ABSTRACT
In the literature, two competing claims can be found on the relationship between political trust and political participation. While some authors argue that trust is a prerequisite for any form of participation to occur, others claim that distrust can be a motivating factor for participation in non-institutionalised forms of participation. The social movement literature suggests that political trust will only have these behavioural consequences if it is associated with sufficiently high levels of political efficacy. In this article, we rely on the results of the 2006 European Social Survey for an in-depth analysis of the relationship between political trust and participation in 25 countries. The multilevel regression shows that while political trust is positively associated with institutionalised participation, it is negatively associated with non-institutionalised participation. Moreover, the effect of political trust on institutionalised participation is dependent on self-confidence about one's capability to understand politics.
1. Introduction
Does it make sense to participate in political life, if one does not trust the political system or political decision-makers? This is one of the perennial questions in political sociology, from the moment on a systematic decline in political trust was first documented in the late 1960s (Levi and Stoker 2000). As Citrin (1974: 980) already argued:
The standard hypothesis, of course, is that political disaffection (cynicism, alienation) is associated with a rejection of conventional or “conformist” models of political participation (…) the politically cynical should be more likely than those who trust the government either to withdraw from political activity altogether or to engage in noncustomer, sometimes illegal, activities such as participating in sit-ins or riots, or organizing for revolution.
In this quote, the words ‘either/or’ – italicised in the original text – are of crucial importance. If low levels of political trust would be associated with low participation levels, one could argue that this phenomenon poses a threat to the stability of the political system, as we can assume that participation is a key characteristic of stable democratic political systems (Barber 1984). If, on the other hand, institutionalised participation would merely be replaced by non-institutionalised forms of participation, this implies that political systems will have to adapt to new forms of interaction with the population (Inglehart and Welzel 2005; Rosanvallon 2008). It is therefore of crucial importance to determine whether low levels of political trust are associated with a complete withdrawal from political life, or whether this ‘only’ leads to a change in the political participation repertoires being used by distrusting citizens. In this article, we want to investigate whether political trust is indeed a necessary resource for political participation.
The relations under investigation are inherently complex. It can be assumed that political trust has a positive impact on some forms of participation, but not on others. Furthermore, we know that political mobilisation processes are the result of a complicated interplay between various elements: issue salience, mobilisation efforts, socio-economic status, civic skills, interest, motivation, political opportunity, political trust and political competence and awareness (Verba et al.1995). From this perspective, political trust can still be considered as an essential prerequisite for political participation to occur, but it should not be considered as a sufficient prerequisite: political trust can only be expected to have a positive effect on participation levels if it is combined with other civic resources. While there is a general consensus that both political trust and efficacy are important resources for participation, there is disagreement on the nature of this relationship: are the effects of political trust and efficacy on political participation simply additive or are both attitudes required? Some authors state that political trust and efficacy are two essential prerequisites and only their joint presence will trigger political action (Gamson 1968; Fraser 1970). Empirical studies testing this theoretical expectation yielded mixed results and the interplay between both attitudes has not yet been tested in a large cross-national study.
In this article, we propose to build on prior research not just by using new comparative data and carefully constructed measurements, but also by paying attention to the occurrence of interaction effects between political trust and elements of political efficacy. In this article we first review the theoretical and empirical literature on the relationship between political trust and political participation, before we move on with our own analysis on the relationship between political trust, institutionalised participation, non-institutionalised participation and voting.
2. Theory
The alleged decline in political trust in some Western societies has led to an intensive debate on the future of political participation (Hetherington 1998). Basically, two different claims can be distinguished. On the one hand, a number of authors assume that citizens need a basic form of trust in the political process before they will embark on various forms of political participation. On the other hand, a number of authors have made the claim that a lack of political trust means that citizens will abandon elite-directed forms of participation, in favour of elite-challenging participation acts (Levi and Stoker 2000).
The first, and oldest line of reasoning, assumes that citizens require political trust in order to participate in politics. Almond and Verba (1963: 27) already stated that in order to become a participant in politics, citizens need positive orientations towards the political system. Negative attitudes towards or negative judgements of the political system lead to alienation which erodes the effectiveness and the legitimacy of the democratic system and will lead to democratic instability over time (Almond and Verba 1963: 22, 230).
The second claim made in the literature, however, states that distrust can also serve as a resource and a motivating factor for political participation. When political decision makers or the political system as a whole are perceived as untrustworthy, citizens will feel compelled to intervene. In this respect Barnes and Kaase (1979: 409–77) introduced a crucial distinction between conventional and unconventional participation. They state that while ‘conventional participation is a manifestation of support for the political order (…). The presence of dissent in democratic societies is often indicated by unconventional modes of action’ (Barnes and Kaase 1979: 444). It has to be noted that by now much of these originally unconventional modes of participation have become largely conventional. In line with the more recent literature we will opt for a distinction between institutionalised and non-institutionalised forms of political participation. While institutionalised participation refers to all acts directly related to the institutional process (campaign activity, contacting elected officials, …), non-institutionalised forms have no direct relation with the electoral process or the functioning of the political institutions (Kingemann and Fuchs 1995). The literature allows us to expect that distrusting citizens are more likely to engage in non-institutionalised forms of political participation which are often rather goal-oriented, issue specific and situated outside the institutions of the political system.
Norris (1999) elaborates on this assertion by claiming that a more critical outlook of citizens towards the political system should not be regarded as a symptom of alienation from the political system. Rather, the rise of a new generation of ‘critical citizens’ will lead to an increase in elite-challenging forms of political participation. Rosanvallon (2008) takes the argument a step further by arguing that distrust is one of the key expectations in most liberal constitutional frameworks. Citizens are not expected to place a blind trust in political authorities, but they are supposed to practice a critical scrutiny towards the actions of politicians and institutions. According to Rosanvallon, the vigilant citizen actively monitors the acts of politicians and decision-makers. A more radical stand is taken by Hibbing and Theiss-Morse (2001: 5) who claim that distrust not only stimulates non-institutionalised ways of political participation but rather all kinds of engagement as ‘People often view their political involvement as medicine they must take in order to keep the disease of greedy politicians and special interests from getting further out of hand’. This claim about the motivating role of distrust is well founded in some of the older literature on social movement participation. Already in 1968, Gamson argued that ‘high trust in authorities implies some lack of necessity for influencing them’ (Gamson 1968: 46–7). Within social movement theory, especially proponents of the grievance theory have claimed already quite early on that without major grievances towards the political system, there is no need to become politically active (Craig and Maggiotto 1981). Based on these insights, one can therefore assume that the relation between political trust and political participation will depend on the exact nature of the participation act. Following the traditional logic of the Almond and Verba approach one can expect that political trust will lead to a more intensive form of participation in conventional, or institutionalised forms of participation:
H1: Political trust will be positively associated with institutionalised political participation.
The more critical approach preferred by authors like Rosanvallon and Norris, on the other hand, allows us to assume that distrust toward the political system can be a strong motivating factor for elite-challenging and non-institutionalised participation:
H2: Political trust will be negatively associated with non-institutionalised political participation
It is clear, however, that the impact of political trust on participation should not be studied in an isolated manner (Kriesi and Westholm 2007). The Civic Voluntarism Model – as developed by Verba, Schlozman and Brady (1995) – emphasises the importance of resources for political participation. A major resource enabling political participation is political efficacy, i.e., ‘the feeling that individual political action does have or can have an impact upon the political process’ (Campbell et al.1954: 187). Political efficacy has been studied extensively since the 1950s and its positive effect on political participation is well-established in empirical research (Campbell et al.1954; Abramson 1983; Bennett 1986; Rosenstone and Hansen 1993; Verba et al.1995; Anderson 2010). A distinction has been made between internal and external political efficacy. Internal efficacy indicates a person's assessment of his or her capabilities to take part in the political process whereas external efficacy indicates a person's assessment of the responsiveness of the political system (Niemi et al.1991; Anderson 2000).
Several authors have argued that especially the combined presence of internal efficacy and political trust functions as an ideal breeding ground for political action. Within the Civic Voluntarism Model, both attitudes are believed to have a positive effect on participation, and their effects are considered as being additive. However, other scholars have argued that both attitudes are needed and that only their joint presence will lead to participation: trust and efficacy thus are both treated as necessary but insufficient conditions for political participation. Already Almond and Verba (1963) considered the development of a sense of political competence and positive orientations towards the political system among the citizenry as two essential prerequisites for the occurrence of a participant political culture. Self-evidently, the reverse argument can be made too: alienation is thus defined as the combination of political powerlessness (i.e., the absence of political efficacy) and the presence of distrust among citizens (Miller 1974).
The theoretical approach introduced by Gamson (1968) is especially relevant in this regard. Gamson (1968) hypothesises that a combination of distrust and internal efficacy triggers political action. One's level of political trust is irrelevant, if one does not feel capable to participate. Conversely, one's level of internal efficacy is irrelevant for political participation if one does not feel the need to intervene on specific policy issues (Gamson 1968; Fraser 1970). Therefore, ‘a combination of high sense of political [internal] efficacy and low political trust is the optimum combination for mobilization-a belief that influence is both possible and necessary’ (Gamson 1968: 48). Some scholars have argued that Gamson's theory should be modified as, they claim, the joint presence of distrust and internal efficacy only fosters participation outside of the conventional arenas, while one should not expect this combination to have an effect on conventional forms of participation (Hawkins et al.1971: 1135; Muller 1977). These observations reinforce the claim that participation is the result of an interaction between political (dis)trust and political efficacy. This would imply that the behavioural consequences of political trust cannot be studied in an isolated manner, as has been done in most of the available research thus far. Following the idea of Gamson therefore allows us to assume that the impact of political trust on participation levels is also dependent on the level of the feeling of political awareness:
H3: The effect of political (dis)trust on political participation is dependent on the interaction with a feeling of political awareness.
3. Prior results on the relation between political trust and participation
The currently available empirical research does not provide us with a clear and unequivocal answer to the questions posed in this theoretical debate. Most studies focus on the United States using the ‘trust in government’ – measurement from the American National Election Studies (NES) and this line of research did not find any consistent relationship between political trust and political participation. A specific problem when using the NES study, however, is that the questions on political trust used in this survey have been heavily criticised on methodological grounds (Citrin 1974; Seligson 1983; Muller and Jukam 1983; Craig et al.1990). It has been argued that the NES-questions measure the confidence in the incumbent administration, rather than to tap diffuse support for the political system as was suggested in the classic study by Easton (1965).
Most comparative surveys rely on a different measurement battery that is expected to tap in a more valid manner the level of trust respondents have in basic political institutions, such as parliament, government, the legal branch of government or various international organisations. When using this operationalisation of diffuse support for the political system, a different picture emerges. Dalton reported a positive correlation between political trust and voter turnout in 12 advanced industrial democracies (Dalton 2004: 171–6). Single-nation studies in the UK and Canada found similar results (Pattie and Johnston 2001: 215; Bélanger and Nadeau 2005: 44). Analyses of the 1995–1997 World Values Survey data also reveal positive correlations between trust in political institutions and party activity (Norris 1999: 257–61; Dalton 2004: 173–6). Most of the available research, therefore, tends to support hypothesis 1.
When turning to non-institutionalised forms of political participation, Kaase (1999) found a negative relation between trust in political institutions and protest participation using the 1981–1996 World Values survey data in nine countries. With regard to demonstrating, Heath (2008) analysed 23 monthly general population surveys (2000–2002) in the UK. The results of this analysis reveal that people who are dissatisfied with the working of democracy are generally more willing to protest than people who are satisfied. While there is some support for hypothesis 2, the evidence thus far remains clearly limited.
Following the publication of Gamson's Power and Discontent (1968) the prediction that both distrust and efficacy are necessary prerequisites for participation to occur was subjected to scrutiny in several studies. Most of the early studies did not find any evidence for the beneficial effects of the presence of political trust and internal efficacy on participation (Fraser 1970; Hawkins et al.1971). However, studies focusing on participation outside the conventional arena (such as riots and demonstrations) confirmed the occurrence of this interaction effect: the joint presence of distrust and internal efficacy was found to trigger ‘unconventional’ participation (Paige 1971; Citrin 1977; Muller 1977; Guterbock and London 1983). More recent research on this topic is scarce as most of the empirical work can be situated in the 1970s and 1980s. Moreover, these studies are based on single-nation studies with a clear focus on the United States. On this account, Bélanger and Nadeau (2005: 137) note:
Our results also challenge the major conclusion reached in American studies that show no relation between political trust and electoral participation. (…) This conclusion further stresses the need to offset the United States bias in the literature on political trust.
There is, therefore, not much recent evidence with regard to the third hypothesis we want to test in the current article.
4. Data and methods
In order to assess the relation between political trust and political participation, we need access to a dataset that includes a comprehensive and reliable measurement of both variables. The third wave of the European Social Survey (2006) provides extensive and up-to-date information on trust in institutions, participation and various political attitudes (Jowell and the Central Coordinating Team 2007). The advantage of using this cross-country dataset is that we can be confident that any results we might find are not dependent on specific characteristics of one country or one political system. As we know there are large country-level differences between participation levels, it is clear that findings from a single country cannot be generalised toward Western societies. As open democratic governance structures tend to be associated with high levels of political participation, it is important to control for basic elements of the political system in order to study political participation levels, as was already noted by Jennings (1990). Comparative data, therefore allow us more certainty that any results we might find are not due to specific country-level characteristics. Data were collected in 2006–2007 by means of face-to-face interviews among representative samples of the population of 25 European countries.1 Response rates varied by country with a median of 63.4%.
4.1. Dependent variable: Political participation
The European Social Survey (2006) provides an extensive list of political activities in which respondents can participate. Institutionalised forms of participation include party membership, working in a political party or action group at least once during the last year and contacting government officials at least once during the last year.2 Furthermore, respondents were also asked whether they had voted during the last national election, and given the very specific characteristics and distribution of this participation act, we decided to analyse it separately. By doing so, we are in line with most of the political participation literature. Voting remains by far the most widespread form of political participation. Contrary to other forms of participation, voting is legally limited in use to one vote for every person and in some countries, voting is compulsory (e.g., Belgium). Non-institutionalised forms of participation, on the other hand, include having signed a petition, boycotted products, and having taken part in a demonstration.3
Our distinction between institutionalised and non-instititutionalised forms of participation closely follows the framework already set out by Barnes and Kaase (1979). Institutionalised forms of participation are defined and organised by members of the political elite (most notably political parties), while non-institutionalised forms of participation in practice are being used predominantly by non-elite actors, in order to challenge the political elite or to gain access to the political agenda (Inglehart 1997). This distinction is in line with the literature in this regard (e.g., Inglehart and Catterberg 2002; Dalton 2008; Marien et al.2010) and it is also empirically valid as a confirmatory factor analysis demonstrates these are distinct dimensions.
To summarise, we have three dependent variables: institutionalised participation, non-institutionalised activities and voting. All three dependent variables are categorical with a value 0 = not done and a value 1 = done.
4.2. Independent variables
The main independent variable is political trust. We operationalise diffuse political trust by measuring trust in the most important institutions of a political system (Muller and Jukam 1983). The question on political trust was phrased: ‘please tell me on a score of 0 to 10 how much you personally trust each of the institutions I read out. 0 means you do not trust an institution at all, and 10 means you have complete trust’. The institutions include: the national parliament, politicians, political parties, the legal system and the police. A principal component analysis indicated that the items load strongly on a single dimension, explaining over half of the total variance with an Eigenvalue of 3.55 and Cronbach's α of 0.89 (Table 1).4
| Trust in … . | Loadings . |
|---|---|
| Country's parliament | 0.867 |
| The legal system | 0.825 |
| The police | 0.757 |
| Politicians | 0.887 |
| Political parties | 0.868 |
| Eigenvalue | 3.55 |
| Percentage explained variance | 70.91 |
| Trust in … . | Loadings . |
|---|---|
| Country's parliament | 0.867 |
| The legal system | 0.825 |
| The police | 0.757 |
| Politicians | 0.887 |
| Political parties | 0.868 |
| Eigenvalue | 3.55 |
| Percentage explained variance | 70.91 |
Note: Results of a principal component analysis on full ESS 2006 sample. (43,733 individuals in 25 countries)
As already argued, political efficacy too can be considered as an important determinant of political participation. Niemi et al. (1991: 1407) note that political efficacy has two components, as it refers to the self-perceived ability to understand politics and to participate in an effective manner. Unfortunately, the ESS questionnaire did not include a full measurement of internal political efficacy, but only includes two questions on the cognitive awareness with regard to politics. Based on the statistical material presented by Niemi et al. (1991) it can be assumed that this cognitive awareness will be closely associated with the efficacy with regard to participation, but given the fact that the ESS dataset does not include any information on this matter, for the time being we refer to these two variables as a form of political awareness.
The following two items were used: ‘How often does politics seem so complicated that you can't really understand what is going on?’ and ‘How difficult or easy do you find it to make your mind up about political issues?’ Both questions tap the feeling that the individual is powerless to understand politics. These variables are correlated (Pearson's correlation: 0.466) and correspond to one underlying factor.
In our analysis, we also include various control variables. Previous research has indicated that education, gender, age, ethnicity and religion are, to varying extents, related to levels or types of political participation (Verba et al. 1995; Van Deth and Elff 2004; Desposato and Norrander 2008; Marien et al. 2010). We included level of education, age, gender, citizenship status and church practice as background variables. Age, level of education and church practice are coded as continuous variables, whereas gender and citizenship status are coded as dichotomy variables (with respectively women = 0 and men = 1; not a citizen of the country = 0 and citizen of the country = 1; see the Appendix for question wording).
Further, we will control for social capital that is often assumed to influence political participation (Hooghe 2003). Based on the literature, we think it is safe to include two elements commonly included in the social capital literature as control variables: generalised trust and networks (Newton 2007). We can assume that trusting people are more likely to engage in collective action. In the ESS, generalised trust was included with three standard questions that strongly correlate and correspond to one factor (with 69.5% explained variance and an Eigenvalue of 2.09). Further, people can be mobilised through networks. Especially, trade unions often mobilise citizens for political action. Therefore, respondents were asked whether they were member of a trade union or a similar organisation. Informal networks were measured by asking respondents: ‘How often do you socially meet with friends, relatives or colleagues?’ Finally, political participation levels tend to be lower in less established democracies (Kostadinova and Power 2007). Some of the more recent democracies in Central and Eastern Europe still struggle to establish a well-functioning democratic political system. In order to take this effect into account, a composite index of good governance was included in the analyses. This index, developed by the World Bank, is based on expert and citizen surveys and is meant to tap the effective functioning of government institutions within a country. The index summarises information on a country with several measures including voice and accountability, upholding the rule of law, fighting corruption, government effectiveness, regulatory quality and offering political stability (Kaufmann et al.2008).
Given that we are interested in three dependent variables (institutionalised participation, non-institutionalised participation and voting), we performed three regression analyses. Because all dependent variables are coded as dichotomous variables, we used logistic regressions techniques. Moreover, it is a reasonable expectation that the scores of respondents sampled in the same country will tend to resemble one another, and this nested structure of the data can bias the standard errors which results in spurious significant results. Multilevel analysis allows taking this intra class-correlation and the variance between countries into account (Hox 2002).
5. Results
First we explain the occurrence of institutionalised participation (Table 2). The null-model demonstrates that 5% of the variation can be situated at the country level. In Model I, all independent variables are included. In line with hypothesis 1, we can observe that political trust is positively related to participation in institutionalised participation. The results show that men, highly educated and older citizens participate more often in this form of political action. Political awareness, too, proves to have a significant and very strong effect on institutionalised participation. Further, union membership and integration in informal networks boost institutionalised participation. As expected, participation is higher in countries that have higher values on the good governance index. In this model, approximately 2% of the variance between the countries is explained.5 Following hypothesis 3, in Model II we allow for an interaction effect between political awareness and political trust. This interaction effect proves to be significant suggesting that a combination of political trust and confidence in the ability to understand politics is highly conducive for institutionalised political participation.
| . | MODEL I . | MODEL II . | ||
|---|---|---|---|---|
| . | Coefficient . | S.E. . | Coefficient . | S.E. . |
| Political trust | 0.043*** | 0.008 | 0.032*** | 0.008 |
| Political awareness | 0.355*** | 0.016 | 0.353*** | 0.016 |
| Education | 0.261*** | 0.015 | 0.262*** | 0.015 |
| Gender (Male = 1) | 0.220*** | 0.027 | 0.217*** | 0.027 |
| Age | 0.010*** | 0.001 | 0.010*** | 0.001 |
| Citizen status | 0.515*** | 0.084 | 0.511*** | 0.084 |
| Church practice | 0.116*** | 0.010 | 0.117*** | 0.010 |
| Generalised trust | −0.024** | 0.008 | −0.023* | 0.008 |
| Union membership | 0.272*** | 0.033 | 0.272*** | 0.033 |
| Informal networks | 0.194*** | 0.017 | 0.195*** | 0.017 |
| Good governance (country level) | 0.298** | 0.093 | 0.301** | 0.093 |
| Political trust * political awareness | 0.028*** | 0.007 | ||
| Intercept | −2.756*** | 0.109 | −2.754*** | 0.109 |
| Intra-Class Correlation (%) | 2.80% | 2.80% | ||
| Between country error variance | 0.094 | 0.028 | 0.094 | 0.028 |
| . | MODEL I . | MODEL II . | ||
|---|---|---|---|---|
| . | Coefficient . | S.E. . | Coefficient . | S.E. . |
| Political trust | 0.043*** | 0.008 | 0.032*** | 0.008 |
| Political awareness | 0.355*** | 0.016 | 0.353*** | 0.016 |
| Education | 0.261*** | 0.015 | 0.262*** | 0.015 |
| Gender (Male = 1) | 0.220*** | 0.027 | 0.217*** | 0.027 |
| Age | 0.010*** | 0.001 | 0.010*** | 0.001 |
| Citizen status | 0.515*** | 0.084 | 0.511*** | 0.084 |
| Church practice | 0.116*** | 0.010 | 0.117*** | 0.010 |
| Generalised trust | −0.024** | 0.008 | −0.023* | 0.008 |
| Union membership | 0.272*** | 0.033 | 0.272*** | 0.033 |
| Informal networks | 0.194*** | 0.017 | 0.195*** | 0.017 |
| Good governance (country level) | 0.298** | 0.093 | 0.301** | 0.093 |
| Political trust * political awareness | 0.028*** | 0.007 | ||
| Intercept | −2.756*** | 0.109 | −2.754*** | 0.109 |
| Intra-Class Correlation (%) | 2.80% | 2.80% | ||
| Between country error variance | 0.094 | 0.028 | 0.094 | 0.028 |
Source: ESS (2006). N = 40,222 individuals in 25 countries. Notes: Results of logistic multilevel analyses. Method of estimation: PQL. The variables were centred. Variance at country level = 5.48% in null-model. Sign.: *p < 0.05; **p < 0.01; ***p < 0.001. Only ‘good governance’ added as country-level variable, all other variables individual level.
Second, we explain the occurrence of non-institutionalised participation and here we proceed in a similar manner (Table 3). The null-model shows that 22% of the variation in the dataset can be situated at the country level. In Model III, all independent variables are included. In line with hypothesis 2, we can observe a negative relation between political trust and non-institutionalised participation. Furthermore, contrary to institutionalised participation, women and younger people participate more often in non-institutionalised than men and older people. Education however remains an important source of inequality in participation. Also for non-institutionalised participation, political awareness plays an important role: the probability that a woman with the lowest level of political awareness will participate is 0.49, while this is 0.77 for a woman with the highest level of political awareness (holding all other variables constant at their mean). Also social capital proves to be important given that generalised trust, union membership and informal networks all positively contribute to non-institutionalised participation.
| . | MODEL III . | MODEL IV . | ||
|---|---|---|---|---|
| . | Coefficients . | S.E. . | Coefficients . | S.E. . |
| Political trust | −0.047*** | 0.007 | −0.045*** | 0.008 |
| Political awareness | 0.305*** | 0.015 | 0.306*** | 0.015 |
| Education | 0.367*** | 0.014 | 0.367*** | 0.014 |
| Gender (Male = 1) | −0.223*** | 0.025 | −0.222*** | 0.025 |
| Age | −0.008*** | 0.001 | −0.008*** | 0.001 |
| Citizen status | 0.542*** | 0.070 | 0.544*** | 0.070 |
| Church practice | 0.019* | 0.010 | 0.019 | 0.010 |
| Generalised trust | 0.029*** | 0.008 | 0.029*** | 0.008 |
| Union membership | 0.370*** | 0.031 | 0.370*** | 0.031 |
| Informal networks | 0.116*** | 0.016 | 0.116*** | 0.016 |
| Good governance (country level) | 1.114*** | 0.181 | 1.115*** | 0.181 |
| Political trust * political awareness | −0.011 | 0.006 | ||
| Intercept | −2.359*** | 0.144 | −2.361*** | 0.254 |
| Intra-Class Correlation (%) | 10.19% | 10.24% | ||
| Between country error variance | 0.370 | 0.106 | 0.372 | 0.107 |
| . | MODEL III . | MODEL IV . | ||
|---|---|---|---|---|
| . | Coefficients . | S.E. . | Coefficients . | S.E. . |
| Political trust | −0.047*** | 0.007 | −0.045*** | 0.008 |
| Political awareness | 0.305*** | 0.015 | 0.306*** | 0.015 |
| Education | 0.367*** | 0.014 | 0.367*** | 0.014 |
| Gender (Male = 1) | −0.223*** | 0.025 | −0.222*** | 0.025 |
| Age | −0.008*** | 0.001 | −0.008*** | 0.001 |
| Citizen status | 0.542*** | 0.070 | 0.544*** | 0.070 |
| Church practice | 0.019* | 0.010 | 0.019 | 0.010 |
| Generalised trust | 0.029*** | 0.008 | 0.029*** | 0.008 |
| Union membership | 0.370*** | 0.031 | 0.370*** | 0.031 |
| Informal networks | 0.116*** | 0.016 | 0.116*** | 0.016 |
| Good governance (country level) | 1.114*** | 0.181 | 1.115*** | 0.181 |
| Political trust * political awareness | −0.011 | 0.006 | ||
| Intercept | −2.359*** | 0.144 | −2.361*** | 0.254 |
| Intra-Class Correlation (%) | 10.19% | 10.24% | ||
| Between country error variance | 0.370 | 0.106 | 0.372 | 0.107 |
Source: ESS (2006). N = 40,086 in 25 countries. Notes: Results of logistic multilevel analyses (PQL estimation). The variables were centred. Variance at country level = 22.08% in null-model. Sign.: *p < 0.05; **p < 0.01; ***p < 0.001. Only ‘good governance’ added as country-level variable, all other variables individual level.
Similar to institutionalised participation, non-institutionalised participation is substantially higher in countries with higher values on the good governance index. The interaction between political trust and political awareness, entered in Model IV does not reach statistical significance. This finding contradicts the earlier studies on this interaction effect (Paige 1971; Citrin 1977; Muller 1977; Guterbock and London 1983). In these 25 societies, a combination of political trust and political awareness does not increase the likelihood to engage in non-institutionalised activity.
In the third analysis, we examine citizens’ turnout during elections. Therefore, we conduct our analysis on the electorate and eliminated respondents who were not eligible to vote due to their age or citizenship status (Table 4). Again, we start with a null-model, demonstrating that 12.4% of the variation in the dataset can be situated at the country level. Model V shows that trusting citizens turn out more often. Furthermore, highly educated respondents and older citizens are more likely to vote. Church practice and social capital increase the likelihood to vote. Also for voting, political awareness proves to be important. We can observe, however, that in this case the good governance indicator does not help us to explain the observed differences between countries. Similar to other forms of institutionalised participation, there is a positive interaction effect between political trust and political awareness (Model VI), suggesting that both attitudes motivate voting and that a combination of political trust and awareness boosts voter participation even more strongly.
| . | MODEL V . | MODEL VI . | ||
|---|---|---|---|---|
| . | Coefficients . | S.E. . | Coefficients . | S.E. . |
| Political trust | 0.156*** | 0.008 | 0.160*** | 0.009 |
| Political awareness | 0.290*** | 0.017 | 0.328*** | 0.018 |
| Education | 0.297*** | 0.017 | 0.296*** | 0.017 |
| Gender (Male = 1) | −0.045 | 0.030 | −0.050 | 0.030 |
| Age | 0.030*** | 0.001 | 0.029*** | 0.001 |
| Church practice | 0.168*** | 0.012 | 0.169*** | 0.012 |
| Generalised trust | 0.048*** | 0.008 | 0.049*** | 0.008 |
| Union membership | 0.461*** | 0.044 | 0.463*** | 0.044 |
| Informal networks | 0.070*** | 0.017 | 0.071 *** | 0.017 |
| Good governance (country level) | 0.043 | 0.186 | 0.045 | 0.185 |
| Political trust * political awareness | 0.046*** | 0.007 | ||
| Intercept | 0.795*** | 0.132 | 0.798*** | 0.131 |
| Intra-Class Correlation (%) | 10.76% | 10.63% | ||
| Between country error variance | 0.393 | 0.113 | 0.388 | 0.112 |
| . | MODEL V . | MODEL VI . | ||
|---|---|---|---|---|
| . | Coefficients . | S.E. . | Coefficients . | S.E. . |
| Political trust | 0.156*** | 0.008 | 0.160*** | 0.009 |
| Political awareness | 0.290*** | 0.017 | 0.328*** | 0.018 |
| Education | 0.297*** | 0.017 | 0.296*** | 0.017 |
| Gender (Male = 1) | −0.045 | 0.030 | −0.050 | 0.030 |
| Age | 0.030*** | 0.001 | 0.029*** | 0.001 |
| Church practice | 0.168*** | 0.012 | 0.169*** | 0.012 |
| Generalised trust | 0.048*** | 0.008 | 0.049*** | 0.008 |
| Union membership | 0.461*** | 0.044 | 0.463*** | 0.044 |
| Informal networks | 0.070*** | 0.017 | 0.071 *** | 0.017 |
| Good governance (country level) | 0.043 | 0.186 | 0.045 | 0.185 |
| Political trust * political awareness | 0.046*** | 0.007 | ||
| Intercept | 0.795*** | 0.132 | 0.798*** | 0.131 |
| Intra-Class Correlation (%) | 10.76% | 10.63% | ||
| Between country error variance | 0.393 | 0.113 | 0.388 | 0.112 |
Source: ESS (2006). N = 36,518 individuals in 25 countries. Notes: Results of logistic multilevel analyses (PQL estimation). The variables were centred. Variance at country level = 12.36% in null-model. Sign.: *p < 0.05; **p < 0.01; ***p < 0.001. Only ‘good governance’ added as country-level variable, all other variables individual level.
The results reported in the previous tables were obtained using the full ESS sample. We can thus be confident that the findings are not dependent on the characteristics of specific country settings. In order to test for the robustness of the results, we also conducted the analysis country by country. In a large majority of cases, the same patterns were found, although conventional standards of significance were not always reached, due to smaller sample sizes. The relations did prove to be most meaningful in the more established democracies, while they were more volatile in some of the more recent democracies in Central and Eastern Europe. This might suggest that the pattern we describe in the analysis might be mostly applicable to established democracies, but this is something that should be investigated more in depth in future research.
6. Discussion
In the literature, competing claims have been put forward with regard to the relation between political trust and various forms of political participation. While some authors claim that political trust is a prerequisite for any form of political participation to occur, others assume that a lack of trust can lead to a more intensive form of citizens’ participation. In this article, we report on a comprehensive analysis of this relation, using the results of the 2006 European Social Survey. The data that are available in this survey allow us to formulate an answer to the question Citrin already posed 37 years ago: will low levels of political trust be associated with less participation, or rather with different patterns of participation? The evidence presented in this article suggests that the latter option is the case: while citizens with high levels of political trust are more likely to engage in institutionalised forms of political participation (thus confirming hypothesis 1), political trust is negatively associated with participation in non-institutionalised forms (confirming hypothesis 2). For voting, a positive relation with political trust could be observed. Our first two hypotheses are thus clearly confirmed: political trust seems to boost institutionalised participation and voting in particular but reduces non-institutionalised participation. It has to be acknowledged, however, that the results of the analysis were strongest for the act of voting and this lends some support to the hypothesis that the decline of political trust could be regarded as an important causal factor for the observed decline in voter turnout in a number of liberal democracies. Whether the loss in voter turnout can be compensated by a simultaneous rise in non-institutionalised participation as a method for citizens to gain political influence is a question that falls beyond the scope of the current article.
Self-confidence about one's capability to understand politics also proved to have a strong effect on all forms of political participation. In 1968, William Gamson stated that a combination of distrust and high efficacy would serve as an ideal breeding ground for political action. Partly, the current analysis confirms the notion that both variables are important and interact with one another. The interaction effects, however, are different than Gamson would have assumed. For institutionalised participation and voting the analysis revealed that a combination of trust and awareness leads to high participation figures. For this kind of participation repertoire, we can indeed conclude that an interaction occurs between both resources for participation, thus confirming hypothesis 3. This implies that participation levels will be highest among those that trust the system and also have confidence in their own way capability to understand the political system. But while distrust and awareness, on their own, are related to non-institutionalised participation, for this kind of participation we did not observe any significant interaction effect – contrary to Gamson's predictions and hypothesis 3. Distrusting citizens are more likely to engage in non-institutionalised politics but their participation levels do not receive a stronger boost from their capability to understand politics than is the case among trusting citizens. However, the findings of this study are limited to one element of internal efficacy, the confidence in one's ability to understand politics. Data limitations did not allow us to investigate the effect of a second element of internal efficacy, the belief in one's own capability to participate.
The result of the analysis is that institutionalised participation and voting are rather straightforward to predict. Those who trust the system and feel capable to understand it, opt for this kind of participation repertoire, and they do so even more intensively if both feelings are present simultaneously. For non-institutionalised participation, on the other hand, the image remains less clear. Those who distrust the system indeed more often join these kinds of action, but this effect is not moderated by the feeling of political awareness. This might suggest that political distrust might be a strong mobilising force across society, and that it is not dependent on the capability of citizens to actually understand politics. While Gamson in the 1960s quite strongly argued that protest should be seen as an instrumental way of changing society and the political system, the current analysis suggest that non-institutionalised participation is less strongly dependent on one's ability to understand political life. As such, it seems less strongly correlated to the explicit goal of bringing about changes in the political system than more institutionalised forms of political participation. To ascertain whether this is indeed the cause for the changes we detected in this analysis, however, needs to be further investigated. The fact that these are only European data limits the variance with regard to country-level variables, so it would be extremely interesting to investigate this relation using a most different case design, also including countries with closed political systems.
The analysis strongly supports the idea that low levels of political trust should not be equated with low across the board levels of citizen participation and maybe even on the contrary: distrusting citizens are more likely to participate in a whole range of political activities. The current analysis therefore seems to strengthen the claim put forward by Norris, Rosanvallon and other authors: low levels of political trust do not imply an alienation from the political system as such, but rather indicate a structural trend towards different forms of interaction between citizens and the political system. This does not imply, self-evidently, that this trend towards new forms of citizen-state interaction could not entail new problems. To start with, we do not know how effective these forms of non-institutionalised politics actually are. While it might be worthwhile that citizens participate in demonstrations and boycott products for political reasons, we do not know yet how effective these forms of participation are. There is some anecdotal evidence suggesting that demonstrations might have changed the nature of political decision-making (Deneckere 1993; Tilly 2008), but currently there is no research available demonstrating that on a routine basis too this might be considered as an effective form of input into the political decision-making process. The transition towards new interaction patterns between citizens and the political system inevitably creates a number of uncertainties: we know what we lose (i.e., the quasi general character of voting) but we do not know yet whether the emerging alternatives are really effective in ensuring that the view of citizens get represented in the political decision-making process.
Footnotes
Countries included are: Austria, Belgium, Bulgaria, Cyprus, Denmark, Estonia, Finland, France, Germany, Hungary, Ireland, Latvia, the Netherlands, Norway, Poland, Portugal, Romania, the Russian Federation, Slovakia, Slovenia, Spain, Sweden, Switzerland, Ukraine and the United Kingdom.
The item on working in a political party or action group might be rather ambiguous in its formulation. This item, however, correlates strongly with party membership, and therefore we assume that this item refers mainly to active involvement in political parties and other forms of institutionalised engagement.
Boycotting products and services for political and ideological reasons is considered as a hallmark of political consumerism, that is now generally regard as a non-institutionalised form of political participation. Stolle et al. (2005) offer empirical evidence for the claim that political consumerism can be considered as an integral part of contemporary political participation repertoires.
The one-factor solution is preferred given the fact that only one factor emerged with an Eigenvalue over 1.0.
This percentage is the estimated Intra-Class Correlation (ICC) which is computed by dividing the ‘between country error variance’ by the ‘total error variance’ and multiplying this by 100 (in the case of logistic regression the total error variance can be computed by adding the country error variance with π2/3 (See Snijders and Bosker 1999: 224).
Appendix
| . | N . | Min. . | Max. . | Mean . | SD . |
|---|---|---|---|---|---|
| Institutionalised participation (dummy) | 46,660 | 0 | 1 | 0.18 | 0.38 |
| Non-institutionalised participation (dummy) | 46,522 | 0 | 1 | 0.28 | 0.45 |
| Vote | 42,391 | 0 | 3 | 0.78 | 0.41 |
| Institutionalised participation | 46,660 | 0 | 3 | 0.23 | 0.56 |
| Non-institutionalised participation | 46,522 | 0 | 3 | 0.39 | 0.70 |
| Political trust | 43,733 | 0 | 10 | 4.39 | 2.13 |
| Education level | 46,878 | 0 | 3 | 1.79 | 1.06 |
| Gender (1 = Male) | 46.999 | 0 | 1 | 0.45 | 0.50 |
| Religiousness | 46,656 | 0 | 4 | 1.57 | 1.42 |
| Age (in years) | 46,752 | 14 | 101 | 47.40 | 18.61 |
| Citizenship status | 47,068 | 0 | 1 | 0.96 | 0.19 |
| Political Awareness | 44,849 | 1 | 5 | 2.92 | 0.93 |
| Generalised trust | 46,094 | 0 | 10 | 5.08 | 2.03 |
| Member of trade union or similar organisation | 46,663 | 0 | 1 | 0.19 | 0.39 |
| Informal networks | 46,783 | 0 | 3 | 1.71 | 0.89 |
| Interaction political trust * political awareness | 42,263 | 0 | 50 | 13.31 | 8.12 |
| . | N . | Min. . | Max. . | Mean . | SD . |
|---|---|---|---|---|---|
| Institutionalised participation (dummy) | 46,660 | 0 | 1 | 0.18 | 0.38 |
| Non-institutionalised participation (dummy) | 46,522 | 0 | 1 | 0.28 | 0.45 |
| Vote | 42,391 | 0 | 3 | 0.78 | 0.41 |
| Institutionalised participation | 46,660 | 0 | 3 | 0.23 | 0.56 |
| Non-institutionalised participation | 46,522 | 0 | 3 | 0.39 | 0.70 |
| Political trust | 43,733 | 0 | 10 | 4.39 | 2.13 |
| Education level | 46,878 | 0 | 3 | 1.79 | 1.06 |
| Gender (1 = Male) | 46.999 | 0 | 1 | 0.45 | 0.50 |
| Religiousness | 46,656 | 0 | 4 | 1.57 | 1.42 |
| Age (in years) | 46,752 | 14 | 101 | 47.40 | 18.61 |
| Citizenship status | 47,068 | 0 | 1 | 0.96 | 0.19 |
| Political Awareness | 44,849 | 1 | 5 | 2.92 | 0.93 |
| Generalised trust | 46,094 | 0 | 10 | 5.08 | 2.03 |
| Member of trade union or similar organisation | 46,663 | 0 | 1 | 0.19 | 0.39 |
| Informal networks | 46,783 | 0 | 3 | 1.71 | 0.89 |
| Interaction political trust * political awareness | 42,263 | 0 | 50 | 13.31 | 8.12 |
Questions wording
Political Participation
‘During the last 12 months, have you (…)
… contacted a politician, government or local government official?
… worked in a political party or action group?
… signed a petition?
… taken part in a lawful public demonstration?
… boycotted certain products?
‘Are you a member of any political party?’ (Yes/No)
‘Did you vote in the last national election?’ (Yes/No/Not eligible to vote)
Political Trust
‘Please tell me on a score of 0–10 how much you personally trust each of the institutions.’
… [country]'s parliament?
… the legal system?
… the police?
… politicians?
… political parties?
Background variables
‘What is the highest level of education you have achieved?’
0: Primary
1: Lower secondary
2: Upper secondary
3: Post secondary
‘Apart from special occasions such as weddings and funerals, about how often do you attend religious services?’
0 Never
1 Less often
2 Only on special holy days
3 At least once a month
4 Once a week or more
‘Are you a citizen of [country]?’ (Yes/No)
Political awareness
‘How often does politics seem so complicated that you can't really understand what is going on?’
‘How difficult or easy do you find it to make your mind up about political issues?’
Generalised Trust (3 items)
‘Generally speaking, would you say that most people can be trusted, or that you can't be too careful in dealing with people?’
‘Do you think that most people would try to take advantage of you if they got the chance, or would they try to be fair?’
‘Would you say that most of the time people try to be helpful or that they are mostly looking out for themselves?’
Networks
‘Are you or have you ever been a member of a trade union or similar organisation?
‘How often do you socially meet with friends, relatives or colleagues?
References
Marc Hooghe is a Professor of Political Science at the University of Leuven (Belgium) and a Visiting professor at the University of Mannheim (Germany). He holds an ERC Advanced Grant to investigate the changing linkage between citizens and the state.
Sofie Marien is a postdoctoral researcher at the University of Leuven, where she holds an FWO-fellowship. She has published mainly on political trust.