Populism from below:

The ideological and social correlates of populist attitudes in Greece

Emmanouil Tsatsanis

CIES-Instituto Universitário de Lisboa, Portugal

[etsats@gmail.com]

Ioannis Andreadis

Aristotle University of Thessaloniki, Greece

[john@polsci.auth.gr]

Eftichia Teperoglou

Aristotle University of Thessaloniki, Greece

[efteperoglou@polsci.auth.gr]

Paper prepared for the 11th ECPR General Conference

Oslo, 6-9 September 2017

Please do not cite or distribute without permission of the authors

2

Abstract

Even though Greece has a long tradition in populist politics, the paper offers a first time

investigation of populism at the mass level. The study relies on a new battery of questions

specifically designed for this purpose and included in the Hellenic National Election Study

(ELNES) of January 2015. The objectives of the paper are to explore the ideological and

social correlates of populism, to investigate how populist attitudes relate to vote choice in

the post-crisis political landscape in Greece, and to highlight possible differences between

right and left-wing populist attitudes. Findings reveal that the strongest predictors of

populism at the mass level are ideological and not structural factors and the divide

between populist and mainstream parties that exists at the party system level is replicated

at the voter level. Furthermore, the impact of xenophobic (and specifically anti-immigrant)

attitudes on populism appear to be mediated by the classic categories of left-right

ideological identification, thus providing evidence that “populism from below” in Greece

comes in both inclusionary and exclusionary varieties.

Keywords: Greece, populism, SYRIZA, economic crisis, austerity; left-wing populism, right-

wing populism

Introduction

The literature on populism has experienced rapid growth in recent years. Following the

global economic crisis of 2007-2008, there has been a notable surge of anti-establishment

political forces, mainly in Europe and North America, which are almost universally

identified as “populist” by academic scholars and observers in the media. In most of the

studies, populism is analyzed from a “political supply-side” perspective, i.e. it is approached

as a communicative strategy, discursive style, or a set of organizational features and policy

preferences that pertain primarily to political parties and leaders. Attempts to examine

populism from the perspective of “political demand” have been far more rare. Even though

the first attempt to analyze populist attitudes at the micro level goes back to 1967 (see

3

Axelrod, 1967; for more details see Akkerman, Mudde and Zaslove, 2014: 1328) only

recently we find more systematic and analytical approaches (Hawkins, Riding and Mudde,

2012; Stanley, 2011; Akkerman, Mudde and Zaslove, 2014). Finally, there are some recent

comparative studies attempting to create a profile of voters of populist parties (see

Pauwels, 2014; Rooduijn 2017).

Our study aims to contribute to this ongoing research about studying populism at the micro

level. This is accomplished by focusing on the Greek case using data from the national

election study of January 2015. The study of Greek populism has received considerable

scholarly attention after the onset of the economic crisis and the electoral success of

populist parties. The two primary exponents of populism in Greece nowadays are the two

partners of the coalition government; the Coalition of Radical Left (SYRIZA) and the

nationalist right-wing party of Independent Greeks (ANEL). These two parties despite their

big programmatic differences on economic and, particularly, on cultural and social issues,

have formed coalition governments in the aftermath of the two national elections of 2015.

The paper pursues three main research objectives. The first and overarching objective is to

explore the ideological and social correlates of populism. The populist attitudes become the

explanandum as we try to identify the factors which facilitate the adoption of populist

worldviews at the individual level. The second objective is to investigate how populist

attitudes relate to vote choice in the post-crisis political landscape in Greece, while the

third one is to highlight possible differences between right and left-wing populist attitudes

(considering the distinction between “exclusionary” vs “inclusionary” populism by Mudde

and Kaltwasser 2013).

The structure of the paper is as follows. In the next section, we will briefly present the

theoretical discussion regarding populism and its conceptual components. Following the

conceptual discussion, we present the roots of Greek populism and its evolution during the

economic crisis. This is followed by a presentation of our hypotheses, data and method of

analysis. In the final part of the paper we present the main findings of our analyses,

followed by some concluding remarks.

4

Conceptualizing populism: From definitional proliferation to the

minimal definition

In the pioneering study on populism Ionescu and Gellner (1967) address that even though

there is no doubt about the importance of populism, no one is quite sure what it is. Over the

subsequent decades even if there is a massive wave in the literature on studies about

populism, there is lack of consensus about its definition (Pappas 2016).The notorious

difficulties in arriving at a widely accepted definition of populism stem from a failure to

settle basic ontological questions about the concept as well as the diversity of historical

phenomena to which the label of populism has been attached. The two problems are not

unrelated to one another, as the former is primarily driven by the latter, whilst the wide

(and mostly pejorative) use of the term in colloquial language further obfuscates the

conceptual landscape. The concept of populism has been deployed to describe phenomena

as disparate as the Narodnik movement in Russia, Alberto Fujimori’s presidency in Peru, or

the National Front party in France. The incongruity between such phenomena exists not

just at the historical and contextual level, but also in terms of units of analysis. Thus,

depending on the specific phenomenon under study, populism has been conceived as a

discourse (Laclau, 2005), mode of political incorporation (Mouzelis, 1985), communication

style (Jagers and Walgrave, 2007), political style (Moffitt and Tormey, 2014), thin-centred

ideology (Stanley 2008, Mudde 2007), interpretive frame (Tsatsanis, 2011: 13-15;

Aslanidis, 2015), or mode of political mobilization (Jansen, 2011). The list is by no means

exhaustive; similar lists are offered in almost every other publication on populism and are

a testament to the heterogeneity that characterizes this particular area of research.

There have been some recent attempts to provide a minimal definition of populism which

could identify a common core among the different conceptualizations, with the hope of

facilitating more comparative approaches that transcend narrow understandings of the

phenomenon associated with specific historical and regional contexts. The most influential

attempt, which has the advantage of “travelling well” across regional or temporal contexts,

belongs to Cas Mudde. According to Mudde (2004), populism can be conceived as a thin-

centred ideology that considers society to be ultimately separated into two homogeneous

5

and antagonistic camps, the ‘pure people’ versus the ‘corrupt elite’, and which argues that

politics should be an expression of the general will of the people (Mudde, 2004: 543). The

three core conceptual components of the definition, namely the corrupted elite, the

virtuous people, and the general will of the homogeneous, undifferentiated “people”

constitute a “lowest common denominator” supposedly present in all manifestations of

populism, thus addressing with the famous chameleonic character of the phenomenon.

The conceptualization of populism as a thin-centered ideology permits the alignment of

populism with different fully fledged ideologies, such as socialism, conservatism or

neoliberalism. The advantages of a minimal definition permits a reduced number of

attributes, while the range of cases is more amplified (Mudde and Kaltwasser 2013: 149).

Additionally, it allows the identification of different subtypes of populism that flourish in

different contexts. These different subtypes can have different peripheral characteristics

but maintain a common core. Specifically, Cas Mudde and Cristóbal Kaltwasser (2013) used

the terms “exclusionary” vs. “inclusionary” populism to originally distinguish between the

right-wing populism that has become prevalent in Europe from the left-wing populism that

is more common in Latin America. The basis of their framework of analysis is composed by

three dimensions of exclusion and inclusion: material, political and symbolic. The first one

is about the distribution of state resources. In the case of exclusionary populism, particular

groups are excluded from state resources (e.g welfare provisions), while the opposite

occurs for inclusionary populism. In the latter case, underprivileged social groups are

targeted to receive more state resources, thus creating a link between the concepts of

populism and clientilism through patterns of material inclusion (Mudde and Kaltwasser,

2013: 158-159). Similarly, the political dimension refers to patterns of exclusion versus

inclusion of particular groups in terms of the political representation and participation.

Finally, the symbolic dimension is not related to any specific policy domain but is

responsible for outlining the conceptual boundaries between ‘the people’ and ‘the elite’. In

the case of inclusionary populism, the criteria used to define the people and construct the

friend/foe distinction are primarily socioeconomic: definitions of the ‘people’ typically

include middle and lower socioeconomic strata, and they can also include socially excluded

groups, such as ethnic minorities and immigrants. In contrast, exclusionary populism

6

emphasizes cultural elements and constructs a double boundary around the notion of the

“common people”: it excludes not just the political and economic elites (including

“intellectuals”) but also specific cultural outgroups (which can be ethnic, religious or racial

minorities, legal and/or illegal immigrants etc.). Therefore, the notion of the people in this

case is not primarily socioeconomic but ethnicized, excluding ‘alien’ people and their values

(Mudde and Kaltwasser, 2013: 166). The perceived common characteristic between “the

elite” and the other excluded groups is a shared lack of connection and commitment to the

national community. The alleged disenfranchisement of the “common people” as well as the

threat posed to their physical and economic security by foreign elements is often explained

in terms of an “elite conspiracy”, through which the latter pursue their own interests and

further a nebulous “globalist agenda” at the expense of the national community.

Even though the inspiration for the distinction between inclusionary and exclusionary

populism have been the two main varieties of populist supply that are met in Latin

American and West European party systems respectively, our goal is to examine the

applicability at the mass level in a South European country where both types of populism

have been able to recently flourish at the party system level.

The roots of Greek populism and its evolution during the economic crisis

Scholarly attention to populism in Greece predates the post-crisis explosion of interest on

populism in Europe. Its study has historical roots which could be traced at least back to the

legacy of the seven-year military regime (Voulgaris, 2002). The interest in the phenomenon

of Greek populism in the 1980s and 1990s was intertwined with the post-authoritarian

political landscape and the ascend of PASOK to power in 1981. The legacy of the first two

terms of PASOK in government, from 1981 to 1989, has led some scholars to

retrospectively label the 1980s as the ´populist decade´ (Clogg 1993). The party expressed

(and developed further) a people-centric rhetoric more effectively than any other political

party during the 1980s and early 1990s. However, some scholars saw the populist

phenomenon as a more durable component of Greek political culture. For some, populism

constituted a type of ideological and political discourse introduced into Greek political life

7

by PASOK and its leader (e.g. Lyrintzis, 1987; Mavrogordatos, 1983), while in another study

the populist phenomenon is associated with an enduring component of Greek political

culture, named as “underdog culture” (Diamandouros, 1994)1.

The electoral success of PASOK in the 1980s coincided with the “populist decade” (see

Lyrintzis and Spourdalakis, 1993) and it was attributed to a large extent to the deployment

of a typical populist or “national-populist” discourse (e.g Pantazopoulos, 2001: 305-322)

that interpreted the social and political spheres as sites of struggle between elites and the

people or, in the rhetoric of PASOK, between the “privileged” and the “underprivileged”,

clearly distinct from more typical class-based leftist rhetoric. The causes of the resonance

of this rhetoric were to be found in the characteristics of the Greek social structure. Late

industrialization, responsible for the absence of a robust working class in Greece, combined

with the large size of the traditional petty bourgeoisie and the existence of a bloated public

sector linked to clientilist recruitment practices have been common sociological

explanations of the populist phenomenon in Greece (e.g. Lyrintzis, 1987). The election of

Kostas Simitis as leader of PASOK in 1996 signaled the abandonment of populist rhetoric in

favor of a more technocratic and modernizing discourse, more closely resembling West

European social democratic discourse of the period.

In the 1990s, Greece experienced new waves of populist mobilization with more

pronounced nationalist overtones, combining the usual anti-elitist and majoritarian

rhetorical devices of populist discourse, with xenophobic, conspiratorial and culturally

1 He distinguished between a “reform” and an “underdog” culture and he developed a theoretical framework

of “cultural dualism”. The “underdog culture” is influenced by Orthodox Christianity, is characterized by

isolationism, by an aversion towards the free market, by paternalism, statism and latent authoritarianism

and, finally, by overt xenophobia. The xenophobic component includes in its core, inter alia, a penchant for

conspiratorial interpretations of events, a siege mentality concerning the place of Greece within the

international environment, a Manichean division of the world into “philhellenes” (those who are considered

as friends of Greece) and “mishellenes” (those who are enemies of Greece) along with an exaggerated sense of

the importance of Greece in the international system and in the history of western civilization. This culture

slowly entrenched itself among the more traditional, inward looking and least competitive strata and sectors

of Greek society. Conversely, the intellectual roots of the “reformist” culture harken back to the European

Enlightenment: it is secular, friendly towards the free market, supports modernizing reforms in every domain

of social life, is more outward looking, promoting the creation and expansion of international links (see

further Diamandouros 1994).

8

defensive themes (Tsatsanis, 2011: 15-16). During this period the acceleration of the

European integration process together with a massive influx of immigrants to the country

following the collapse of communist regimes in Eastern Europe reinforced the feelings of

uncertainty and insecurity arising from the “threat of multiculturalism”, allowing

xenophobic and nationalistic sentiments to surface more openly than before. The decade

was marked by massive mobilization campaigns over issues such as the demand for Greek

exclusivity to the name “Macedonia” (which came under question after the former Yugoslav

republic of the same name acquired independence), blatantly one-sided pro- Serbian rallies

during the Bosnia and Kosovo wars and opposition to the decision to adhere to EU legal

standards and remove the registration of religious affiliation from state identity cards

(Stavrakakis, 2004).

A legacy of this period has been the creation of the Popular Orthodox Rally (LAOS), the first

political party in Greece that squarely belonged to the new populist radical right party

family in Europe (Mudde, 2007). The party produced typical right-wing populist rhetoric,

focusing on issues such immigration, antisemitism, and anti-americanism, along with a

conspiratorial view of history and contemporaneous political developments. Moreover, the

party repeatedly defended the nationalist and religious identity of Greek orthodoxy

(Pappas and Aslanidis, 2015). The presence of LAOS on the political scene increased the

salience of its flagship issues in Greek political discourse – hostility towards immigration,

protection of the national and religious culture, as well as Euroscepticism (represented also

by KKE).

Populism in the years of crisis: ubiquitous and polyprismatic

A new phase in the study of Greek populism starts with the onset of the economic crisis

(Pappas and Aslanidis, 2015). The political consequences of the economic crisis in Greece

have been seismic and have been documented by various studies (e.g Teperoglou &

Tsatsanis, 2014; Verney, 2014; Teperoglou et al., 2015; Karyotis and Rüdig, 2015; Tsatsanis

& Teperoglou 2016). One of the effects of the crisis has been the intensification and

renewal of populist discourses in the country. The effects of the austerity measures

9

prescribed by Greece's international lenders and implemented by consecutive Greek

governments has been an intensely debated topic. It has dominated political discourse from

the spring of 2010 onwards, giving rise to the neologisms “mnemoniakoi” and “anti-

mnemoniakoi” to distinguish between those that saw the 'bailout' and the attached MoU2 as

a lifeline for the Greek economy and those that viewed them as the actual reason why the

economic crisis has been so deep and protracted. The rapidly shifting socio-economic

landscape coupled with the massive loss of confidence towards political institutions led to

the eradication of stable party loyalties and the emergence of new fault lines in the

electorate (Teperoglou et al., 2015: 336). The intensification of populist rhetoric went hand

in hand with what appeared to be in the first years of the crisis a full blown regime

legitimacy crisis, manifested in historic lows of trust in government and representative

institutions, the rise of antidemocratic parties such as Golden Dawn - a party that according

to Hawkins, Rovira Kaltwasser and Andreadis (2016), among others, is against democracy

and therefore not populist - as well as the massive populist mobilizations of the

“aganaktismenoi” (indignants) movement (Karyotis and Rüdig, 2015).

In the transformed Greek party system one can find representatives of populism both from

the left and the right of the political spectrum, creating a more pronounced inter-party

dynamic along the populist dimension. A coalition government between a party of the

radical left (SYRIZA) and a party of the nationalist right (ANEL) would have been

unthinkable before the economic crisis upended the old party system. However, such a

coalition indirectly provides some credence to the claim that “inclusionary” and

“exclusionary” forms of populism share a common core. Their shared rejection of the

bailout agreements, as well as their common critique of high-level corruption and of vested

interests connected to the two former major parties, had made them behave as potential

coalition partners for months in advance (Tsatsanis and Teperoglou, 2016). The

unprecedented coalition governments between PASOK and New Democracy between 2012

and 2015, which were both a product and a catalyst for the further restructuring of the

ideological and political space in Greece, had unwittingly already set the stage for the

2 Memorandum of Understanding, mnemonio in Greek.

10

emergence and collaboration of “anti-establishment” parties against the parties of the old

establishment.

However, points of differentiation still remain that underline the diametrically opposed

ideological legacy and political pedigree of the two parties. The populist discourse

produced by Alexis Tsipras, the leader of SYRIZA and Greek PM, and other high level

members of the Greek government that belong to SYRIZA, even after the radical left party

became the main party in government, continues to emphasize themes of socioeconomic

inequalities between ‘the people’ and the ‘elites’, by which he mostly refers to the

previously governing political class and their allies .The struggle against the foreign

creditors and the EU/ECB/IMF ‘troika’ is framed as a struggle primarily against

neoliberalism. The notion of the people is mostly de-ethnicized and inclusionary, with

frequent inclusive references to immigrants and minority groups. On the other hand, the

perception of the ‘people’ in the narrative of the leader of ANEL Panos Kammenos (and

minister of defense in the government) is typically exclusionary and ethnicized (Markou

2017). On the material dimension, Kammenos adopts the familiar right-wing populist

language of welfare chauvinism, often targeting immigrants and minorities as groups that

drain resources from the Greek welfare state and away from the more deserving ‘Greeks’.

Moreover, the struggle of the Greek government against its foreign creditors is primarily

framed in nationalist and conspiratorial terms combined with a stringent anti-German tone

(with frequent references to a supposed establishment of a sinister “Fourth Reich” in

Europe). However, the discourse of both political camps converges when singling out the

‘corrupt domestic oligarchy’ and the old political establishment as the main foes of the new

coalition government. Thus a main objective this paper is to identify the degree to which

such elective ideological and rhetorical affinities - that already exist at the level of party

elites - can be traced at the mass level as well.

11

Hypotheses

As stated in the beginning, the goal of the paper is to investigate the relationship between

populist attitudes with social and ideological factors, as well as to gauge the association

between populist attitudes and voting. Based on the theoretical discussion and empirical

findings from comparative or other case-studies, as well as the historical and qualitative

research literature about populism in Greece, we formulate certain testable hypotheses.

The first four hypotheses relate to the effects of social and ideological factors on populist

attitudes. Our first main expectation is that individual-level characteristics that are

associated with less dynamic, competitive and outward-looking social strata, such low

levels of education, employment in low-skilled occupations (e.g. unskilled manual or

service labor) or in protected industries or sectors of the economy (e.g. public sector) will

tend to be positively associated with populist attitudes. This hypothesis largely rests on

the (mostly speculative) scholarly work concerning the social penetration and the carriers

of populist ideology in Greece, as well as the comparative literature concerning the

increasingly politicized divide between the so-called “winners” and “losers” of globalization

(Kriesi et al. 2008). According to this approach, the losers of globalization are those people

whose life chances were relatively protected before the advent of globalization by the

existence of national institutional frameworks, such as welfare states, neo-corporatist

structures of interest intermediation or clientelist modes of political inclusion, that were

designed to mediate the worst aspects of market competition. Globalization has brought

increased capital and labor mobility, periods of increased access to capital markets but also

periods of harsh conditionality in the form of austerity policies as Greek society has been

made painfully aware in the aftermath of the 2008 global financial crisis. This means that

the individuals most ill-equipped to deal in an environment of increased pressures from

global competition and decreased public spending are the most likely to experience a

decline in life chances and turn against “establishment” parties, politicians and political

platforms. One of the main empirical referents of this divide between the “winners” and

“losers” of globalization are levels of education and specialization that provide “exit

options” in an age where mobility becomes a crucial determinant of social stratification

(Kriesi et al., 2008: 5).

12

H1. Individual-level characteristics that are associated with the less dynamic and competitive

strata of Greek society are expected to be positively associated with populist attitudes

Even though it has been argued that populism in Greece is a phenomenon that historically

cross-cuts the left-right ideological axis and that political entrepreneurs employing

populist rhetoric often tend to reject classic left and right ideological labels, the political

forces that are currently associated the most with populist and generally anti-

establishment rhetoric hail either from the left (e.g. SYRIZA) or the right (ANEL, Golden

Dawn) ideological camp. Furthermore, given that these forces tend to jointly reject the

mainstream politics of centrist political forces (mainly represented by the two major

parties of the old two-party system, i.e. PASOK and New Democracy), we expect that the

relationship between populism and left-right ideological identification will be a curvilinear

one, with left-wing and right-wing ideological identification to be associated with stronger

populist attitudes in relation to the degree of populism for individuals that place

themselves on or near the center of the ideological spectrum.

H2. Populist attitudes will tend to increase at the two ends (left and right) of the ideological

spectrum

Similarly, following the literature on recent and historical populism in Greece, we expect

that statist attitudes on matters of economic ideology and policy will tend to be overall

positively associated with populist attitudes, as well as anti-immigrant and Eurosceptic

attitudes. The latter expectation is based on the theoretical discussion concerning the links

of populism to xenophobic, protectionist and nationalist sentiments. However, given the

comparative literature on subtypes of populism, we expect that statist attitudes will have a

stronger impact among left-wing individuals (inclusive populism) whereas anti-immigrant

attitudes will have a stronger impact on right-wing individuals (exclusive populism).

H3. Statist, anti-immigrant and Eurosceptic attitudes are expected to be positively associated

with populist attitudes

13

H4a. The effect of statist attitudes on populism will tend to be stronger among left-wing

individuals

H4b. The effect of anti-immigrant attitudes on populism will tend to be stronger among right-

wing individuals

The second part of the empirical investigation relates to the association between party

choice in elections and populist attitudes. One aspect of this investigation concerns the

degree of congruence between populist parties and voters (Andreadis and Stavrakakis,

2017). It is expected that voters of parties that have become the primary exponents of

populist discourse in the years of the crisis, such as SYRIZA and ANEL (Stavrakakis and

Katsabekis, 2014; Katsabekis, 2015; Pappas and Aslanidis, 2015), will be more likely to

adopt and express populist attitudes, especially in relation to the two parties of the former

coalition government, New Democracy and PASOK. Furthermore, following the discussion

on the inclusive and exclusive subtypes of populism, we expect that the association

between statist and populist attitudes will be stronger among voters of left-of-center

political parties and that the association between anti-immigrant and populist attitudes

will be stronger among voters of right-of-center parties.

H5. SYRIZA and ANEL voters display a higher propensity for adopting populist attitudes in

relation to voters of the parties of the ND-PASOK former coalition government

H6a. Statist attitudes will be positively associated with populism for voters of all parties, but

the association will be stronger for voters of left-wing and center-left parties (KKE, SYRIZA,

PASOK)

H6b. Anti-immigrant attitudes will be positively associated with populism for voters of all

parties, but the association will be stronger for voters of right-wing and center-right parties

(New Democracy, ANEL, Golden Dawn)

14

Data and Methodology

In order to test our hypotheses, we use data from the Hellenic (Greek) Voter Study for the

Parliamentary elections of January 2015 (ELNES 2015)3. To allow the study of the impact

of populism on Greek electoral politics, ELNES 2015 includes a battery of populist attitudes

questions (see Table A1 in the Appendix) which takes into account the minimal definition

of populism (Mudde, 2007), while it relies mainly on the work by Hawkins & Riding (2010)

(for an overview of the progress towards the final formulation of this battery, see

Stavrakakis, Andreadis and Katsambekis, 2016; Hawkins, Riding, and Mudde, 2012;

Hawkins, Rovira Kaltwasser and Andreadis, 2016). In order to gauge the dimensionality,

scalability and reliability of the populism scale, we include in our analysis additional issue

items that were originally used for the voting advice application HelpMeVote 2015 and

where re-included in ELNES 2015. These capture the most important issues of political

competition in Greece at the beginning of 2015 and include economic attitudes,

sociocultural attitudes, attitudes towards the EU, as well as attitudes towards the first two

bailout deals4 (see Table A2 in the Appendix).

Given that we use Likert type items, we follow van der Eijk and Rose (2015) and we apply

Mokken Scale Analysis (van Schuur, 2003) using the R package mokken (Van der Ark,

2012). Mokken scale analysis (MSA) is a scaling technique for ordinal data, and it consists

of two parts: The first part, known as automated item selection procedure (AISP), is an

exploratory method for constructing scales from ordinal variables. AISP is a hierarchical

clustering algorithm that starts by constructing the first scale with the two items that have

the highest item-pair scalability coefficient (Hij) and it keeps adding items to the scale until

3 The survey is part of the Hellenic National Elections Studies (ELNES) (see Andreadis et al. 2015 and

Andreadis 2016a). It is a mixed-mode survey conducted by the Laboratory of Applied Political Research,

School of Political Sciences at Aristotle University of Thessaloniki. The recruitment process for ELNES 2015

lasted from June 12 until July 16, 2015 using RDD (Random Digit Dialing). The 1008 completed cases were

collected either as web-based self-administered questionnaires or using telephone interviews (CATI). The

data are weighted according to population parameters such as gender, age, education level, region and vote

choice (Andreadis, 2016b). Table A4 in the appendix shows the distribution of valid votes in the weighted

sample and it is compared with the distribution of the valid votes from the official election results. 4 We have included questions about the bailout to the high politicization and visibility of this issue since the

agreement of the first bailout in May 2010. This issue has saturated political discourse in Greece during the

years of the crisis and was the dominant issue during the January 2015 electoral campaign.

15

there are no more items that meet the criteria (the homogeneity coefficient lower bound

for inclusion in a scale is usually set at Hi>0.3). Then, the procedure is repeated for the

remaining unselected items, until no more scales can be formed. The second part of MSA

consists of methods to investigate the goodness-of-fit for each of the Mokken scales

constructed in the first part. If each item of the scale has a coefficient Hi>0.3, then the scale

will be unidimensional. For the reliability of the populist scale, in addition to Cronbach's

alpha, we present three more reliability statistics available in the mokken package.

The Mokken scale analysis returns three scales (see Table A3 in the appendix). The first

scale includes issues related to the economy, the bailout agreements between Greece and

the so-called Troika as well as attitudes towards the European Union. The second scale

includes seven of the eight populist attitudes items with larger values indicating stronger

populist attitudes. The populism scale can be considered unidimensional as it passes the

test of homogeneity (the H-score is well above 0.3) and it can be considered reliable (all

reliability statistics are above 0.8).The third scale includes socio-cultural items: higher

values indicate conservative and authoritarian attitudes, including anti-immigrant

attitudes.

In order to test H1 and H3 we perform a linear regression model with the populism scale

returned from the Mokken scale analysis used as the dependent variable (rescaled from 0

to 10). Given the limited research on the predictors of populist attitudes at the mass level,

we employ in our explanatory model a range of demographic, occupational, cognitive and

ideological variables. We use gender, age and education (9 categories) as controls, as well

as a several occupational dummy variables (6 occupational groups, employment in the

public sector, union membership, unemployment), area of residence (1: rural area, 4: large

town or city), religiosity (1: no religious beliefs, 4: very religious), political knowledge (0:

no correct responses, 4: 4 correct responses), household income (1: up to €10,000, 5:

€40,000+), and left-right self-placement (0:left, 10: right). We also include two scales for

statist (items eco1, eco2, eco3, eco4 and eco5 in Table A2, all reversed; Cronbach’s alpha:

0.713) and anti-immigrant attitudes (items soc3, soc6, soc7, soc12 in Table A2; Cronbach’s

alpha: 0.728). For Eurosceptic attitudes we include the item which captures degree of

16

agreement with the proposition to increase the power of the European parliament (item

euro3 in Table A2; 1: completely agree, 5: completely disagree). All non-binary variables

(with the exception of left-right self-placement that is used in the next model to plot

interactions) are mean subtracted and divided by two standard deviations in order to

facilitate comparisons between coefficients (Gelman, 2008).

We have run the model both on the weighted and on the unweighted sample and have

gotten similar results. But as Freedman and Berk (2008) argue, if a regression model is

properly specified, it is better to fit the model on the unweighted sample, because

weighting is likely to increase the bias in estimated causal parameters. Thus, we have

selected to present the linear regression output on the unweighted sample. After exploring

the direct effects of our variables, we run the same model including interaction terms

between left-right ideology with statist and anti-immigrant attitudes respectively in order

to test H4a and H4b. Finally, we gauge the curvilinearity of the relationship between

populism and left-right ideology (H2) by creating a quadratic prediction plot.

In order to examine H5, we trace the mean locations of voters by party choice on the scale

of populist attitudes that is constructed following the Mokken scale analysis. The goal is to

gauge in a straightforward way the relationship of voting choice and populist attitudes and

see if our hypotheses concerning a divide on the populist dimension between the parties of

the former and the current coalition governments in Greece holds. In order to test H6a and

H6b we run additional quadratic prediction plots on the subsamples of the voters of each of

the main 7 parties (KKE, SYRIZA, PASOK, To Potami, New Democracy, ANEL and Golden

Dawn) for the relationship of statist and anti-immigrant attitudes with populist attitudes.

Findings

Our baseline regression model provides some evidence that only marginally confirm

aspects of H1 as well as evidence of more unequivocal confirmation of H3 (see Figure 1 and

Table A5 in the appendix). Our main expectation, following the relevant but mostly

17

speculative literature, is that individual-level characteristics that tend to place someone in

the less dynamic and competitive strata of society will tend to be associated with populist

attitudes. Such characteristics are low-levels of education, employment in low-skilled

occupations or in protected industries or sectors of the economy (e.g. public sector).

However, none of the occupational variables present a statistically significant impact on

populist attitudes, failing to confirm H1. Education is negatively associated with populist

attitudes as expected, but this relationship is not significant at the p<.05 level (even if very

close to reaching significance at the p<.10 level). The only sociodemographic variable that

has a clear negative effect on populism is that of income. As expected, household income

presents a strong negative relationship with populist attitudes.

[Figure 1 about here]

Conversely, statist, anti-immigrant and Eurosceptic attitudes (along with religiosity) all

present very strong positive effects on populist attitudes, confirming H3. Of the three, the

effect size of statist attitudes is by far the largest as a movement of one standard deviation

in statist attitudes produces an increase of almost 13 percentage points on the populism

scale.

H2 is only partially confirmed as the relationship between left-right ideology and populist

attitudes presented in Figure 2 clearly shows that populism in Greece is primarily

associated with left-wing ideological identification in relation to both right-wing and

centrist ideological self-placement. The strong negative effect of the left-right scale on

populist attitudes in our baseline model is thus confirmed even if the relationship is not

completely linear. This relationship between leftist identification and populism could

perhaps be explained by the fact that anti-elitism and the ‘underdog’ political culture has

been historically more closely linked to political forces on the left of the political spectrum

(e.g. PASOK in the early post-authoritarian period) and right-wing populism is a more

recent phenomenon on the Greek political stage. An additional explanation is that right-

wing populist entrepreneurs tend to reject self-categorizations on the left-right dimensions

more strongly than their leftist counterparts (for example, the public rejection of the left-

18

right dimension is common among both Golden Dawn and ANEL leadership members) and

that resistance to an identification with classic ideological categories at the right end of the

political spectrum might filter down to the mass level.

[Figure 2 about here]

Moving on the regression model which includes interaction terms (see Figure 3 and Table

A5 in the appendix), H4a can clearly be rejected, as the effect of statism on populist

attitudes remains equally strong across the left-right ideological spectrum, for both left-

wing as well as right-wing respondents. Conversely, the interaction between anti-

immigrant attitudes and left-right ideology reveals a clear pattern of difference between

inclusive and exclusive populism in Greece, confirming H4b. The effect of anti-immigrant

attitudes on populism becomes statistically significant and stronger as we move towards

the right-end of the ideological spectrum, whereas the effect is completely neutralized at

the left-end.

[Figure 3 about here]

Moving on to the examination of the association between vote choice and populist

attitudes, we see that H5 is confirmed (see Figure 5 and Table A6 in the appendix). SYRIZA

and ANEL voters score on average the highest on the dimension of populist attitudes,

confirming H2, whereas New Democracy and PASOK (along with To Potami) voters score

the lowest, which reproduces the distance between the two groups of parties that exist at

the level of political elites (Stavrakakis et al., 2016). However, the gap between

“establishment” and “populist” voters should not be exaggerated given that the distribution

of voters of all parties remains for the most part within the “populist” half of the scale. In

other words, there is no populist vs anti-populist divide within the Greek electorate. A

better way to describe this divide would be between “more” and “less populist” voters. In

fact, the populism polarization index is rather low (0.16) when compared with the

Economic/EU/Bailout polarization index (0.29), which is the main polarization line of the

Greek electorate in 2015 (Stavrakakis and Andreadis, 2017)

19

[Figure 4 about here]

Another interesting finding relates to the position of Golden Dawn and KKE voters. These

two parties often employ populist rhetoric but are not classified as populist in the

literature: Golden Dawn is an antidemocratic fascist party and KKE is an orthodox

communist one. However, the voters for Golden Dawn and KKE occupy a space closer to the

populist end of the index together with SYRIZA and ANEL voters, meaning that their voters

share to a large extent the same anti-elitist attitudes with the voters of the two parties

whose leadership is more often classified as “populist” in the relevant literature (e.g.

Pappas and Aslanidis, 2015). One explanation for this is the level of anti-elitist rhetoric

adopted by the two parties. Even though anti-elitist rhetoric is not synonymous with

populism (anti-elitism is after all only one component of populist rhetoric), according to

2014 Chapel Hill Expert Survey data, on the 0-10 scale which measures salience of anti-

elitist party discourse, KKE and Golden Dawn recorded the highest mean values.

[Figure 5 about here]

[Figure 6 about here]

Finally, the evidence fails to confirm either H6a or H6b (see Figure 5 and Figure 6).

Admittedly, the quadratic prediction plots provide more cursory evidence instead of

rigorous tests of the hypotheses. Even though the findings per party are informative, no

clear pattern emerges in terms of right-wing and left-wing parties. Statist attitudes are of

course associated with voting for left-wing parties, but the effect of statism on populism

appears to be strong in both left-wing (SYRIZA) and center to right-wing parties (New

Democracy, ANEL, To Potami) but for other parties (PASOK, KKE, Golden Dawn) this

relationship does not hold. Of course, the small sample sizes, especially concerning the

smaller parties, do not permit a high level of certainty concerning the replicability of the

observed patterns in bigger samples. The effect of anti-immigrant attitudes in populism

appear to be more uniform across parties, but clearly more pronounced for the two parties

20

located at the right-end of the political spectrum, ANEL and Golden Dawn. The effect is

much weaker for voters of New Democracy, KKE and PASOK, whereas a U-shaped

curvilinear relationship appears to exist for voters of SYRIZA and To Potami (both extreme

anti-immigrant and pro-immigrant attitudes are associated with higher populism values),

showing evidence that both inclusive and exclusive populism can coexist within the same

party voter base.

Concluding remarks

This paper offered one of the first investigations of populist attitudes at the mass level in a

country with a long pedigree of populist politics as well as populism studies. Taking

advantage of the growing interest on populist internationally and the ensuing development

of new research tools for the study of populism at the mass level, we were able to test

empirically some long-standing assumptions about populism in Greece. Our first finding is

that structural factors do little to explain variation in populist attitudes, unlike certain

economic and cultural ideological orientations that seem to constitute far better predictors

of populism at the mass level. Furthermore, the impact of xenophobic (and specifically anti-

immigrant) attitudes on populism appear to be mediated by the classic categories of left-

right ideological identification, thus providing evidence that “populism from below” in

Greece comes in both inclusionary and exclusionary varieties.

This relationship between populist attitudes and vote choice constitutes our second major

finding. It confirms the expectation that voters of the two paradigmatic populist parties in

the post-crisis political landscape, SYRIZA and ANEL, display a higher level of populist

attitudes in relation to voters of the “establishment” parties of the former coalition

government (ND and PASOK), thereby providing some initial evidence concerning the

emergence of a new political divide between populist and anti-populist that cuts across the

historically dominant left-right dimension of party competition in Greece. In general, the

evidence makes clear that populism in Greece is not simply a binary worldview that

informs the thinking and shapes the communication strategies of certain political

21

entrepreneurs but frames the way in which a very large part of the Greek electorate relates

to politics and understands the way society works. The lack of comparable historical data

cannot allow to easily answer questions about whether such high levels of “populism from

below” existed in Greece prior to the crisis or to what extent it is the creation (rather than

the stimulus) of political supply factors. However, future studies using comparable data can

help answer questions as to whether populism at the mass level constitutes an enduring

characteristic of Greek politics or whether, like other attitudinal trends, it is sensitive to

changes in the political environment.

References

Akkerman, A., Mudde, C., & Zaslove, A. (2014), ´How Populist Are the People? Measuring

Populist Attitudes in Voters´, Comparative Political Studies, 47(9): 1324–1353.

Andreadis, I. & Stavrakakis, Y. (2017), ‘European Populist Parties in Government: How Well

are Voters Represented? Evidence from Greece’, Swiss Political Science Review,

doi: 10.1111/spsr.12255

Andreadis, I. (2016a), Web based surveys Greek Election 2015, Ann Arbor, MI: ICPSR

[distributor]: http://doi.org/10.3886/E100074V16

Andreadis, I. (2016b), ‘Weights for the 2015 Hellenic Voter Study’. Ann Arbor, MI: Inter-

university Consortium for Political and Social Research.

http://doi.org/10.3886/E74764V1

Andreadis, I., Kartsounidou, E. and Chatzimallis, C. (2015), ‘Innovation, an Answer to Lack

of Funding: The 2015 Hellenic National Election Voter Study.’

http://www.polres.gr/en/sites/default/files/elnes2015a.pdf.

Aslanidis, P. (2015), ‘Is populism an ideology? A refutation and a new perspective’, Political

Studies, published online before print December 11, 2015, doi: 10.1111/1467-

9248.12224

Axelrod, R. (1967), ´The structure of public opinion on policy issues´, The Public Opinion

Quarterly, 31(1): 51–60.

22

Clogg, R., ed.(1993), Greece, 1981-89: The Populist Decade. London: Palgrave Macmillan

Diamandouros, N. (1994), ´Cultural dualism and political change in postauthoritarian

Greece´, Estudios Centro de Estudios Avanzados en Ciencias Sociales, Instituto Juan

March de Estudios e Investigaciones Working Paper 50.

Freedman, D. A., & Berk, R. A. (2008), ‘Weighting regressions by propensity scores’,

Evaluation Review, 32(4), 392-409.

Gelman, A. (2008), ‘Scaling regression inputs by dividing by two standard deviations’.

Statistics in Medicine, 27(15): 2865-2873.

Hawkins, K. A., & Riding, S. (2010), ‘Populist Attitudes and their Correlates among Citizens:

Survey Evidence from the Americas’, Paper presented at the ECPR Workshop

‘Disassembling Populism (and Putting It Back Together Again): Collaborative

Empirical Research on Interactions among Populism’s Attributes’, March 22-25,

2010 ,Muenster, Germany.

Hawkins, K. A., Rovira Kaltwasser, C. and Andreadis, I. (2016), ‘The Activation of Populist

Attitudes: Evidence from Contemporary Chile and Greece’, Team Populism January

Conference: The Causes of Populism. January 28-30, 2016, Provo, Utah.

Hawkins, K., Riding, S. & Mudde, C. (2012), Measuring Populist Attitudes. Political Concepts,

Committee on Concepts and Methods, Working Paper Series, 55.

Ionescu, G., & Gellner, E. (eds) (1969). Populism: Its meanings and national characteristics.

London: Weidenfeld and Nicolson.

Jagers, J. and Walgrave, S. (2007), ‘Populism as political communication style: An empirical

study of political parties’ discourse in Belgium’, European Journal of Political

Research, 46(3): 319–345.

Jansen, R. S. (2011), ‘Populist mobilization: A new theoretical approach to populism’,

Sociological Theory, 29(2): 75–96.

Karyotis, G. and Rüdig, W. (2015), ‘Blame and punishment? The electoral politics of

extreme austerity in Greece’, Political Studies, 63: 2–24.

23

Katsambekis, G. (2015), ‘Left-wing and Right-wing populisms. A comparison of SYRIZA and

the Independent Greeks (ANEL) in contemporary Greece’, Paper presented at the

PSA 65th Annual International Conference, 30 March-1 April, Sheffield.

Kriesi, H., Grande, E., Lachat, R., Dolezal, M., Bornschier, S. and Frey, T. (2008), West

European Politics in the Age of Globalization. Cambridge: Cambridge University

Press

Laclau, E. (2005), On populist reason. London: Verso.

Lyrintzis, Ch. (1987), 'The power of populism: the Greek case', European Journal of Political

Research 15(6): 667-686.

Lyrintzis, Chr. and Spourdalakis, M. (1993), 'On Populism: A Synthesis à propos of the

Greek Bibliography', Hellenic Political Science Review, 1: 133-62. (In Greek).

Markou, G. (2017), 'The Rise of Inclusionary Populism in Europe: The Case of SYRIZA ',

Contemporary Southeastern Europe, 4(1): 54-71.

Mavrogordatos, G. Th. (1983), ´Rise of the Green Sun: The Greek Election of 1981´, Centre of

Contemporary Greek Studies, King's College London, Occasional Paper I.

Moffitt, B. and Tormey, S. (2014), `Rethinking Populism: Politics, Mediatisation and Political

Style´, Political Studies, 62: 381–397.

Mouzelis, N. (1985), ´On the Concept of Populism: Populist and Clientelist Modes of

Incorporation in Semiperipheral Polities´, Politics & Society 14: 329-348.

Mudde, C. (2007), Populist Radical Right Parties in Europe. Cambridge: Cambridge

University Press.

Mudde, C. and Kaltwasser, C.R. (2013), ‘Exclusionary vs. inclusionary populism: Comparing

contemporary Europe and Latin America.’ Government & Opposition, 48(2): 147-

174.

Muddle, C. (2004), ‘The populist zeitgeist.’ Government & Opposition, 30(3): 541-563.

Pantazopoulos, A. (2001), 'For the People and the Nation'. The Moment Andreas Papandreou

1965-1989, Athens: Polis Publications (in Greek).

24

Pappas, T. (2016), ‘Modern populism: Research advances, conceptual and methodological

pitfalls, and the minimal definition.’ Politics: Oxford Research Encyclopedias. Online

publication date: March 2016, DOI: 10.1093/acrefore/9780190228637.013.17

Pappas, T. S., & Aslanidis, P. (2015), ´Greek populism: A political drama in five acts.´ In H.

Kriesi & T. S. Pappas (Eds.), Populism in the shadow of the great recession,

Colchester, U.K.: ECPR Press. pp. 181–196.

Pauwels, T. (2014), Populism in Western Europe: Comparing Belgium, Germany and the

Netherlands, New York: Routledge.

Rooduijn, M. (2017), ‘What unites the voter bases of populist parties? Comparing the

electorates of 15 populist parties’, European Political Science Review, doi:

https://doi.org/10.1017/S1755773917000145

Stanley, B. (2008), ´The thin ideology of populism´, Journal of Political Ideologies, 13(1): 95–

110.

Stanley, B. (2011), ´Populism, nationalism, or national populism? An analysis of Slovak

voting behavior at the 2010 parliamentary election´, Communist and Post-

Communist Studies, 44: 257-270.

Stavrakakis, Y. & Katsambekis, G. (2014), ‘Left-wing Populism in the European Periphery:

The Case of SYRIZA’, Journal of Political Ideologies, 19(2): 119-142.

Stavrakakis, Y. (2004), ‘Antinomies of formalism: Laclau’s theory of populism and the

lessons from religious populism in Greece’, Journal of Political Ideologies, 9(3), pp.

253–267.

Stavrakakis, Y. and Andreadis, I. (2017), ‘Dynamics of Polarization in the Greek Case’. Paper

presented at the ‘Workshop on Polarized Polities: Causes, Consequences and

Solutions’. Central European University, Budapest.

Stavrakakis, Y., Andreadis, I. and Katsambekis, G. (2016), ‘A new populism index at work:

identifying populist candidates and parties in the contemporary Greek context’,

European Politics and Society, pp. 1–19. doi: 10.1080/23745118.2016.1261434.

25

Teperoglou, E. Tsatsanis, E. and Nicolacopoulos, El. (2015), ´Habituating to the new

normal in a post-earthquake party system: The 2014 European Election in Greece´,

South European Society and Politics, 20 (3): 333-355.

Teperoglou, E. and Tsatsanis, E. (2014), ´Dealignment, de-legitimation and the implosion of

the two-party system in Greece: The earthquake election of 6 May 2012´, Journal of

Elections, Public Opinion, and Parties, 24 (2): 222-242.

Tsatsanis, E. (2011), ‘Hellenism under siege: The national-populist logic of antiglobalization

rhetoric in Greece.’ Journal of Political Ideologies, 16(1): 11-31.

Tsatsanis, E. and Teperoglou, E. (2016), ´Realignment under Stress: The July 2015

Referendum and the September Parliamentary Election in Greece´, South European

Society and Politics.

Van der Ark, L. A. (2012), ‘New Developments in Mokken Scale analysis in R’, Journal of

Statistical Software, 48(5), 1-27.

Van der Eijk, C., & Rose, J. (2015), Risky business: factor analysis of survey data–assessing

the probability of incorrect dimensionalisation. PloS one, 10(3),

doi:10.1371/journal.pone.0118900

Van Schuur, W. H. (2003), ‘Mokken Scale Analysis: Between the Guttman Scale and

Parametric Item Response Theory’, Political Analysis, 11(2): 139-163.

Verney, S. (2014), ‘Broken and can’t be fixed: the impact of the economic crisis on the Greek

party system’, The International Spectator: Italian Journal of International Affairs,

vol. 49, no. 1, pp. 18–35.

Voulgaris, Y. (2002), Post-Authoritarian Greece 1974–1990, Athens: Themelio (in Greek).

26

Figures

Figure 1. Effects of predictors on populist attitudes

Data: ELNES 2015

Note: Point estimates and error bars represent regression coefficients and 95% confidence intervals

27

Figure 2. Populist attitudes by left-right ideological identification

Data: ELNES 2015

Note: In the quadratic prediction plot the solid line represents fitted values and the dashed lines represent

95% confidence intervals

28

Figure 3. Marginal effects of statist and anti-immigrant attitudes on populism by left-right

ideological identification

Data: ELNES 2015

Note: Solid lines represents average marginal effects and dashed lines represent 95% confidence intervals

Figure 4. Populism by January 2015 vote choice

Data: ELNES 2015

Note: Point estimates and error bars represent mean values and 95% confidence intervals. Results are

weighted by gender, age, education, region and vote.

29

Figure 5. Populism by level of statist attitudes (by January 2015 vote choice)

Data: ELNES 2015

Note: In the quadratic prediction plots the solid lines represent fitted values and the dashed lines represent

95% confidence intervals

30

Figure 6. Populism by level of anti-immigrant attitudes (by party)

Data: ELNES 2015

Note: In the quadratic prediction plots the solid lines represent fitted values and the dashed lines represent

95% confidence intervals

31

Appendix

Table Α1. Populist attitudes items Item Statement

pop1 The politicians in parliament need to follow the will of the people.

pop2 The people, and not politicians, should make our most important policy

decisions.

pop3 The political differences between the elite and the people are larger than the

differences among the people.

pop4 I would rather be represented by a citizen than by a specialized politician.

pop5 Elected officials talk too much and take too little action.

pop6 What people call ‘compromise’ in politics is really just selling out on one’s

principles.

pop7 Popular demands are today ignored in favour of what benefits the

establishment.

pop8 Political forces representing the people should adopt a more confrontational

attitude in order to make their voice heard and influence decision-making. Source: ELNES 2015

32

Table Α2. Other items included in Mokken Scale Analysis Variable Reversed Text

eco1 We should have more flexible forms of work in order to combat

unemployment.

eco2 The government should take measures to reduce income inequalities

eco3 It must be possible to operate non-governmental, non-profit institutions of

higher education.

eco4 The national health system can become more efficient through partial

privatization.

eco5 Yes Banks and utilities must be under public control.

mem1 The Memorandum has not caused the economic crisis; the economic crisis

has resulted in the Memorandum

mem2 Yes With the Memoranda we accumulate debts without any visible benefits.

mem3 Memoranda of Understanding with the Troika were necessary to avoid the

bankruptcy of Greece.

mem4 Yes We have every right to cancel the debt without consulting anyone else

mem5 We ought to have done many of the changes provisioned in the Memoranda

on our own long ago.

euro1 It is better for Greece to be in the European Union rather than outside.

euro2 Yes The economy of Greece would have been better if we had our own currency

instead of Euro.

euro3 The decision power of the European Parliament should be increased on all

matters of internal and foreign policy.

euro4 Yes The probability of GRexit should not be considered as a disaster

soc1 Yes The right for temporary leaves from prison is more important than the risk

of escape

soc2 People who break the law should be given stiffer sentences.

soc3 Yes Immigrants are good for [country’s] economy.

soc4 The police should use stricter enforcement measures to protect the property

of citizens.

soc5 Defense spending should not be reduced to avoid becoming a vulnerable

country.

soc6 The requirements for asylum and citizenship must be tightened.

soc7 Yes The existence of multiculturalism in Greece is a positive phenomenon.

soc8 Same-sex marriages should be prohibited by law.

soc9 Yes Women should be free to decide on matters of abortion.

soc10 Yes The church and the state should be completely separated.

soc11 There should be legislation to limit protests

soc12 Immigrants should be required to adapt to the customs of Greece Source: ELNES 2015

33

Table Α3. Mokken scale analysis and reliability tests

Economy/EU/

Bailout

Populism Sociocultural

H se H se H se

eco1 0.314 0.025

eco2*

eco3 0.366 0.025

eco4 0.416 0.022

eco5 0.507 0.018

mem1 0.391 0.022

euro4 0.479 0.020

mem2 0.530 0.018

euro1 0.502 0.022

mem5 0.479 0.020

euro2 0.495 0.018

euro3*

mem3 0.557 0.017

mem4 0.503 0.019

pop1 0.464 0.021

pop2 0.498 0.019

pop3 0.379 0.024

pop4 0.349 0.027

pop5*

pop6 0.449 0.022

pop7 0.513 0.019

pop8 0.453 0.023

soc1*

soc10 0.311 0.023

soc11**

soc12 0.302 0.023

soc2*

soc3 0.371 0.020

soc4 0.352 0.023

soc5 0.332 0.024

soc6 0.463 0.018

soc7 0.440 0.019

soc8 0.417 0.019

soc9*

34

Scale H 0.459 0.015 0.444 0.016 0.374 0.016

MS 0.899 0.822 0.805

Alpha 0.894 0.817 0.805

lambda.2 0.896 0.821 0.809

LCRC 0.905 0.818 0.819

* the item has not met the Mokken scale criteria

**the item has entered the first scale

Table Α4. Distribution of valid votes in the sample Party N % (of valid

votes)

Actual % vote in

election

SYRIZA 274 36.6 36.3

ND 193 25.8 27.8

GD 30 4.0 6.3

RIVER 67 8.9 6.1

KKE 47 6.3 5.5

ANEL 40 5.3 4.8

PASOK 43 5.7 4.7

Other 55 7.3 8.5

Source: ELNES 2015

35

Table A5. Determinants of populism Variables Baseline With interactions

Β (s.e.) Β (s.e.)

Male .085(.108) .078(.108)

Age .401(.126)*** .388(.127)***

Education -.220(.142) -.231(.141)

Class (Ref=“Other”)

Managers -.028(.232) -.041(.229)

Professionals .021(.179) .021(.174)

Technicians .362(.269) .363(.266)

Clerical support workers -.077(.234) -.064(.229)

Service and sales workers -.073(.273) -.065(.268)

Manual workers -.066(.257) -.053(.256)

Public sector -.164(.147) -.167(.149)

Unemployed .091(.167) .100(.169)

Union membership -.227(.208) -.243(.211)

Urban residence -.123(.113) -.120(.114)

Religiosity .385(.118)*** .424(.118)***

Income -.549(.128)*** -.537(.128)***

Left-Right self-placement -.138(.031)*** -.143(.032)***

Political knowledge -.195(.132) -.192(.132)

Euroscepticism .252(.110)** .258(.109)**

Statism 1.28(.128)*** 1.18(.280)***

Anti-immigrant attitudes .501(.125)*** .039(.295)

Left-Right*Statism .010(.053)

Left-Right*Anti-immigrant .102(.061)*

Constant 7.54(.257)*** 7.21(.229)***

N 644 644

R2 .370 .375

***:p <0.01; **:p < 0.05; *:p < 0.10

Data: ELNES 2015

Note: Robust standard errors in parentheses

36

Table A6. Populist mean scores by vote-choice (voter-level) and anti-elitism scores (party-

level)

Party ELNES 2015 CHES 2014

Populism (Mean) [95% confidence interval] Anti-elitism (mean)

KKE 7.33 6.67 7.99 9.78

SYRIZA 7.72 7.45 8.00 8.56

PASOK 6.08 5.22 6.94 2.78

To Potami 5.60 5.27 5.93 5.33

New Democracy 5.88 5.46 6.30 2.33

ANEL 8.42 7.52 9.32 9.22

Golden Dawn 7.26 6.66 7.85 10

Data: ELNES 2015 and Chapel Hill Expert Survey (CHES) 2014

Note: 0-10 scales for both populism and anti-elitist scores. ELNES 2015 results are weighted by gender, age,

education, region and vote.