The populists

Who are the populists? We analysed variables that might explain or account for populist attitudes across the seven nations. Prior studies have found populist attitudes are shaped by demographic factors and political orientations. Major demographic correlates of populism have been older age and less education. Relevant political orientations have included a stronger political ideology and higher anti-immigrant (nationalist) sentiments (Hawkins et al. 2012; Bernhard and Hanggli 2018; Rico and Anduiza 2019; Tsatsanis et al. 2018; Spruyt et al. 2016).

Following this research, we looked at the multivariate relationships between populism, as indicated by our summary scale, and sets of demographic and political orientation variables. Populist attitudes were entered as the dependent variable in multiple regression analyses, with demographic and political antecedents entered as explanatory variables (see Table 40.2). The results of our multiple regression analyses show that populist attitudes tend to be most closely associated with older age and less education, in line with previous research, as well as higher levels of political participation and a stronger political ideology. Accounting for these associations, less education may be associated with populism due to a perceived elite/non-elite divide between those with greater and lesser education and stronger political attitudes due to populism’s links to more radical politics (Hawkins et al. 2012; Bernhard and Hanggli 2018; Tsatsanis et al. 2018).

That populism is associated with greater levels of political participation may be reflective of the political discontent and desire for political action or change that populist rhetoric can stir. Scholars have also noted a relationships between political knowledge, political interest, and populism (Rico and Anduiza 2019; Bernhard and Hanggli 2018), associations which may emerge from a sense that citizens are in a better position now, due to higher levels of education, to pay attention to politics, judge politicians, and think for themselves (Mudde 2004).

Table 39.2 Multiple regressions predicting populist attitudes1

DE

ES

FR

IT

PL

UK

US

ß

ß

ß

ß

ß

ß

ß

Demographic factors Age

.107*“

.056*

.094"

.028

.094"

.050

.069**

Female

.007

.024

.021

.005

-.004

-.002

-.016

Education

-.121“*

-.035

-.061*

-.145***

-.063*

-.153***

-.086**

Income

-.038

-.013

-.052

-.056*

.020

-.002

-.028

White (UK/US only)

.009

.014

Born in country

Political orientation

-.007

.066**

-.023

-.037

.038

-.038

.038

Political interest

.070*

-.001

.015

.008

-.022

.005

.089**

Political participation

-.029

.104"*

.113*"

.068*

.045

.113***

.178"*

Strength of ideology

.105*“

.153"*

.171***

.058*

-.016

.080"

.057*

R2

.046”*

.048"*

.068***

.034***

.017*

.043***

.081*"

N =

2000

2007

2000

2000

2005

2000

2018

Notes: Standardised coefficients displayed; ' p < .05; ** p < .01; "* p < .001

1 Populism scale ranges from 0 (non-populist) to 4 (strong populist)

Models were also tested with marital status (single, married, lives with partner, divorced/separated, widowed) and life stage (student, employed, retired, unemployed), but these variables were non-significant and therefore excluded.

In short, while populists conform with some stereotypes, such as being older and less educated, they are more engaged and interested in politics than their non-populist counterparts. But are populists trapped in echo chambers or filter bubbles?

Populism, political engagement, and access to political information

We have two approaches to questions about the relationship between populism and online political echo chambers and filter bubbles (Gerbaudo 2018). First, our respondents were asked if ‘most people you communicate with online tend to have political beliefs similar to yours, different political beliefs from you, or a mix of various political beliefs’ and also how often they agreed with the political opinions or political content posted by friends on social media (l=almost never to 5=nearly aways). The relationships between these responses and populist attitudes are shown in Tables 40.3 and 40.4.

Generally, the results indicate that populist attitudes are not associated with higher levels of agreement with the political opinions or content posted by friends on social media or with a higher likelihood of communicating primarily with politically similar others online. Populist attitudes are positively associated with agreeing with content posted by friends on social media, but this relationship is not statistically significant except in Germany = .072, p < .01) and the United States (ß = .077, p < .01). Meanwhile, among populists, there is only a higher likelihood of communicating with similar others in Germany (B = .268, p < .01). These findings go against arguments that online echo chambers foster populist sentiments but are in line with Groshek and Koc-Michalska (2017), who found that homogeneous online networks were not related to support for populist political candidates.

Agreement with the opinions or content posted by friends on social media is more closely associated with younger internet users and those most interested and involved in politics, as

Table 39.3 Multiple regressions predicting level of agreement with the political opinions or content posted by friends on social media

DE

ES

PR

/7'

PL

UK

US

Demographic factors

P

P

P

P

P

P

P

Age

-.286**'

-.092"

-.219***

-.073*

-.046

-.215*"

-.209"*

Female

-.027

-.006

.005

.061*

.055

.007

-.051*

Education

-.057*

-.021

.031

.005

-.023

.028

.021

Income

-.021

.034

-.006

.044

.022

-.038

-.046

Online ability

Political orientation

.072*

.138***

.080“

.095“

.040

.093"

.089"

Political interest

.083**

.113***

.110*“

.020

.132*“

.226*"

.158"*

Political participation

.266***

.188***

.305***

.266***

.281***

.210"*

.203*"

Strength of ideology’

.022

.012

.015

.033

-.074*

.007

.022

Populist attitudes

.072“

.047

.013

.033

.037

-.026

.077"

1<-

.202***

.129***

.213***

.118“*

.139***

.227*"

.198"*

N =

2000

2007

2000

2000

2005

2000

2018

Notes: Standardised coefficients displayed; * p < .05; ** p < .01; “* p < .001

Table 39.4 Logistic regressions predicting communication primarily with politically similar others online1

DE

L’S

PR

IT

PL

UK

US

Demographic factors

P

P

P

P

P

P

P

Age

-.008

-.003

-.003

.000

-.009

-023***

-.006

Female

.220

.145

-.050

.312

.513"

-.163

-.109

Education

-.030

-.069

.168

.106

-.090

-.053

.174

Income

-.057

.086

-.077

-.010

-.049

-.002

-.056

Online ability

Political orientation

.038

.294*

.288

.206

.170

.313*

.174

Political interest

.104

.254*

.031

.280

.238

.606***

.385*"

Political participation

.074*

.067*

.135*"

.123*"

.129***

081"

.056*

Strength of ideology’

.197*

.208*

.342***

.360**

-.047

.437***

.455***

Populist attitudes

.268"

.049

.103

.158

.195

.063

.166

Nagelkerke R2

.049

.070

.090

.096

.076

.197

.150

N =

2000

2007

2000

2000

2005

2000

2018

Notes: Standardised coefficients displayed; * p < .05; ** p < .01; “* p < .001

1 Outcome variable is binary 1 (communicates primarily with politically similar others) and 0 (communicates primarily with politically mixed or different others)

measured by participation (Table 40.3). Communicating primarily with similar others online is most closely associated with political participation and strength of ideology (Table 40.4). Thus, polarisation is more likely to be related to echo chambers than populism.

Secondly, we explored the diversity of populists’ media use, assessing whether there is a risk of individuals with populist attitudes being trapped in online echo chambers or filter bubbles. If true, we might expect individuals with populist attitudes to visit fewer and less diverse sources of information. In this case as well, our findings counter the thesis of populists cocooned in echo chambers or filter bubbles.

We developed a number of indicators of the number and diversity of sources consulted online and offline. In all seven nations, respondents were asked how often they used seven different online sources of information about politics and public affairs, including social media sites, search engines, online-only news sites, legacy online news sites, email, political websites (e.g. for politicians or online political groups), and online video platforms (l=never to 5=very often). The more sources consulted frequently online, for example, the less likely it would be for them to be trapped in a filter bubble or echo chamber.

We then focused on the relationship between populist attitudes and this set of media use variables. We used multiple regressions, looking at the association between populism and source diversity, controlling for demographic and political variables that might moderate this relationship. We also included a measure of online ability as a control (‘How would you rate your ability to do things online?’ l=bad to 5=excellent) as this may have been an obstacle to consulting more sources.

The results reported in Table 40.5 show that when controlling for demographic and political moderating variables, those with populist attitudes more often consult online news sources, contrary to what a populist narrative would suggest. Frequent consultation of more online sources is also associated with being younger, greater ability to use the internet, and more interest and participation in politics. In four of the countries, those with a strong ideological position were less likely to frequently consult more online sources.

Figure 40.1 shows the overall pattern of findings, indicating that as populist attitudes increase, the number of online political news sources consulted often or very often increases. Respondents were split into groups with high (+ 1 standard deviation above the mean), moderate, and

Table 39.5 Multiple regressions predicting number of online political news sources consulted often or very often1

DE

ES

ER

IT

PL

UK

US

Demographic factors

ß

ß

ß

ß

ß

ß

ß

Age

-.229***

-.101***

-.164“*

-.096*“

-.011

-.212*"

-.269***

Female

-.024

.008

-.045

.006

.017

-.041

-.062"

Education

-.034

.013

-.014

.021

-.074"

.011

-.030

Income

.000

.050*

.013

-.026

.028

.017

-.008

Online ability

Political orientation

.180***

.184***

.160***

.179***

.105"

.173***

.117***

Political interest

.154***

.193***

.134***

.137***

.210*"

.189***

.154***

Political participation

.287***

.230***

.349***

.244***

.273*"

.235*"

.338***

Strength of ideology

-.048*

-.070**

-.039

.010

.041

-.052*

-.075***

Populist attitudes

.098***

.055*

.067**

.080"

.124*"

.122*"

.128***

R2

.264***

.226***

.279***

.200***

.217*"

.287*"

.345***

N =

2000

2007

2000

2000

2005

2000

2018

Notes: Standardised coefficients displayed; ' p < .05; ** p < .01; "* p < .001

1 DE M=1.92, SD=1.92; ES M=2.82, SD=2.20; FR M=1.91, SD=2.05; IT M=2.32, SD=2.12; PL M=2.73, SD=2.18; UK M=1.80, SD=2.05; US M=2.18, SD=2.19

‘White’ and ‘born in country’ were removed from these analyses due to their lack of predictive power

Average number of online political news sources consulted often or very often

Figure 39.1 Average number of online political news sources consulted often or very often

Notes: Low populist = < populism mean score — 1 SD; high populist = > populism mean score + 1 SD

Means statistically different at ‘ p < .05; " p < .01; "* p < .001

One-way ANOVAs - DE: F(2,1856) = 3.80, p = .022; ES: F(2,1905) = 8.05, p < .001; FR: F(2,1823) = 7.37, p = .01; IT: F(2,1877) = 3.83, p = .022; PL: F(2,1872) = 20.10, p < .001; UK: F(2,1835) = 16.37, p < .001; US: F(2,1897) = 39.08, p < .001

low (- 1 standard deviation below the mean) levels of populist attitudes, with these groupings becoming factors in one-way ANOVAs. The average number of online political news sources that strong populists used often or very often is significantly different than the number of sources consulted frequently by weak or non-populists in all nations.

This finding adds further evidence that populists may not be trapped in online echo chambers or filter bubbles. Instead, they seek out a greater range of political information from different sources. This finding is in line with that of Schulz (2019), who found that populists in Europe and the United States were also more likely to frequently consult multiple sources of news.

We also found that populist attitudes are significantly associated with more frequently reading disagreeable news or political information in all seven nations (Table 40.6). The pattern of results, represented graphically in Figure 40.2 (with one-way ANOVAs included), is the same as with diverse political news consumption: stronger populist attitudes are related to more frequent consumption of disagreeable political content. Also, those less likely to look at disagreeable news or information about politics are younger, less skilled in using the internet, and less interested and participative in political activity (Table 40.6).

While not reported in tables here, we also found that, when controlling for demographic and political orientation variables, in five of the nations surveyed, those with populist attitudes more frequently accessed a more diverse set of online and offline political news sources. In four nations, populist responders said they more frequently checked sources different from what they normally read. And in all nations surveyed, they indicated more frequent and diverse use of online searches as well as more frequent participation in a more diverse set of online activities. All these findings reinforce the basic theme of populism not being a determinant of individuals being trapped in filter bubbles or echo chambers.

Table 39.6 Multiple regressions predicting frequency of‘reading something you disagree with’when looking for news or political information1

DE

ES

PR

IT

PL

UK

US

Demographic factors

ß

ß

ß

ß

ß

ß

ß

Age

-.140**'

1 1 c*** -.115

-.036

-.091”

-.107**

-.056*

-.078”

Female

.002

.008

-.074**

.016

-.001

.026

.008

Education

.022

.034

.006

.012

.017

.035

-.009

Income

.002

.007

-.005

-.009

.027

.029

.026

Online ability

Political orientation

.076**

.106***

.102***

.071*

.029

.085**

.031

Political interest

.228***

.318***

.322***

.200***

.263***

.231***

.158"*

Political participation

.192***

.122***

.174***

.161***

.133***

.184***

.213"*

Strength of ideology

.014

-.083***

-.050*

-.007

-.001

-.015

-.045

Populist attitudes

.065**

.067**

.062*

.096***

.089**

.110***

.086"*

R?

.169***

.206***

.237***

.133***

.139***

.177***

.133"*

N =

2000

2007

2000

2000

2005

2000

2018

Notes: Standardised coefficients displayed; * p < .05; ** p < .01; *** p < .001

1 DE M=3.11, SD=.98; ES M=3.20, SD=1.05; FR M=2.94, SD=1.09; IT M=3.27, SD=1.02; PL M=3.17, SD=. 99; UK M=3.15, SD=1.05; US M=3.26, SD=1.10

‘White’ and ‘born in country’ were removed from these analyses due to their lack of predictive power

Low populist ■ Moderate populist ■ High populist

Figure 39.2 Average frequency of‘reading something you disagree with’ when looking for news or political information

Notes: Scale is l=never, 2=rarely, 3=sometimes, 4=often, 5=very often

Low populist = < populism mean score — 1 SD; high populist = > populism mean score + / SD

Means statistically different at * p < .05; "p < .01; *** p < .001

One-way ANOVAs - DE: F(2,1853) = 1.71, p = .180; ES: F(2,1903) = 7.39, p = .01; FR: F(2,1852) = 8.71, p < .001; IT: F(2,1886) = 6.18, p = .002; PL: F(2,1882) = 8.46, p < .001; UK: F(2,1839) = 13.94, p < .001; US: F(2,1905) = 17.07, p <.001

 
Source
< Prev   CONTENTS   Source   Next >