Interaction between Parliamentary Position and Experience FIXED PART: Left-Right Ideology (0= left, 1= right) Parliamentary position (0= opposition, 1=majority) Experience of protests (0= low, 1= high) Parliamentary position X Experience of protests Left-Right Ideology X Experience of protests *** *** -.563*** -.526***.363***.293***.735***.535*** -.447** -.797** Intercept 2.719*** 3.678*** 3.272*** 3.035*** RANDOM PART: Municipality level variance Individual level variance 4.161*** 3.895*** 3.970*** 4.036*** Random slope variance Experience Random slope variance Parliamentary Position.215**.216***.214*** Number of municipalities Number of individuals Bayesian Information Criteria Comment: * p<.10 ** p<.05 *** p<.01 33