Populism is one of the most frequently used and discussed terms in the past years. Given that there is no academic consensus on a definition, the following paper aims to empirically research the way populism levels in speeches are influenced by certain contextual factors, and what this does say about the character of populism. For that purpose, two opposite hypotheses are drawn from the ideational and political-strategic approach and tested by analyzing the Global Populism Database. I find that speech type, office, term, political orientation, and region influence the populism level, but the analysis can be interpreted in favor of both definitions of populism.
Keywords: Populism, Ideational Approach, Political-Strategic Approach, Global Populism Database, Political Speeches
Introduction: Populism everywhere?
Populism is without a doubt one of the most discussed terms when it comes to politics nowadays. From accusations against political adversaries to journalistic essays, we can easily get the impression that populism is everywhere. Usually correlated with a negative connotation, the word is so easily used in everyday life that when we ask the question “What is populism?”, no clear answer comes to mind. The idea of populism became even more prominent since the campaign and presidency of Donald Trump, with some even blaming populism as being the driving force that led to the assault of the Capitol in January 2021 (Viala-Gaudefroy, 2021). Some have even accused Trump’s successor, Joe Biden, of being more populist than his predecessor (Ganesh, 2021). Populism became a buzzword, which think tanks and research foundations started to pay attention to and publish about (Hartleb, 2018).
If the term is often underdefined in day-to-day politics and media, one might turn to the academic sphere in search of a clear explanation of what populism is. The scientific community lacks a definitory consensus on populism. The published research papers on the topic employ a multitude of definitions, three of them being dominant in the field: the cultural, ideological, and strategic definition (Espejo et al., 2017, pp. 12, 14). However, it is striking that more than half of the articles published cannot be put into any category, mainly because they do not use a definition, but rather a vague conceptualization (Ibid, p.12). As most of the research is focused on specific countries or regions (Ibid, p.2), the motivation of this paper is to try to find some empirical clues for which definition of populism might be better suited to explain populism generally.
Populism is most often brought to the discussion when looking at speeches, be it Trump’s rallies or far-right speeches in Europe. I use speeches as the unit of observation for this paper. For that, I approach the idea of holistic grading and the Global Populism Database. Speeches can be measured in terms of populism, but there are also a lot of contextual factors that can be observed. By looking at a large number of speeches, these factors, as well as the populism level, varies and the relation between these variations can be researched, especially how contextual factors influence the level of populism. As such, the research question is: What do the influences of contextual factors say about the character (ideology versus strategy) of populism? To answer this question, I statistically analyze the Global Populism Database, while also taking into account the two levels that can be observed: the public of the speech and the politician who holds it.
To do this, I first present the main two theories (strategic and ideational) to draw my hypotheses and the state of methodology when it comes to measuring populism in speeches. Then I turn to the Global Populism Database, the most comprehensive dataset on populism in speeches available to date, to define the contextual factors and populism level I take from it. The analysis is structured in two parts: a descriptive analysis of the variation of populism under each factor, and a multivariate regression analyzing all factors concomitantly. Lastly, I present the conclusions of this paper.
The results I come to are in no way one-sided: There are multiple factors that I test and which truly influence the level of populism in speeches. The type of speech, the office held by the speaker, the term, the world region and the political orientation influence the level of populism. I discuss these influences in my analysis by looking at them separately and then through multivariate regression. When it comes to the character of populism, it goes both ways, but the ideational approach seems to be better at explaining the behavior of the public, while the strategic one does slightly better when it comes to the politician holding the speech.
What do we know? A look at the current state of research and theory
Before starting to make assumptions and conducting research, a look at what the scientific literature has to offer is needed. Populism is one of the most controversially discussed concepts of political science, so defining it and narrowing it down to conduct research, and discussing different research results and theoretical approaches is required. I first present the minimal consensus on populism, then the ideational and strategic approach. Based on them, I draw two opposite hypotheses to test. The second part of this section is reserved for the different methodological developments in speech analysis when it comes to populism.
One important aspect of this paper is the use of the terms character and characteristics, which are different. To clarify them, under character I understand the definition of what populism is. Characteristics are the manifestations of populism, aspects that are empirically verifiable to determine if a speech is populist or not. While different approaches to populism measure it by using similar or the same characteristics, they differ in how they see the character.
To determine the lowest common denominator, I look at the characteristics that are accepted regardless of the theoretical approach. One of those is the divinization of the people (Gidron and Bonikowski, 2013). A more exact common feature is the discourse of a conflict between the will of the people and the will of the elite (Hawkins et al., 2019d). The presence of populism is often determined by calls for more direct elements in democracies, like referendums, and a mistrust of the ideals of liberal democracies like minority rights and professional politicians (Canovan, 2004). The difficulties of discussing populism revolve around the ambivalent relationship it has with democracies, also known as the democratic paradox: It calls for more democratic elements, but it also rejects the core feature of liberal democracies, the most common form nowadays (Canovan, 2002). It is a democratic driving force in autocracies, but a democracy decelerating or de-democratizing force in established democracies (Mudde and Kaltwasser, 2017, pp. 86-93). These ideas are influenced by the ideational approach, which I use to determine the general characteristics of populism, but shared by the other approaches. However, the main research of this paper involves the character of populism, which I also discuss through two theories.
Mudde defines populism as “an ideology that considers society to be ultimately separated into two homogeneous and antagonistic groups, ‘the pure people’ versus ‘the corrupt elite’, and which argues that politics should be an expression of the volonté générale (general will) of the people” (Mudde, 2004). The three main ideas this approach brings to the table are a Manichean worldview, the presentation of “the people” as a homogenous and moral group, and the existence of a powerful and corrupt elite (Hawkins and Kaltwasser, 2018). As this definition has common elements with the others and the explicit focus lies on speeches in this paper, I define the concept of populism as a discourse style that includes ideas such as a popular will (that is fought for by the speaker), the existence of a corrupt elite and the fight of the people against the elite. This establishes the characteristics of the concept, but not its character. Based on these three empirically verifiable speech themes the presence of populism in a certain speech can be confirmed. Conclusions about the character can be made only by adding and interpreting the variation of the contextual factors.
The ideational approach defines the character of populism as an idea or ideology. Because it does not have the explanatory power of a normal ideology, like nationalism or socialism, it is considered to be a thin ideology (Stanley, 2008). It only appears together with a host-ideology, like nativism, this being the reason why populism is so diverse and appears all around the political spectrum. Because of the vagueness of “the people”, the actual identity varies according to the host ideology. As “the people” is not an inclusive concept, as it has to be differentiated from the elite, it is rather an imagined community, much like how nationalists define the nation (Mudde, 2004; Anderson, 2020). Another term used in the scientific literature to define “the people” is the heartland, an imagined community where “a virtuous and unified people reside” (Taggart, 2000, p. 95). Populism as an ideology does not aim to change the “the people”, but rather the impact they have on politics: “The people” are seen as morally pure and the basis for political legitimacy, their will has to play the main role in policy-making (Mudde, 2004). When it comes to political leaders, they are a catalyst of already existing populist sentiments in the population (Hawkins and Kaltwasser, 2018, pp. 10-11). According to Hawkins and Kaltwasser, there are three main organizational forms in which populism manifests: institutionalized parties, social movements, and charismatic or personalistic movements (Ibid.). This means that the general population believes in populist ideas, which are only activated by the right context and/or leader.
Due to my focus on speeches, this means that the audience and the leader believe in the populist ideas that appear in the speech. Being part of their worldview, I expect politicians to use relatively similar levels of populism across all speeches, independent from the speech-specific context. This is an exaggerated claim, as there are virtually unlimited factors that can influence the content of discourses, but for the sake of this research paper I determine the hypothesis as H₁: The level of populism does not fluctuate based on different speech factors.
To further deepen the understanding of populism and the possibility of studying it empirically, I bring a second hypothesis to the table, one which should be different than or ideally opposite to H₁. For that, I turn towards the political-strategic approach to populism, but focus only on its differences from the ideational approach. According to Weyland, populism is “a political strategy through which a personalistic leader seeks or exercises government power based on direct, unmediated, uninstitutionalized support from large numbers of mostly unorganized followers” (Weyland, 2001). In this case, populism is more of a set of tools through which leaders try to gain, maintain and exercise power (Weyland, 2017, p. 8). It manifests itself mostly in big crowds like rallies, where the politician can show the support he or she has from “the people”. The main characteristics are a direct connection to “the people”, mostly through TV and social media, and the claim to be the embodiment of the will of the people (Ibid. p.12-13). While this theory makes extensive claims about the way populism works, I only focus on the implication for speech analysis. As said, populism tends to manifest itself when the politician is talking to “the people”, so in this case, in speeches that have a large audience, or that are expected to go viral through social media. As populism is a strategy for political gains, I expect higher levels of populism in speeches held in campaigns or those with large audiences. Therefore, I determine the opposite hypothesis as H₂: The level of populism fluctuates based on different speech factors, increasing when the circumstances allow for political and electoral gains.
The theories I presented are far more complex and explain more than just the way populism should function in speeches, but for my research question, this aspect plays the dominant role. The hypotheses are exaggerated to be opposite claims, as I want to get an idea of which one has more explanatory power when tested empirically.
When it comes to researching populism across different discourse types like speeches and party manifestoes, there are two main approaches in the scientific community: Content analysis and holistic grading. Content analysis usually involves coders who look for populist elements in discourses and measure them by paragraphs or computers that look for populist words based on a codebook and measure the exact number of words (Rooduijn and Pauwels, 2011). One of the more interesting ways qualitative content analysis has been used is in analyzing Trump’s tweets. Because a tweet is 140 characters or less, the way the analysis works is by looking for certain themes, like crime reduction or fake news (Lacatus, 2021). The main aspect of this method is the focus on words or paragraphs and not the speech as a whole.
The other main way to look at speeches is through holistic grading. The newest innovation in this area is the Global Populism Database (GPD), which is a continuation of quantitative research in this field. The approach of holistic grading has been used in the past couple of years in an ever-increasing effort. The first instance of it being used at the mass level in populism is a comparative study mainly focused on Latin America but also including a couple of politicians in office from other parts of the world (Hawkins, 2009). The measured categories are similar to the ones used in the GPD and ideal for categorizing speeches by certain characteristics like speech type. The current GPD was also used by The Guardian for researching the trends and changes in populist discourse over time (Lewis et al., 2019) and appears in the newest effort to measure and analyze populism together with nationalism (Jenne et al., 2021). The GPD uses holistic grading, which is a method used mostly in educational psychology in which a certain text is analyzed as a whole by human coders, who can notice idea patterns in speeches and then assign a grade in comparison to an anchor text (White, 1985). It is most suitable for measuring latent aspects of discourses that are only insufficiently captured by counting certain words or other methods. In comparison to content analysis, the GPD does not quantify words or paragraphs, but lets the coder interpret the message of the speech as a whole.
Most of the existing literature on populist speeches has been either used to determine whether a speech is populist or to gain an overlook of the global situation. The scope of this paper is to take it one step further, by using the already existing data but taking all possible influencing factors into account.
Operationalization through the Global Populism Database
To test the hypotheses, I use the Global Populism Database (Hawkins et al., 2019a). It analyses the level of populist discourse in 1.113 speeches held by politicians in office (Hawkins et al., 2019b). The reason I use this database is the multitude and diversity of speeches it analyses, allowing me to test the influence certain factors have on the level of populism, and the relative novelty of the database. The research of this paper is mostly driven by the GPD, as it is the most comprehensive dataset on populism in speeches worldwide available at the time of writing. Early tests also indicate a convergent validity and potential for new findings on the causes and consequences of populism (Hawkins et al., 2019d, pp. 9-12). As such, I try to explore what the contextual factors included in the dataset and the populism level can say about the explanatory power of the two approaches to populism.
Because the unifying aspect of all definitions of populism is ideas, populism also counts as a latent aspect. In this case, populism is measured on a 0-2 scale, with 0, 1, and 2 as anchor values: 0 means a speech lacks populist elements and most importantly does not contain a reference to popular will, 1 means the speech has clear populist elements that are tempered, and 2 is the closest it gets to the ideal populist discourse and includes most if not all characteristics of populism (Hawkins et al., 2019b, Hawkins, 2009). Speeches are analyzed in original form by multiple coders who understand the original language, but I only use a speech’s average grade in my analysis.
The independent variable in the hypotheses has been defined broadly as the contextual factors of a speech. While such a term can include a lot of measurable aspects, my analysis only takes into account categories that have been measured by the GPD:
Political Ideology: The left-right scale separates speeches by the political orientation of the leader in comparison to the unweighted mean of the database and includes the categories left, center, and right. Populism appears across the political spectrum, so this variable is rather exploratory. I connect it to the geographic location and check the findings against the published literature.
Office of the politician: either president or prime minister (PM). The dataset does not differentiate between parliamentary, presidential, and mixed polities. As such, conclusions regarding the office will be limited. Based on H2, I expect presidents to be more populist than PMs, as they are directly legitimated by the electorate. I also expect the results to be distorted because of the lack of differentiation in mixed systems. I will further discuss this in the analysis.
Term: The term in which the speech was delivered. I interact the variable with the office and according to H2, I expect more variation across terms for presidents than for PMs. In the case of H1, I expect no major differences.
Speech type: international speeches in a diplomatic gathering with foreign officials, ribbon-cutting speeches which have a relatively small audience, famous speeches, and campaign speeches (Hawkins et al., 2019a). According to H2, I expect the level of populism to increase in the order I presented the categories. According to H1, I expect no major variation across speech types.
Geographic Location: The database also includes a regional differentiation into 9 categories: Central and Eastern Europe, Central Asia, East Asia & Pacific, Latin America & Caribbean, Middle East & North Africa, North America, South Asia, Sub-Saharan Africa, and Western Europe (Hawkins et al., 2019c). The analysis will also consider an artificial variable of regions based on these 9 categories, with Central and Eastern Europe, Western Europe, and North America being considered as the Global North and the remaining 6 regions the Global South. Because it was coded using the 9 regions, it is a rather arbitrary variable that resembles the Brandt line (Tait, 2016, p. 1) only to a certain degree. While it is not a perfect geographical match and represents a simplified image of the differentiation between developing and developed countries nowadays, this variable has the sole purpose of facilitating the research of macro trends regarding populism.
Even though this database follows the tradition of the ideational approach, it is created specifically to be useful regardless of the approach used, as the unifying element across all theories of populism is the presence of ideas (Hawkins et al., 2019d). As a consequence, it is fit for testing the two hypotheses, even though they are drawn from different theories.
The variation of these variables is analyzed in comparison to the variance of the populism level. First, each one is considered separately through descriptive mean comparison and if an interaction effect is discovered, it is also visually presented and compared. In the end, all of them are brought together through multivariate regression analysis. The whole analysis has been done through R (R Core Team, 2021) and RStudio (RStudio Team, 2021), and the supporting graphs and tables have been created with the help of ggplot2 (Wickham, 2016) and stargaze (Hlavac, 2018).
Analysis: Which factors influence populism?
I first begin by looking at the characteristics of speeches and their effect on the populism level separately and possible interaction effects between them. As most of them are coded as categorical variables, the main tool for analysis is the mean comparison, aided by visual depictions. I interpret the mathematical differences by assessing how well the hypothetical predictions have been confirmed or contradicted, to get insight into the character of populism. While the main focus of the analysis is the perspective of the politician holding the speech, in the places where the data allows it, I also conclude what populism is to the audience/ regular citizens. For interpretation of the populism level, I use the Codebook description and the 4 categories established by The Guardian: 0-0,5 not populist, 0,5-1 somewhat populist, 1-1,5 populist, and 1,5-2 very populist (Lewis et al., 2019).
4.1. Speech Type
At the first look at Figure 1, there seems to be a significant fluctuation in the level of populism by speech type. International speeches have the lowest mean of 0,198 and the median of 0. The standard deviation lies at 0,43 so most speeches lie under the 0,5 threshold and are not populist. The same applies to ribbon-cutting speeches, which have only a slightly higher mean of 0,21. Based thereon, I conclude that politicians rarely use populist elements like popular will in international and ribbon-cutting speeches, and if they do, they are usually tempered by dominant non-populist elements.
The reality of the other two categories is different: famous speeches have a mean of 0,43 and a median of 0,1. The standard deviation is also significantly higher, at 0,59. As a consequence, while most of them are still relatively low on the populism scale, some famous speeches surpass the 1 threshold and slightly move in the direction of 1,5, meaning that they have clear populist elements that are usually, but not always, tempered. The increase is even more striking in campaign speeches. With a mean of 0,5995 and a standard deviation of 0,62, the average campaign speech analyzed has references to popular will and other elements. While the overall level of populism across all categories is fairly moderate with no mean surpassing 1, there is an important gap between the first and last two categories.
All these insights support the second hypothesis and point in the direction of populism as a political strategy. The first two categories of speeches, where the audience is either small or consists of foreign officials, are way less populist than the other two. This can be explained by a lack of opportunity for political gains, so the incentive to strategically use populism is also lower. For international speeches, the lack of populism is also confirmed by research on populism’s impact on foreign policy, which showed that the host ideology has a great influence but populism has a rather marginal role (Verbeek and Zaslove, 2017). It follows logically that campaign speeches are the most populist. The database used either the announcement or closings (Hawkins et al., 2019b) which usually have the most media coverage, so the possibility of gaining momentum is high. On the other hand, it also implies that there is an expected readiness or proneness to a populist discourse in the audience, suggesting that populist ideas are part of the citizen’s mindset and implicit ideology. The same applies to famous speeches, as the politician cannot know beforehand which speech is going to be viral. This shows that there is at least more appreciation, discussion, and sharing of rather populist speeches, hinting in the direction of ideology in the audience.
4.2.
Office of the Speaker
The second characteristic I look at has more to do with the politician rather than the speech. The Global Populism Database includes speeches by presidents and prime ministers. Figure 2 shows a clear distinction between them: The PM category has a mean of 0,25, the President a mean of 0,447, and a higher standard deviation (0,6). A prime minister is also less likely to hold a populist speech in comparison to a president. Interestingly enough, the median for president is 0,15 and the mean-median difference is high, meaning that most presidents do not have a significantly populist discourse, but the ones who do are extremely populist. I also found an interaction effect between term and office as explanatory variables that can be seen in Figure 3. Due to the properties of the variable term, the graph can also be interpreted as a time plot. For PM, there is relative stability over the terms, as the average difference for the mean from one term to the other is 0,017. There is a striking difference when it comes to presidents. It seems that a president’s populism level increases with the term. It is continuous for the first three terms, gets lower at the fourth, and skyrockets at the sixth. As there are only three units of investigation in the sixth term with extremely high populism values, they can be interpreted as distortions through extreme values.
The interpretation of these values leads to mixed results, especially because it has to take the political system of the country into account and the selection by the GPD is rather unclear: For presidential systems like the US there are speeches by presidents, for parliamentary ones like the United Kingdom and Germany, speeches by the prime minister or chancellor. The confusion appears in rather mixed systems: For France, only the president is included, but for Romania, a semi-presidential system that modeled its constitution after the French, both the president and the prime minister are present. What the data does show is that indirectly elected leaders (prime ministers) have a constant level of populism, meaning that populism is rather an ideology. In presidential systems, the level of populism tends to increase over time, which is fairly strange as populism is an anti-elitist and outsider discourse. But this insight leads in the direction of populism as a way for elected officials to hold on to power in elections when they lack political results during the term, blaming failures on a hidden economic and political string-drawing elite and promoting conspiracy theories, such as Trump’s administration and campaign showed concerning the deep state or rigged elections (Porter, 2020, Bossie, 2020). In the research literature, there have been empirical findings indicating that belief in such conspiracy theories is correlated to populism (Castanho Silva et al., 2017), even regarding the Trump example (Langer, 2017), which further suggests an ideological component for ordinary citizens.
4.3. Geography and Political Orientation
When it comes to geography, there are also clear differences between world regions. While analyzing each region at a time would require a lot of focus on the characteristics of each one and is not the scope of this research paper, I created the artificial variables Global North and Global South to be able to analyze a simplified version of the regions. What can be certainly observed is that there is a clear difference between the 9 regions, meaning that cultural and societal aspects influence the level of populism politicians use in speeches. When it comes to the two big regions as seen in Figure 5, the Global South’s average speech is more populist (0,43) than in the North (0,3), but the standard deviation in the Global South is also higher and both categories have a mean in the non-populist interval.
Comparing the Global North and South can better be done by involving another variable, political ideology, with which it interacts. The variable is categorical, but I put it on
an axis in Figures 6 and 7 for better visual clarity. Nonetheless, only the mean values can be interpreted. Generally speaking, the average speech by a leftist politician is the most populist at 0,413, but the difference is not that high (0,323 for center and 0,327 for right). A better picture is revealed when checking for an interaction effect, as shown in Figure 7. There are inverted trends: Populism levels in the Global South tend to decrease the more right-wing the politician is, as opposed to the Global North, where they increase. While the average value for all categories in the Global North and center and right in the Global South do not surpass the 0,5 threshold, there’s a striking difference when considering the left in the Global South. The mean value of 0,697 and the 3. Quartile at 1,5 show extremely high values compared to the rest, which also means that most of the speeches have clear populist elements and one-third of them are close to the ideal populist discourse. This is further confirmed by the extensive literature on recent left-wing populism in the regions of the Global South, especially Latin America (Conniff, 2020, pp. 5-6).
This insight can be interpreted both ways: Because populism and especially left-wing populism are so common in the Latin American society, there is a higher chance that the politician has a populist mindset, but they also could use populism as a strategy for political gains because of the existence of populist sentiments in the population. As the Global South is a vague variable, and there are extensive differences between the countries it encompasses, this only shows a general macro trend.
4.4. All Factors Considered Together – The Multivariate Regression
Finally, to bring the explanatory variables together and assess the importance of every one of them, I create a multivariate regression as shown in Table 1. As most of the variables were categorical, I coded them as dummy variables and only considered certain ones. For political ideology, office, and geography, I include all categories. As most variables are dummies, for each factor there needs to be a reference variable: Center for political ideology, PM for office, Global North for geographic locations. In the case of the speech type, the categories international and ribbon-cutting behave rather similar and without significant variation, so I use them both as a reference and only directly include the categories famous and campaign. The main focus lies on the two interaction effects.
Using those variables, a significant linear regression was found, as can be seen in Table 1. The regression has a 20% degree of determination, which is satisfactory, being given that there are virtually unlimited characteristics that could play a role in the level of populism.
When it comes to political ideology, the variables Right and Left are not statistically significant, but that might be because of the interaction effect. When taking the geographic location into account, the level of populism is the lowest in the Global South, when the speech is not held by a center or right-wing politician and the highest level of populism appears when left-wing politicians in the Global South hold speeches, if all other variables are held constant.
The second interaction effect shows that statistically, Presidents with more terms tend to hold more populist speeches. Speech categories behave as expected: Famous speeches are more populist than the reference, and campaign speeches are the most populist.
The variables behave the way I expect after looking at them separately. The two interaction effects, the two speech categories included, and the Global South variable are all statistically significant and with R2 at 20%, meaning that these factors explain an important part of the variation of the level of populism.
One point of critique I have to address regarding my research and the Global Populism Database is the absence of speeches held by opposition politicians. Populist parties nowadays usually do not win the absolute majority in elections, and if they join the government, they are usually the junior coalition partner. So, there are relatively few populist presidents or prime ministers in comparison to the overwhelming presence populism has in protests and smaller parties. This could be one reason for the low level of populism throughout the analysis. For further research, an updated database that includes opposition politicians and personalities from protest movements and/or civil society would be beneficial.
Conclusion
The character of populism is one of the most discussed aspects of political science, with no clear consensus emerging in the scientific literature. Even though the ideational approach has established itself as the definition that is mostly used in research, alternative definitions are still present. This paper did aim to ultimately demonstrate which definition is the most valid but tried to test two idealized opposite hypotheses based on a limited dataset that measured a finite number of characteristics of political speeches to find more clues regarding the meaningfulness of the ideational and strategic approach. The paper does not claim to completely analyze populist speeches empirically but rather test the two theories with the most extensive quantitative dataset available at the time of writing.
The main findings of this research are not pointing in a certain direction, they are rather pointing in both directions. The tendency of popular speeches to be populist points to an appreciation of populist elements by the public, while the difference between campaign and other speeches shows an instrumentalization of populism for political gains. When looking at political offices and, implicitly, political systems, I can conclude that presidential systems or directly elected offices tend to foster increasingly populist discourses, while parliamentary systems show relatively stable levels of populism. This can also mean that in the case of presidents, populism is used as a tool for elections, while for prime ministers it is rather an ideology because it remains constant. The difference in different regions of the world can be interpreted in both ways: Different sociopolitical spaces have different ideological compositions, meaning populism is rather an ideological component, but at the same time, leaders can instrumentalize the existing populism to gain power. This is also shown when considering the host-ideology or host-direction (left, right, or center) and the political history of the Global North and South.
Based on this, emerges the question: Can populism not be both? This is a question that should be considered in further research by looking at the two different levels of leaders and the public. This paper shows that in the case of leaders, the strategic approach tends to explain the phenomenon slightly better, while the ideological approach has way more explanatory power when it comes to the public. This implies that populism is even more of a complex phenomenon, and therefore requires further research.
Literature
ANDERSON, B. 2020. Imagined Communities: Reflections on the origin and spread of nationalism. The New Social Theory Reader. Routledge.
BOSSIE, D. 2020. Anonymous NY Times op-ed shows Deep State thugs are working against the will of the American people. Fox News.
CANOVAN, M. 2002. Taking Politics to the People:Populism as the Ideology of Democracy.
CANOVAN, M. 2004. Populism for political theorists? Journal of Political Ideologies, 9, 241-252.
CASTANHO SILVA, B., VEGETTI, F. & LITTVAY, L. 2017. The Elite Is Up to Something: Exploring the Relation Between Populism and Belief in Conspiracy Theories. Swiss Political Science Review, 23.
CONNIFF, M. L. 2020. A Historiography of Populism and Neopopulism in Latin America. History Compass, 18, e12621.
ESPEJO, P. O., KALTWASSER, C. R., TAGGART, P. & OSTIGUY, P. 2017. Populism: An Overview of the Concept and the State of the Art. Oxford University Press.
GANESH, J. 2021. The controlled populism of Joe Biden. Financial Times.
GIDRON, N. & BONIKOWSKI, B. 2013. Varieties of populism: Literature review and research agenda.
HARTLEB, F. 2018. The Rise of New-Populism in Europe and Asia. Konrad-Adenauer-Stiftung.
HAWKINS, K. 2009. Is Chávez Populist? Comparative Political Studies, 42, 1040-1067.
HAWKINS, K. A., AGUILAR, R., CASTANHO SILVA, B., JENNE, E. K., KOCIJAN, B. & ROVIRA KALTWASSER, C. 2019a. Global Populism Database, v1. V1 ed.: Harvard Dataverse.
HAWKINS, K. A., AGUILAR, R., CASTANHO SILVA, B., JENNE, E. K., KOCIJAN, B. & ROVIRA KALTWASSER, C. 2019b. GPD Codebook.pdf. Global Populism Database, v1. V1 ed.: Harvard Dataverse.
HAWKINS, K. A., AGUILAR, R., CASTANHO SILVA, B., JENNE, E. K., KOCIJAN, B. & ROVIRA KALTWASSER, C. 2019c. GPD_20190625.tab. Global Populism Database, v1. V1 ed.: Harvard Dataverse.
HAWKINS, K. A., AGUILAR, R., SILVA, B. C., JENNE, E. K., KOCIJAN, B. & KALTWASSER, C. R. Measuring Populist Discourse: The Global Populism Database. EPSA Annual Conference in Belfast, UK, June, 2019d. 20-22.
HAWKINS, K. A. & KALTWASSER, C. R. 2018. Introduction: the ideational approach. The ideational approach to populism. Routledge.
HLAVAC, M. 2018. stargazer: Well-Formatted Regression and Summary Statistics Tables.
JENNE, E. K., HAWKINS, K. A. & SILVA, B. C. 2021. Mapping Populism and Nationalism in Leader Rhetoric Across North America and Europe. Studies in Comparative International Development, 56, 170-196.
LACATUS, C. 2021. Populism and President Trump’s approach to foreign policy: An analysis of tweets and rally speeches. Politics, 41, 31-47.
LANGER, G. 2017. Nearly half of Americans think there’s a ‘deep state’: Poll. ABC News.
LEWIS, P., BARR, C., CLARKE, S., VOCE, A., LEVETT, C. & GUTIÉRREZ, P. 2019. Revealed: the rise and rise of populist rhetoric. The Guardian.
MUDDE, C. 2004. The Populist Zeitgeist. Government and Opposition, 39, 541-563.
MUDDE, C. & KALTWASSER, C. R. 2017. Populism: A very short introduction, Oxford University Press.
PORTER, T. 2020. Mick Mulvaney said the deep state is ‘100% true,’ and government officials should obey Trump or quit. Business Insider.
R CORE TEAM 2021. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.
ROODUIJN, M. & PAUWELS, T. 2011. Measuring Populism: Comparing Two Methods of Content Analysis. West European Politics, 34, 1272-1283.
RSTUDIO TEAM 2021. RStudio: Integrated Development Environment for R. Boston: RStudio.
STANLEY, B. 2008. The thin ideology of populism. Journal of Political Ideologies, 13, 95-110.
TAGGART, P. A. 2000. Populism / Paul Taggart, Buckingham [u.a, Open Univ. Press.
TAIT, O. 2016. The North-South Divide: How Do We Measure Development Around the World? To What Extent is the Brandt Line Outdated? Available: https://intranet.kes.hants.sch.uk/resource.aspx?id=145038.
VERBEEK, B. & ZASLOVE, A. 2017. Populism and Foreign Policy. The Oxford Handbook of Populism, 384-405.
VIALA-GAUDEFROY, J. 2021. How Donald Trump’s populist narrative led directly to the assault on the US Capitol. The Conversation.
WEYLAND, K. 2001. Clarifying a Contested Concept: Populism in the Study of Latin American Politics. Comparative Politics, 34, 1-22.
WEYLAND, K. 2017. Populism. A political-strategic approach. The Oxford handbook of populism, 48-73.
WHITE, E. M. 1985. Teaching and Assessing Writing: Recent Advances in Understanding, Evaluating, and Improving Student Performance. The Jossey-Bass Higher Education Series, ERIC.
WICKHAM, H. 2016. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York.
Creative Commons Attribution 4.0 International License (CC-BY 4.0)
