Populist Attitudes and Conspiracy Beliefs: Exploring the Relation Between the Latent Structures of Populist Attitudes and Conspiracy Beliefs

Despite the alleged affinity between populism and conspiracy theories, how they relate on the individual level remains relatively unknown. This study explores the relation between populist attitudes and conspiracy beliefs at the individual level. First, I test whether the conspiracist facets, which directly involve governmental participation, are associated with the dimensions of populist attitudes. Further, I examine the relation of political trust with the dimensions and facets of both constructs as well as their predictive power of the self-reported propensity to vote for a populist party. To test these assumptions, a cross-sectional study was conducted in Germany. Confirmatory factor analyses indicate a strong association between conspiracist facets that directly involve governmental participation and the anti-elitism and sovereignty dimensions of populist attitudes. Findings further show that low political trust is related to all dimensions of populist attitudes–especially anti-elitism–and to the conspiracist facets. Furthermore, the sovereignty dimension of populist attitudes and low political trust predict the propensity to vote for the right-wing populist party AfD. These findings provide new insights to a more nuanced understanding of populism on the individual level and the relation to conspiracy beliefs. SOV and AE. The results revealed strong correlations between the three facets of conspiracy beliefs with governmental participation with the populist AE dimension. The highest r was for the information control facet ( r = .69, p < .001). The correlations of the three participatory facets with the populist sovereignty dimension were somewhat lower, with one exception. The sovereignty dimension was strongly correlated with the GM facet ( r = .61, p < .001). Contrary to the expectation, the three participatory facets were not strongly connected with the anti-pluralistic dimension of populism, namely higher beliefs in the homogeneity of the people. In line with the assumptions, correlations between the ET facet and the populist attitudes dimensions


Populism and Belief in Conspiracy Theories
Populism is widely defined as thin-centered ideology "that considers society to be ultimately separated into two homogenous 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, p. 543). Three elements characterize the ideational core of populism (Akkerman et al., 2014;Jagers & Walgrave, 2007;Schulz et al., 2018): first, a strong antagonism between the good people and the corrupt elite (Akkerman et al., 2014;Hawkins, 2012). This division is pri marily moral and is often referred to as Manichaean discourse, a 'good-versus-evil' understanding of politics (Hawkins, 2012;Oliver & Wood, 2014). Second, populism refers closely to the honest, hard-working people and emphasizes an unrestricted popular sovereignty (Schulz et al., 2018). Third, people are seen as being homogeneously virtuous (Schulz et al., 2018), whereas specific population segments are stigmatized as a threat or burden on society and are excluded (Jagers & Walgrave, 2007). This ideological core is independent of political ideology and can be observed in both rightand left-wing actors (Mudde & Kaltwasser, 2017;Van Hauwaert & van Kessel, 2018). To measure populism among individuals, the concept of populist attitudes used in this study builds on the multidimensional definition of populism. Populist attitudes are operationalized and measured as a latent construct made up of three dimensions: demand for popular sovereignty, anti-elitism, and belief in the homogeneous virtuousness of the people (e.g., Akkerman et al., 2014;Schulz et al., 2018).
Belief in conspiracy theories represents the degree to which individuals consider a conspiracy theory to accurately describe and explain a specific event (e.g., Abalakina-Paap et al., 1999;Douglas et al., 2017). A key finding of psycho logical research is that the endorsement of one conspiracy theory strongly intercorrelates with the endorsement of other theories, as well as with unrelated (Goertzel, 1994;Swami et al., 2010;Swami et al., 2011), hypothetical (Swami et al., 2011), and mutually contradictory conspiracy theories (Wood et al., 2012). Therefore, it has been previously argued that conspiracy theories form a monological belief system (Goertzel, 1994) and that people do not evaluate the evidence of a specific conspiracist claim in a rational manner. Rather individuals show relatively stable differences in conspiracist ideation which describes a general tendency to deal with conspiracy theories (Brotherton et al., 2013;Swami et al., 2011) and functions as a generalized political attitude distinct from other political attitudes, such as right-wing authoritarianism and populist attitudes (Imhoff & Bruder, 2014).
Conspiracy theories become especially prevalent in situations of crisis and change (van Prooijen & Douglas, 2017). Research investigating the psychological factors that drive belief in conspiracy theories differentiate three main motives  that provide an explanation for why conspiracy theories are particularly appealing in crisis situations and to those individuals who feel disconnected from authorities and people in power (Thórisdóttir et al., 2020;Uscinski & Parent, 2014). First, conspiracy theories provide simple explanations and help people to understand the world around them when facing uncertainty and contradiction (epistemic motives). Second, conspiracy theories can offer people a sense of security in threatening situations (existential motives), and third, they may help people to maintain a positive image of the (threatened) in-group (social motives) .
Much like conspiracy beliefs, populist attitudes are also linked to crisis situations and to the individuals who are on the losing side of current changes such as growing modernization and globalization (Mudde, 2010). Consequently, the question arises as to how populist attitudes and conspiracy beliefs are related. The relation can be drawn upon the three core ideological elements of populism (Castanho Silva et al., 2017). Populism and conspiracies share a similar worldview; one manifested in politics, the other in society (Castanho Silva et al., 2017). The (political) world is divided into well-defined sides on moral grounds (Manichean worldview) in which political or societal elites are accused of betraying the good and honest people. Populism accuses the (political) elite of being evil or corrupt, disinterested in the needs of the people, and working for their own gain, whereas conspiracy theories assume that conspirators with power and resources control the unknowing people or victims of the conspiracy (Castanho Silva et al., 2017;Imhoff & Bruder, 2014;Mudde, 2004). Moreover, both populism and conspiracy theories try to reduce the complexity of world events using rhetorical simplification and explaining them as an effect of the actions of powerful groups (Byford, 2014;Castanho Silva et al., 2017;Hawkins et al., 2018).
In line with empirical findings highlighting the relation between populist attitudes and conspiracy beliefs (Bergmann, 2018;Castanho Silva et al., 2017), I assume that citizens who have formed an attitude about one attitudinal concept will also be likely to have developed a corresponding opinion on the other. Yet, research has shown that both are generalized political attitudes and theoretically distinct concepts (Brotherton et al., 2013;Imhoff & Bruder, 2014). Based on these findings, I hypothesize that populist attitudes and conspiracy beliefs are two closely related but distinct concepts.
H1: Populist attitudes will show strong and positive correlations with conspiracy beliefs.

The Relation Between Populist Dimensions and Conspiracist Facets
Several studies present evidence for the internally coherent and unidimensional structure of conspiracy beliefs (Goertzel, 1994;Wood et al., 2012; for an overview, see Swami et al., 2017). Brotherton et al. (2013) suggest five different facets of conspiracy beliefs which are strongly intercorrelated, thus also supporting the unidimensional structure of conspiracy beliefs. At the same time, these facets reflect different assumptions underlying conspiracy beliefs. The five facets are: (1) government malfeasance (GM), which comprises allegations of the involvement of governments in routine criminal conspiracy; (2) malevolent global conspiracies (MG), which reflect the idea that global events are controlled and carried out by small, secret groups; (3) extraterrestrial cover-up (ET), which reflects beliefs in the existence and concealment of extraterrestrial life; (4) personal well-being (PW), which comprises suspicions or concerns that personal health is endangered by conspiracies using mind-control technology or the spreading of diseases; and (5) control of information (CI), which comprises conspiracies about deliberately withholding information from the public by organizations (Brotherton et al., 2013, p. 6).
The above facets not only allow to differentiate the most important aspects of conspiracy theories (Brotherton et al., 2013), but also to analyze the relation between populist attitudes and conspiracy beliefs in greater detail. Three facets include governmental participation and directly manifest the anti-elitist attitude: GM and MG both focus on criminal actions either by national or supranational governments. CI talks about the suppression of information by organizations which also includes governments. In contrast, the two other facets, ET and PW, do not include governmental participa tion and do not directly manifest the anti-elitist attitude (Castanho Silva et al., 2017). Consequently, I assume that they do not have the same appeal to populists and therefore hypothesize a stronger relation between governmental participatory facets compared to facets where this attitude is not inherent.
H2: All dimensions of populist attitudes will show strong and positive correlations with the three facets of conspiracy beliefs which directly involve governmental participation (GM, MG, CI) and weaker correlations with the two facets of conspiracy beliefs, ET and PW.

Political Trust
Political trust addresses the core elements that unite populist attitudes and conspiracy beliefs, namely the opposition towards the elite. Political trust measures the extent to which the outputs of political institutions relate to the individual expectations (Craig et al., 1990). Although political trust can also refer to a broad evaluation of policy outcomes and regime satisfaction, it is predominantly used with regard to the evaluation of the functioning of political institutions, such as political parties, governments, or parliaments (e.g., Thomassen et al., 2017).
As argued above, populist attitudes and conspiracy beliefs show overlaps regarding the belief that (political) elites are willing and able to act (conspiratorially) against the people. Populist attitudes and conspiracy beliefs can therefore be viewed as a function of increasing political distrust (Swami et al., 2010). Several empirical studies show that low political trust is related to both populist attitudes (Doyle, 2011;Geurkink et al., 2020) and belief in (specific) conspiracy theories (Miller et al., 2016;Swami et al., 2010), but also to more general conspiratorial predispositions (Imhoff & Bruder, 2014;Uscinski et al., 2016). Further, some experimental studies show that exposure to conspiracy theories can lower political trust (Einstein & Glick, 2015;Kim & Cao, 2016). Nevertheless, the causal direction of the relationship between political trust and populist attitudes and conspiracy beliefs is highly debated (see e.g., Geurkink et al., 2020;Thórisdóttir et al., 2020).
Here I investigate the association between political trust and subdimensions of populist attitudes and conspiracy beliefs. I assume political trust to be negatively associated with all dimensions of populist attitudes and facets of conspiracy beliefs and that political trust is more strongly associated with some dimensions and facets than others. Low levels of trust in political institutions taps into the anti-elitism dimension of populist attitudes "with anti-elitist sentiments defined as anti-politician or anti-institution" (Geurkink et al., 2020, p. 248). However, political trust does not directly capture the demand for sovereignty nor defines the people as homogeneously virtuous (Geurkink et al., 2020). Political trust taps into the GM facet of conspiracy beliefs which directly refers to conspiratorial acts by the government. Further, political trust addresses the MG and CI facet of conspiracy beliefs as they both also directly relate to governmental action. It does not address the ET and PW facet of conspiracy beliefs. Just because an individual has lower levels of trust, this does not entail a belief in the existence and the cover-up of extraterrestrials or that one's own well-being is threatened by mind-controlling technologies or secretly administered drugs.
H3: Political trust will show strong and negative correlation with all dimensions of populist atti tudes and the facets of conspiracy beliefs but vary in strength.

Voting Preference
The recent success of populist parties in different countries has led researchers to increasingly examine why voters support populist parties. Studies have found evidence for a strong link between populist attitudes and the preference to vote for a populist party (Bélanger & Aarts, 2006;Geurkink et al., 2020;Werts et al., 2013) for both left-and right-wing populist parties Van Hauwaert & van Kessel, 2018). To date, findings on the relation between belief in conspiracy theories, ideology, and voting preference are mixed (e.g., Thórisdóttir et al., 2020). While some studies link belief in conspiracy theories to right-wing political ideology (e.g., Galliford & Furnham, 2017), others show that this relation is present both on the right and on the left, especially on the political extremes (e.g., van Prooijen et al., 2015). A similar picture emerges with regard to electoral preferences: while some studies have found a relation between belief in conspiracy theories and voting for right-wing (populist) parties (e.g., Oliver & Rahn, 2016), findings from multi-party systems in Europe show that belief in conspiracy theories is associated with voting for populist parties at both ends of the ideological spectrum (see e.g., Thórisdóttir et al., 2020). This might be explained by the fact that leftand right-wing populist voters (similarly) oppose elites (Mudde & Kaltwasser, 2017;Thórisdóttir et al., 2020) although each might endorse different types of conspiracy theories (van Prooijen et al., 2015).
Germany presents a particularly interesting case to examine the extent to which populist attitudes and conspiracy beliefs are related to the preference to vote for a populist party. First, since 2017 the right-wing populist party AfD has had representation in the parliament. Second, the other side of the ideological spectrum is also represented in the parliament by the left-wing party Die Linke. Although the question of whether Die Linke should be considered a populist party is debated, the party still employs populist discourse (Hough & Keith, 2020;Rooduijn & Akkerman, 2017). To add to the question of how populist attitudes and conspiracy beliefs contribute to the voting preference for different populist parties and to analyze the relation in more detail, I investigate the predictive power of populist attitudes dimensions and conspiracist facets. In line with previous research, I expect the dimensions of populist attitudes and facets of conspiracy beliefs to predict the voting propensity for a populist party (Van Hauwaert & van Kessel, 2018). I hypothesize that: H4: Facets of conspiracy beliefs and dimensions of populist attitudes will positively predict the propensity to vote for a populist party.

Method Sample
The participants in this study were German citizens over 18 years recruited online between May and June 2018. Participants completed an online questionnaire and in return could take part in a lottery involving gift cards. A total of 295 individuals filled in the questionnaire. Prior to the analysis, data quality was thoroughly checked and ensured with regard to speeding and straight-lining, age, and incomplete questionnaires (see Table S1 of the Online Supplementary Materials). The final sample includes 243 cases (109 females, 131 males; no gender data for 3 participants) which is in accordance with the requirements to establish robust correlation coefficients as described by Schönbrodt and Perugini (2013). Age ranged from 19 to 80 years (M = 35.46, SD = 14.35). The descriptive statistics show an overrepresentation of higher levels of education with 9.1% of participants with a medium level of education, 87.2% with a high level of education. The descriptive information of all variables used in this study are presented in Table S2 of the Online Supplementary Materials.

Instruments
Likelihood to vote for a populist party was measured with the propensity to vote question to capture the probability to vote for a certain party (van der Eijk et al., 2006). Responses were measured on an 11-point scale ranging from (1) I would never to (11) I would most definitely vote for this party. Participants were asked to evaluate all parties represented in the German Bundestag (CDU/CSU, SPD, FDP, B90/Die Grünen, AfD, Die Linke).
For populist attitudes, I used the multidimensional 9-item scale by Schulz et al. (2018) consisting of three subscales which represent the three dimensions: (1) demand for popular sovereignty (SOV), (2) anti-elitism attitudes (AE), (3) belief in the homogenous virtuousness of the people (HOM) (for all item wordings see Table S3 of the Online Supplementary Materials). Participants' agreement with each item was measured using a 5-point Likert scale ranging from (1) completely disagree to (5)  To examine conspiracy beliefs the generic conspiracist beliefs (GCB) scale by Brotherton et al. (2013) was used. The 15-item scale measures conspiracy beliefs at a general level on five facets without reference to real-world conspiracy theories. Although the GCB scale is widely used to measure individual differences in conspiracy beliefs (Swami et al., 2017), so far it has never been used in a German speaking context. Therefore, the items were translated into German and subsequently discussed and revised to ensure an accurate translation (for all item wordings see Table S4 of  Political trust was measured as a latent variable with four indicators: confidence in the European Union, the national government, politicians and political parties. Participants rated their trust on a 5-point Likert scale from (1) not at all to (5) completely. Responses were aggregated to a political trust scale (M = 3.10, SD = .86, α = .86). Finally, participants were asked to report their age, gender and level of education.

Analysis
To investigate the hypotheses, I applied confirmatory factor analyses using R (version 3.6.3) with the lavaan package (Rosseel, 2012) to assess the strength of the relation between populist attitudes and conspiracy beliefs (H1) and to examine which populist attitudes dimensions and conspiracist facets are strongly associated (H2). I applied first order Pearson's correlations to examine how political trust is associated with the dimensions of populist attitudes and conspiracy conspiracist facets (H3) and multivariate regression analyses to explain voting for a populist party (H4).

Results
A first confirmatory factor analysis was conducted to check the measurement models and the strength of the correlation between populist attitudes and conspiracy beliefs. The model included both constructs modeled as latent variables, with the dimensions and facets as primary factors and the total scores for populist attitudes and conspiracy beliefs as secondary factors (see Figure 1). The model showed an adequate model fit (χ 2 = 476.59, df = 243, p < .001, CFI = .94, TLI = .93, RMSEA = .06, SRMR = .07, N = 243).
As hypothesized in H1 results showed a strong positive correlation (r = .77, p < .001) between populist attitudes and conspiracy beliefs. Individuals high in populist attitudes also showed higher levels of conspiracy beliefs. Due to the high theoretical overlap and the high correlation between the two constructs I addressed discriminant validity to examine whether the two constructs are distinct. The current data fulfills the required criteria of the Heterotrait-Monotrait Ratio to establish discriminant validity (Henseler et al., 2015) (see Table S5 and Table S6 of the Online Supplementary Materials for details).
Next, I investigated the correlation between the latent variables capturing populist attitudes dimensions and con spiracy facets. Results are displayed in the matrix in Table 1. Therefore, a second confirmatory factor analysis model was conducted including only the five conspiracist facets and the three dimensions of populist attitudes. The model showed an acceptable fit, χ 2 = 391.14, df = 224, p < .001, CFI = .96, TLI = .95, RMSEA = .06, SRMR = .05. N = 243. As hypothesized conspiracist facets which directly involve government participation should be strongly associated with all populist attitudes' dimensions. Following the theoretical assumptions, results showed moderate to high correlations between the facets GM, MG, and CI, with the two populist attitudes dimensions SOV and AE. The results revealed strong correlations between the three facets of conspiracy beliefs with governmental participation with the populist AE dimension. The highest r was for the information control facet (r = .69, p < .001). The correlations of the three participatory facets with the populist sovereignty dimension were somewhat lower, with one exception. The sovereignty dimension was strongly correlated with the GM facet (r = .61, p < .001). Contrary to the expectation, the three participatory facets were not strongly connected with the anti-pluralistic dimension of populism, namely higher beliefs in the homogeneity of the people. In line with the assumptions, correlations between the ET facet and the populist attitudes dimensions were clearly lower (.26 < r < .28, p s < .001). However, the results showed significant moderate to strong correlations between the PW facet and all dimensions of populist attitudes (.40 < r < .50, p s < .001). To sum up, these results partially support the theoretical assumption that facets of conspiracy beliefs which directly involve governmental participation are strongly connected to all dimensions of populist attitudes.
Next, I ran first-order Pearson correlations to analyze how political trust is related to the dimensions of populist attitudes and facets of conspiracy belief. Results are displayed in Table 2. Following the theoretical expectations, all populist attitudes dimensions and conspiracist facets are significantly negatively associated with political trust differ, however, with regard to the strength of the association. The sovereignty dimension of populist attitudes and the five facets of conspiracy showed moderate negative correlations with political trust (-.46 < r < -.59, p s < .001). The AE dimension and political trust were found to be strongly negatively correlated (r = -.63, p < .001) whereas the correlation with the HOM dimension was found to be weakly negative (r = -.24, p < .001).    In the last set of hypotheses, I expected facets of conspiracy beliefs and dimensions of populist attitudes to positively predict the voting propensity for a populist party. Multivariate regression analysis was used to test if the populist attitudes dimensions and the conspiracist facets significantly predicted the propensity to vote for the right-wing populist party AfD and for the left-wing party Die Linke. Results for the multivariate regression model are displayed in Table 3. The results indicated that dimensions of populist attitudes, facets of conspiracy beliefs and the control variables political trust, age, gender and education explained 34.8% of the variance for the propensity to vote for AfD and 11,2% for the propensity to vote for Die Linke. It was found that the populist dimension SOV significantly positively predicted voting propensity for AfD (β = .27, p < .001). Results further revealed that individuals scoring low on political trust (β = -.35, p < .001) and male individuals (β = .24, p < .001) showed a higher propensity to vote for the AfD. Contrary to the assumptions, the two populist dimensions AE and HOM did not significantly predict voting propensity for AfD. Further, none of the conspiracist facets predicted the likelihood of voting for the AfD when the populist attitudes dimensions are considered. When only conspiracist facets are considered, the GM facet significantly predicted voting propensity for AfD (β = .27, p < .01) (see Table S7 of the Online Supplementary Materials).
For the voting propensity for Die Linke it was found that neither the dimensions of populist attitudes nor the facets of conspiracy beliefs positively predict voting propensity. For the populist HOM dimension results revealed an effect contrary to the assumption. Individuals with lower values on the HOM dimension of populist attitudes (β = -.28, p < .001) showed a higher voting propensity for Die Linke, and so did female individuals (β = -.24, p < .001).

Discussion
The main objective of this study was to examine the relation between populism and conspiracy theories by systemati cally testing the relationships among the multiple dimensions of populist attitudes and facets of conspiracy beliefs. Although the literature mentions overlaps between populist attitudes and belief in conspiracy theories (Bergmann, 2018;Hawkins et al., 2018;Imhoff & Bruder, 2014;Mudde, 2004), research that systematically analyzes in which specific conceptual areas these two constructs overlap is scarce (for an exception, see Castanho Silva et al., 2017). The study used a survey to test the hypotheses on the relation between the latent structures of populist attitudes and conspiracy beliefs, the relation with political trust as well as their predictive power for the voting propensity for a populist party in Germany (AfD, Die Linke). In doing so, this paper extends recent research of correlates of populism which is essential to understand why populist parties are increasingly gaining approval in today's democracies. I summarize the main findings of this study and discuss them in relation to the research questions. What is the relation between populism and belief in conspiracy theories at the individual level? Populist attitudes and conspiracy beliefs are positively related, but they are distinct constructs. These results are in line with earlier findings (Castanho Silva et al., 2017) that people who have developed populist attitudes tend to give greater credence to conspiracy theories and vice versa, people who have more pronounced conspiracy beliefs tend to also show stronger populist attitudes. More specifically, the findings of this study provided general support for a close relation between the facets of conspiracy beliefs which directly involve governmental participation (GM, MG, CI) and two dimensions of populist attitudes, namely AE and SOV. This supports the theoretical assumption that populism and conspiracy theories both share a Manichean worldview, which is manifested in those dimensions encompassing a good vs. evil dichotomy, such as strong anti-elitism attitudes (Castanho Silva et al., 2017;Imhoff & Bruder, 2014;Mudde, 2004).
I want to highlight two further results. First, the strongest relation was found between the AE populist dimension and the information control conspiracy facet. This is consistent with prior findings on perceptions of the news media held by citizens with populist attitudes (Schulz et al., 2020). The anti-elitist attitudes towards politics are transferred to the media, which is similarly seen as part of the elite and involved in an elitist conspiracy (Fawzi, 2019). Second, contrary to the theoretical expectations, the findings also suggest a moderate relation between all populist attitudes' dimensions and the conspiracist facet on suspicions that personal welfare is endangered by conspiracies. It is plausible that the Manichean discourse is manifested not only on a political or societal level, but also on a personal one due to a perceived threat to one's own life.
How is political trust related to populist attitudes and conspiracy beliefs? This study displayed a relation between low political trust and both populist attitudes dimensions and conspiracist facets. This is in line with previous findings that showed negative relations between political trust and populist attitudes (Doyle, 2011;Geurkink et al., 2020) and conspiracy beliefs independently (Miller et al., 2016;Swami et al., 2010;Uscinski et al., 2016). Low trust in political institutions was particularly related to the AE dimension of populist attitudes. However, it appears that low political trust is similarly related to all facets of conspiracy beliefs, not just those involving direct governmental participation. Individuals who distrust political institutions also tend to believe in the existence and cover-up of extraterrestrial life and the endangerment of their personal health by mind-controlling technologies or deliberately spread diseases. Although, no conclusions could be drawn about the directional relation between populist attitudes, conspiracy beliefs, and political trust, these results demonstrated the potential negative societal effects in a world where conspiracy theories and populism are gaining popularity and political trust is declining.
Do conspiracist facets and dimensions of populist attitudes predict the propensity to vote for a populist party? Previous research has linked political attitudes and the belief in conspiracies to populist voting (e.g., Akkerman et al., 2014;Van Hauwaert & van Kessel, 2018;van Prooijen et al., 2015). This study sought to provide a more nuanced understanding of what specific dimensions and facets predict voting for a populist party. Despite highly correlated, populist attitudes and conspiracy facets predict differentially propensity to vote. The findings indeed indicated that male individuals with higher demands for popular sovereignty and lower political trust had a higher propensity to vote for the right-wing populist party AfD in Germany. These findings support previous research by showing that stronger populist attitudes lead to a higher propensity to vote for a right-wing populist party (e.g. Van Hauwaert & van Kessel, 2018). The findings also highlight the link between the decline in trust in politicians and political institutions and the propensity to vote for AfD.
Notably, for support for Die Linke, there was a negative relation with the belief in the homogeneous virtue of the people. Contrary to the theoretical assumption, support for the left-wing party was not based on a notion of shared values and interests of 'the people'. Rather, a pluralistic perception of the society may be what drives the support for the left-wing party, as suggested by the finding that the HOM dimension of populist attitudes was negatively related to the propensity to vote for Die Linke. Further, these results contribute to the discussion of whether and to what extent the party Die Linke can be classified as a populist party (Hough & Keith, 2020).
Surprisingly, conspiracist facets did not affect voting intention for the right-wing populist party AfD and the left-wing party Die Linke. This seems to contradict previous research which has found that conspiracy beliefs predicted voting intentions for populist parties across Europe (see Thórisdóttir et al., 2020). However, unlike in this study, populist attitudes were not considered. In the present study, the conspiracist GM facet positively also predicted the voting intention for AfD prior to adding populist attitudes to the model, but the results no longer held once populist attitudes were simultaneously considered. Further, the investigation of the populist dimensions showed that especially demand for popular sovereignty explained the voting intention for AfD. The rejection of the elite, for which the strongest overlap between populist attitudes and conspiracy beliefs was found, seemed to be less important here.
This study however is not without limitations. The first one is the small sample size. Determining the required sample size for confirmatory factor analyses is challenging (Wolf et al., 2013) and the recommendation is to apply more stringent criteria regarding the ratio of observations to estimated parameters (Schreiber et al., 2006). However, the study fulfills the rule-of-thumb of a minimum sample size of 200 (Boomsma, 1985). Future studies should investigate the relations using a larger sample. Since the sample is not representative of the German population, any inferences about average levels of attitudes should be made with caution. A second limitation is the correlational nature of our data which does not allow us to draw conclusions about the direction of the relationship in terms of causality. Although negligible for this study-as I was interested the latent relations of populist attitudes and conspiracy beliefs-future studies should analyze the direction of the relation using longitudinal or experimental designs.
Despite the above-mentioned limitations, there is value in this study's findings as this is one of the first studies to formally establish the relationship between populist attitudes and conspiracy beliefs in a multi-party political system. AE and SOV dimensions of populist attitudes are strongly associated with the three facets of conspiracy beliefs that include direct governmental participation, whereas the associations with the two facets ET and PW are less strong. However, the association of the third dimension of populist attitudes, belief in the homogeneous virtuousness of the people, with all facets of conspiracy beliefs is lower. While low political trust is especially associated with the AE dimension of populist attitudes, political trust is similarly strongly associated with all facets of conspiracy beliefs. This is advantageous as the alleged affinity for conspiracy theories among populist sympathizers until now had been mostly assumed rather than tested. Thereby this study contributes to a better understanding on how populism and conspiracy beliefs are related on the individual level.
Funding: This research was supported by a research grant from the research focus "Communication, Media, and Politics" which is part of the Forschungsinitiative Rheinland-Pfalz, Germany.
Acknowledgments: I gratefully thank Michaela Maier, Tobias Rothmund and Teresa Gil-Lopez for their useful comments that really helped to improve the manuscript. Further, I would like to thank the editor and the anonymous reviewers for their excellent reviews and suggestions.

Competing Interests:
The author has declared that no competing interests exist.

Supplementary Materials
The Supplementary Materials contain the exclusion criteria (Table S1), descriptive statistics of the sample (Table S2), the item wording for the populist attitudes scale (Table S3) and the general conspiracist belief scale (Table S4), the discriminant validity criteria (Table  S5 and S6), and the summary of the hierarchical regression analysis for voting propensity for AfD and Die Linke (Table S7). (For access see Index of Supplementary Materials below).