Voting and elections
TRUEDEM Work Package 2 explores changes in the electoral behavior of citizens witnessed in the last decades with the aim of identifying patterns of electoral behavior that imply the relationship with trust in political institutions as such. The relationship between voter turnout and trust has been investigated by a wide range of studies. To name only a few, Bjørnskov (2009) found a positive correlation between social trust and voter turnout. Trusting voters were more likely to hold politicians accountable through voting, indicating that trust can influence voter behaviour. Crepaz et al (2016) found that electoral turnout was significantly and positively related to the level of trust in political institutions. This suggests that trust in political institutions can motivate individuals to participate in the electoral process. A worldwide comparative analysis developed by Norris (2022) shows that the popular belief in the steady decline of trust is not supported by empirical evidence. Trust trends are instead marked by their complexity and contextual nature, which makes it necessary to take multiple factors into account when evaluating trust levels across different societies and their relationship with democracy, political participation, and voter turnout. In our model, trust is supposed to be an intermediary factor associated both with societal development and government performance. Although we are aware that interpersonal trust is not necessarily a direct antecedent of political trust since it can stimulate unconventional or non-institutionalized political participation and used against the state (Rose, Mishler, Haerpfer 1997; Kaase 1999; Cox 2003), we follow the Putnam's argument which connects social trust to political trust (Putnam, 1993).
We start the WP2's workplan by examining turnout and its predictors. A non-exhaustive list of determinants of voter turnout identified by the existing literature includes:
Institutional factors: institutional variables such as the legal characteristics of elections (compulsory voting in particular), and the type of electoral system have featured most strongly in turnout models (see for instance Norris, 2004; Arzheimer & Carter, 2006; Stockemer, 2017). Compulsory voting with sanctions shows an uncontroversial boosting effect on turnout (Stockemer, 2017) whereas the theoretical argument that PR should foster turnout is not always empirically supported (Collier and Vicente 2012; Stokes et al. 2013). Although important, other institutional factors such as the voting arrangements and requirements, the registration laws, the voting age, the concurrent elections occur in a minority of studies with mixed results. Some other studies introduce also corruption and trust in the turnout function (see for instance Kostadinova, 2009; Birch, 2010; Norris, 2017; Gutiérrez-Romero & LeBas, 2020; Stokemer et al 2013).
Political factors: variables pertaining to political competition such as the number of parties that win seats also features frequently in turnout models although with an unclear relationship (see for instance Huddy et al., 2018; Leininger & Meijers, 2020).
Socio-economic factors: following modernization theories (Inglehart, 1997), the existing literature suggests that economic development may have major effects on the political involvement of citizens although this influence on voter turnout appears to be moderate in previous studies (see for instance Norris, 2004; Susánszky et al., 2022; Steiner 2010; Diwar 2008; Indridason 2008).
Socio-demographic factors: voter turnout is supposed to be higher in relatively small political environments which make community relation closer and more direct and empirical research seems to confirm this stipulation (see for instance Tavits (2005); Elff, 2007; Bernauer & Bochsler, 2011; Koch et al., 2023; Kostandinova and Power 2007). Other socio-economic variables such as population density or urbanization do not show clear relationships.
Despite the increasing methodological sophistication of turnout studies, there is still no established core model of electoral turnout and as the meta-analysis of aggregate-level research demonstrate different factors have alternatively been considered predictors for macro-level turnout (Geys, 2006; Cancela, Geys 2016; Smets, van Ham, 2013; Stockemer, 2017). These factors have been rarely examined all together, but studies have alternatively tested the influence of just one or few factors. Thus, we aim to provide a comprehensive model to assess turnout in Europe. This analysis prompted us to develop a heuristic model that seeks to systematically define the relationships between electoral turnout (dependent variable), on the one hand, and a complex set of dimensions among which different aspects of trust play an important role. We consider both explanatory factors find to be robust in the established literature and other potentially important factors which have attracted increasing attention in more recent works.
The heuristic model, which become the framework for the D2.1 database on voting and elections, includes 10 dimensions. Three of them can be located at a macro level since they refer to socio-economic factors:
inequalities defined both in terms of income disparities and opportunity access;
development defined both in terms of income, education and welfare;
demography confined not only to the size of population but extended also to the presence of some groups such as the young, the elderly and the migrants.
Five of them can be located at a meso-level being related to political-institutional environment:
electoral system: this includes legal aspects such as compulsory voting and type of electoral system but also voting arrangements, regulations and criteria for registration;
electoral integrity: additional set of indicators characterizing electoral system as a whole in the opinion of experts included through selected indicators of electoral integrity perceptions.
democratic profiles: in terms of respect of the liberal, electoral, participatory, egalitarian, and deliberative principles of democracy
corruption: it refers to manipulations of election in terms of intimidation, vote-buying, violence and other irregularities;
media pluralism: it refers to the media system, its capability of expressing different opinions, level of digitalization, and media censorship and bias;
political fragmentation: it refers to the level of competition in terms of effective number of parties.
Two of them can be located at the micro level being related to civic engagement and including both trust and individual political culture. Indicators on political and institutional trust will be collected within work package 1 of TRUEDEM and merged with the D2.1 database. The distinction between macro, meso- and micro-level factors is also linked to the data collection. Macro data preserve country as unit of analysis, meso- data are election-related information and micro data are individual survey-based data.
Figure 1. Heuristicl model of predictors of turnout [Source: TRUEDEM UNISA-IT].
The described heuristic model become the framework for the D2.1 database on voting and elections. The database, therefore, allows to explore the context-dependency of turnout. Political and institutional determinants of turnout might be more complex than the institutional theory suggests so the dataset may allow to discover if institutions are still the most important predictors of turnout and under which socioeconomic contexts do institutional factors affect turnout. Moreover, it allows to explore the election-dependency of turnout: turnout determinants may depend on the type of election. By including both national and EU elections, the database may allow to explore the variation in voter turnout and factors affecting it in national Parliament and European Parliament elections and to explain why some countries experience a greater difference in voting at the two levels than other countries.
This database may help to explore the measurement-dependency of turnout. Each dimension is operationally defined by a set of indicators including both variables found to be significant predictors of voter turnout (which appear to be indispensable in any analysis of turnout) and new variables we think can be predictors but have been rarely considered into previous studies or did not provide any clear relationship. This because also the measurements of concepts may matter. Just to make some examples, there is no consensus in the literature on how to operationalize development. GDP per capita is the most used indicator but we think that also education and literacy rate and the type of welfare have a role in the level of development of a country. A high per capita income does not necessarily imply higher educational levels, and this makes the relationships complex. The same for inequalities that is commonly measured with the Gini index but if we are interested in the connection between the distribution of income and the distribution of power maybe the poverty rate is a more appropriate measure to detect those who do not have the means to become politically engaged and so become disenfranchised. By looking at several operationalization of the same concept, the dataset may also help to discover the influence on turnout that might be dependent on the operationalization of some concepts.
