The Political-Economic Correlates of Discursive Engagement in Europe

• DOI: 10.1177/1065912921998457
• Published date: 01 June 2022
• Author: Shane P. Singh,Quinton Mayne
• Date: 01 June 2022
• Subject Matter: Articles

2022, Vol. 75(2) 291 –306

https://doi.org/10.1177/1065912921998457

Political Research Quarterly

© 2021 University of Utah

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/1065912921998457

journals.sagepub.com/home/prq

The health of democracy suffers when citizens disengage

from politics. Voter turnout is on the decline in many

democracies (Blais 2010; Kostelka 2017). Large numbers

of citizens care little about political matters; in Europe

and the United States, for example, well under half of

recent survey respondents report being moderately or

very interested in politics (Russo and Stattin 2017).

Voters with low levels of political interest, who tend to be

less educated (Dostie-Goulet 2009, Jacobs, Phalet, and

Swyngedouw 2004), are less likely to cast meaningful

ballots (Lau and Redlawsk 2006). They are instead prone

to make choices based on misperceptions of economic

and societal conditions (Anderson 2007; Healy and

Malhotra 2013), candidate appearance (Lenz and Lawson

2011), conspiratorial beliefs (Lamberty, Hellmann, and

Oeberst 2018), or the order in which candidates appear on

the ballot (Brockington 2003; Shue and Luttmer 2009).

In this paper, we consider a fundamentally important

and little studied manifestation of political engagement:

citizens choosing to take part in interpersonal political

discussion. Even individuals with relatively low thresh-

olds for general political engagement may take part in

political discussions. This is because, compared to other

kinds of political participation, discussing politics—at

least in the context of free democracies—is a low-cost

activity. Political discussion should, in principle, be an

activity in which anybody can easily engage. Discursive

engagement is also normatively important (Kim and Kim

2008), as even elitist theories of democracy (see, for

example, Schattschneider 1960; Schumpeter 1950) imply

a minimal level of political discussion, albeit intermittent

and short-lived at the time of elections. A democracy in

which citizens rarely or never discuss politics is more

normatively troubling still for defenders of models of

democracy that are more demanding of citizens, includ-

ing accounts rooted in civic republican, deliberative, and

participatory thought (see, for example, Barber 1984;

Pateman 1970).

Political discussion does not always produce demo-

cratic benefits. During political conversations, people

often internalize information that aligns with their beliefs

and disregard messages with which they disagree (e.g.,

998457PRQXXX10.1177/1065912921998457Political Research QuarterlyMayne and Singh

research-article2021

1Harvard University, Cambridge, MA, USA

2University of Georgia, Athens, USA

Corresponding Author:

Quinton Mayne, John F. Kennedy School of Government, Harvard

University, 124 Mount Auburn St. 200N-252, Cambridge, MA 02138,

USA.

Email: quinton_mayne@hks.harvard.edu

The Political-Economic Correlates of

Discursive Engagement in Europe

Quinton Mayne1 and Shane P. Singh2

Abstract

We examine the individual-level characteristics and political and economic conditions associated with political

discussion. To build our model of discursive engagement, we draw on existing research on political participation as

well as our own theoretical reasoning. Our data cover two million individuals in twenty-eight European countries over

forty-five years, and we employ a little-used approach to multilevel analysis that distinguishes variations in engagement

attributable to cross-country differences from those stemming from within-country changes. Our primary findings

reveal that, within countries, citizens are more likely to talk about politics at election time, when there are more

electorally competitive political parties, during periods of recession, when unionization levels are higher, and when

racial and ethnic diversity is greater. Across countries, political discussion is more likely where elections are ongoing

and in countries with lower levels of income inequality and corruption. We also find that men and the higher-educated

are more likely to discuss politics, as are those who are middle aged or employed. Our approach is wide-ranging, but

it is also deliberately correlational. Future observational and experimental studies might expand on and identify the

causal underpinnings of the associations we establish here.

Keywords

political discussion, political engagement, democratic citizenship

Article

292 Political Research Quarterly 75(2)

2 Political Research Quarterly 00(0)

Huckfeldt, Johnson, and Sprague 2004; Huckfeldt and

Sprague 1987, 1995). Still, recent empirical studies pro-

vide evidence of the ways in which discursive engage-

ment may bolster democracy. Partaking in political

discussion can lessen partisan biases (Beck 2002; Lupton,

Singh, and Thornton 2015), increase other forms of polit-

ical participation (Klofstad 2007; McClurg 2003; but see

Mutz 2002b), and foster political tolerance (Mutz 2002a;

Pattie and Johnston 2008). Talking about politics also

provides opportunities to acquire political informati on

(Delli Carpini and Keeter 1996; Finkel and Smith 2011;

Smith 2018), and those who are discursively engaged

are more likely to select parties that align with their

interests (Ryan 2011; Sokhey and McClurg 2012).

Political discussion may also moderate the relationship

between political system characteristics and attitudes

toward democracy (e.g., Hoerner and Hobolt 2020).

We assemble the largest data set ever used to analyze

political discussion, pooling data for roughly two million

respondents from over 2,000 Eurobarometer surveys,

spanning forty-five years and twenty-eight European

democracies. We measure engagement in terms of citi-

zens’ self-reported propensity to talk about politics. To

analyze our data set, we take advantage of an underuti-

lized approach to multilevel modeling that distinguishes

variations in discursive engagement attributable to cross-

country differences from those due to within-country

changes in our explanatory variables. This allows us to

ascertain whether the likelihood of citizens discussing

politics is more or less related to changing circumstances

within their country or enduring characteristics that dis-

tinguish countries from each other.

To determine the model’s inputs, we review the large

literature on political participation and the much smaller

body of existing work on discursive engagement. From

this review, along with our own theoretical reasoning, we

arrive at a group of macro-level political and socio-eco-

nomic predictors of political discussion. On the political

side, we focus on elections, party systems, and political

corruption. On the socio-economic side, we examine the

relationship between discursive engagement and eco-

nomic recessions, unionization, inequality, and ethnic

heterogeneity. While we are constrained by the temporal

and spatial consistency of the questions asked on the

Eurobarometer, we also incorporate a set of well-known

individual-level correlates of participation.

To preview our results, we find that, within countries,

citizens are more likely to talk about politics at election

time, when there are more electorally competitive politi-

cal parties, during periods of recession, when unioniza-

tion levels are higher, and when racial and ethnic diversity

is greater. Across countries, political discussion is more

likely where elections are ongoing and in countries with

lower levels of income inequality and corruption. We also

find that men and the higher-educated are more likely to

discuss politics, as are those who are middle aged or

employed.

While our analysis is broad, our aims are decidedly

non-etiological. We have great regard for the recent

“credibility revolution” in the social sciences (cf. Angrist

and Pischke 2010; Samii 2016), but we see continuing

value in large-scale investigations, such as ours, that

probe the key correlates of important political phenom-

ena. Subsequent studies can and should use design-based

methods to identify the extent to which such correlations

are also causal. Finally, despite its breadth, our study also

leaves room for future associational research to identify

additional correlates of discursive engagement.

A Political-Economic Model of

Discursive Engagement

In what follows, we develop expectations about the

micro- and macro-level correlates of discursive engage-

ment, a term we use to describe taking part in political

discussion or debate with others. We do so with reference

to extant literature on citizen engagement. While this

work places comparatively little focus on the conditions

associated with private political discussion, it provides

important insights into the scale, causes, and correlates of

other important forms of political engagement. The most

sizeable body of work in this area concerns voter turnout,

and we reference this literature frequently. In some cases,

we expect that predictors of voter turnout will have an

opposing relationship with political discussion. Our aim

is to develop a comprehensive but not oversaturated

model, and we thus focus on individual-level, political,

and economic factors with clear empirical and theoretical

links to discursive engagement. We outline the expected

relationship between each explanatory factor and discur-

sive engagement in Table 1.

An individual’s likelihood of voting is often expressed

in terms of the balance between expected benefits and

costs. Turning out is more likely when a party or candi-

date contesting an election is more preferable than others

(b) and when the probability of the individual’s vote

being decisive is higher (p), when the costs of voting (c)

are lower, and when the intrinsic benefits one would

receive from turning out (d) are higher (Downs 1957;

Riker and Ordeshook 1968). In a meta-analysis of 90

studies, Smets and van Ham (2013) find education, mid-

dle age (as opposed to being very young or very old),

income, concern about the outcome of the election, feel-

ings of civic duty, partisan identification, political knowl-

edge, and political interest to be reliably and positively

correlated with voting. Each of these factors has a direct

relationship with the cost-benefit calculation. Those who

are more interested in politics or the election outcome,