The Personality Origins of Positive and Negative Partisanship

Negative partisanship describes the intense disdain for a rival political party. A growing number of political scientists in the US and beyond examine the impact of negative partisanship on citizens’ political behavior, asserting the notion that negative partisanship exerts a strong influence, either on its own or in combination with positive partisanship. Yet we know little about the psychological origins of negative and positive partisanship: Which personality traits are associated with high levels of negative partisanship, and do they differ from the ones that have been linked to positive partisanship? In this article, I address these questions. Utilizing a sample of US adults and a sample of Swedish adults, I examine the influ‐ ence of prominent personality traits—including Authoritarianism, Social Dominance Orientation, the Need for Closure, and the Big Five—on strong negative and positive partisanship. I demonstrate that the personality origins of positive and neg‐ ative partisanship differ not just across the two samples but also across partisans on the left and on the right. I conclude the article with implications for research on polarization and a plea for more comparative work on (positive and nega‐ tive) partisanship.


Introduction
Negative partisanship (NPID) has received increasingly more attention from scholars in the past few years.While a simple Google Scholar search yields only 1,600 hits for the term "negative partisanship," this number cannot convey the exponential growth of scholarship over the past few years.Indeed, 83% of current research on NPID has been contributed since 2018.Thus, NPID is gaining traction among political scientists who study its effects on political behavior, either in combination with positive partisanship (PPID; Abramowitz & Webster, 2016;Bankert, 2021) or, increasingly, on its own (Caruana et al., 2015;Garzia & Ferreira da Silva, 2022;Mayer, 2017;Medeiros & Noël, 2014;Meléndez & Kaltwasser, 2021;Rose & Mishler, 1998), demonstrating that both PPID and NPID shape vote choice, turnout, and various other forms of political participation (Samuels & Zucco, 2018;Tsatsanis et al., 2020).Despite their demonstrated impact on political behavior, we know little about the origins of PPID and NPID.While prior work examines the role of party leaders (e.g., elite-level polarization) as well as institutional features (e.g., the two-party system), I focus on the individual, psychological origins of NPID and PPID that address several important questions: Who is more likely to develop strong NPID?What kind of personality traits are associated with high levels of NPID, and do they differ from the ones that have been linked to PPID?
To tackle these questions, I examine a range of prominent personality traits, including the Need for Closure (NfC), Authoritarianism, Social Dominance Orientation (SDO), and the Big Five.I focus on these traits because of prior work that has provided evidence for their association with stronger partisan attachments.Yet there is little systematic evidence that examines more than just one personality trait at a time or that compares their impact on both PPID and NPID.Moreover, there is little prior comparative work on the relationship between personality and partisanship despite the concern that personality traits and their impact might vary across countries due to cultural as well as political differences.To address this gap, I examine a sample of American and Swedish citizens in this study.The US two-party system differs dramatically from the Swedish coalitional multi-party system.Indeed, Sweden is not just characterized by proportional representation but also by a "fundamentally consensual political culture" and much lower levels of affective polarization across partisans (Oscarsson et al., 2021, p. 5).These features stand in sharp contrast to the American two-party system, thereby allowing for a comparison of the personality origins of positive and negative partisan identity in two vastly different political systems.
Last, in contrast to prior work, I utilize a measure that conceptualizes NPID as a social identity rather than just a negative affect towards the out-party (see Abramowitz & Webster, 2018) or a negative vote (see Caruana et al., 2015).While I do not claim this measure to be better in any way than prior measurement strategies, it does feature a few promising measurement properties, such as a multi-item index that can gauge even subtle variations in negative partisan identity strength, high reliability, as well as good predictive power (see Bankert, 2021)-similar to the positive partisan identity scale (see Bankert et al., 2017).
Using these identity scales, I demonstrate that PPID and NPID have very different psychological origins among partisans in both Sweden as well as the US.I also find significant differences across the ideological aisle whereby NPID and PPID on the left are associated with different traits than NPID and PPID on the right.From this perspective, this article contributes a few insights to the contemporary literature on PPID and NPID: First, it provides a systematic and comprehensive overview of the effects of personality traits on strong PPID and NPID.Second, it compares these effects across two vastly different political systems, thereby providing insights into their generalizability.Third, it compares these effects across the ideological left and the right, which elucidates the different psychological compositions of partisan groups.
In the remainder of this article, I first provide a brief overview of the existing literature on personality and partisanship.I then introduce the data and the measurement strategies, including the decision to measure PPID and NPID as an identity.In the analysis part of the article, I examine four different types of partisans: Negative Partisans, who display high levels of NPID but low levels of PPID; Positive Partisans, who display high levels of PPID but low levels of NPID; Closed Partisans, who display both high levels of NPID and PPID; and last, Apathetic Partisans, who display both low levels of NPID and PPID.This comparison clearly identifies the different origins of PPID and NPID as well as the traits that contribute to their alignment.I conclude the article with implications for research on polarization and a plea for more comparative work on (positive and negative) partisanship.

Partisanship as a Social Identity
In this study, I consider partisanship a "social identity," which is defined as "that part of an individual's selfconcept which derives from his knowledge of his membership in a social group (or groups) together with the value and emotional significance attached to that membership" (Tajfel, 1978).I measure this identity-based conceptualization of partisanship (Green et al., 2004;Greene, 1999Greene, , 2002Greene, , 2004;;Huddy et al., 2015) with a multi-item index which is broadly based on Mael and Tetrick's (1992) "identification with a psychological group scale."With items such as "When I meet somebody who supports this party, I feel connected" and "When people praise this party, it makes me feel good," the scale captures crucial social identity ingredients such as partisans' subjective sense of belonging to the group as well as the importance of the group membership.Since the scale gauges affirmative identification with a political party, I refer to it as the positive partisan identity (PPID) scale.
This social identity framework is also useful for deriving an understanding of NPID.According to social identity theory (Tajfel & Turner, 1979), identities can also form in opposition to groups to which we do not belong.Thus, the identity is negative in the sense that it centers on the rejection of an out-group and its members (Zhong, Galinsky, & Unzueta, 2008;Zhong, Phillips, et al., 2008).In the political arena, Americans form negative identities in response to third parties (Bosson et al., 2006) as well as political organizations like the National Rifle Association (Elsbach & Bhattacharya, 2001), turning the exclusion from a group-the "not being one of them"into a meaningful social identity.I argue that the same can be true for political parties, whereby the strong rejection of a political party can develop into a negative partisan identity (NPID).In my prior work, I designed and validated a multi-item scale that measures this NPID (see Bankert, 2021).To make the PPID and NPID scales as comparable as possible, I flip the items of the PPID scale to capture the emotional significance respondents associate with their rejection of the out-party with items such as "When I meet somebody who supports this party, I feel disconnected" and "I get angry when people praise this party."The PPID and NPID scale items are listed in Table A1 in the Supplementary File.

The (Un-) Alignment of Positive and Negative Partisanship
With two separate measures to capture PPID and NPID, it is possible to create a typology of partisans that can be distinguished by their different PPID and NPID levels.Early work by Rose and Mishler (1998) has already done so with the example of post-communist countries, whereby the authors examine four different types of partisans: (a) Open Partisans with PPID toward their in-party and without NPID toward another party, (b) Negative Partisans with NPID and without PPID, (c) Closed Partisans with both NPID and PPID, and (d) Apathetic Partisans with no identification.Rose and Mishler find that in the four countries they studied, namely Hungary, Poland, Romania, and Slovenia, more than half of respondents held NPID towards at least one party but PPID towards none.Similarly, Mudde and Kaltwasser (2018) note that in many Western European democracies, populist parties are the targets of NPID despite declining levels of PPID.These findings show that NPID and PPID do not always occur together.
While I partly rely on Rose and Mishler's terminology in this study, I slightly alter their typology.Rather than examining whether a partisan has a positive and/or negative identification with a political party, I examine the intensity or strength of that positive and/or negative identification.This leads to four different types of partisans: Positive Partisans with high levels of PPID and low levels of NPID, Negative Partisans with high levels of NPID and low levels of PPID, Closed Partisans with high levels of both PPID and NPID, and Apathetic Partisans with low levels of both PPID and NPID (see Figure 1).In the next section, I will utilize this typology to make predictions about the distinct personality traits that are associated with each type of partisan.

Personality and Partisanship
Researchers have long been interested in the personality origins of political attitudes and behavior (Adorno et al., 1950;Eysenck, 1954;McClosky, 1958).Throughout this article, I define personality traits as "relatively enduring patterns of thoughts, feelings, and behaviors that distinguish individuals from one another" (Roberts & Mroczek, 2008, p. 31) and that are exogenous to their political socialization (McCourt et al., 1999).From this perspective, the focus on personality traits as determinants of partisan attachments offers two distinct advantages: First, despite some developmental changes in dispositional traits during early adulthood, personality traits are relatively stable, which allows for a more generalizable interpretation of their effects on partisanship throughout an individual's life cycle.Second, personality traits temporally precede the development of many political values, attitudes, and behavior, including party attachments.Thus, despite the observational nature of the following analyses, personality traits intuitively are more likely to be a determinant of partisanship rather than vice versa (see Luttig, 2021, for an exception).
Within the large and diverse share of scholarship on the relationship between personality and politics, there are a few select and distinct traits that are featured quite prominently.These traits include Authoritarianism, SDO, the NfC, as well as the Big Five.While most of the prior literature tends to focus on either one or two of these traits at a time, this article examines the impact of all four personality concepts, thereby offering a systematic and comprehensive overview of the relationship between personality and partisanship.Prior scholarship has also focused much more extensively on the personality origins of PPID (Cooper et al., 2013;Gerber et al., 2012;Schoen & Schumann, 2007) and NPID (Webster, 2018).From this perspective, my predictions for Positive and Negative Partisans are most firmly grounded in prior scholarship, while the determinants of Apathetic and Closed Partisans constitute mostly uncharted territory.I thus remain agnostic regarding their personality associations.Yet, their exploratory nature also offers new avenues for future research on the origins of PPID and NPID and their varying intensity.In the following, I will briefly elaborate on each trait and articulate my expectations for their effect on PPID and NPID.

Need for Closure
The NfC is a psychological predisposition that has been used extensively in psychology to describe individuals with a "desire for a firm answer to a question, any firm answer as compared to confusion and/or ambiguity" (Kruglanski, 2004, p. 6).From this perspective, people with high levels of NfC tend to prefer firm and unequivocal assessments of the world and avoid ambiguity and nuance that could negate their need for order and structure (Kruglanski & Webster, 1996).Accordingly, NfC has been associated with heightened in-party favoritism and out-party hostility as well as partisan identity strength (Luttig, 2018).As Luttig (2018, p. 240) explains: Group identification, ingroup bias, and outgroup prejudice are motivated partly by the need for certainty and closure because groups provide members with a social identity and prescribe beliefs about who one is and what they should believe and think.Furthermore…uncertainty as a motivation for group membership can foster extremism, as extreme groups are more distinct and unambiguous.
From this vantage point, NfC might strongly predict high levels of PPID and NPID since they facilitate the rigid categorization of political parties into "good" and "bad," "us" versus "them" (H1a).At the same time, NfC has been linked to political conservatism as well as more right-wing political party preferences (Kossowska & Hiel, 2003), which leads to the expectation that NfC is more strongly related to PPID on the ideological right (H1b).

Authoritarianism
Authoritarianism reflects a general preference for social conformity over individual autonomy (Feldman, 2003;Feldman & Stenner, 1997;Stenner, 2005), driven by a strong dispositional need for order, certainty, and security as well as a general commitment to conventions and norms (Hetherington & Weiler, 2009;Jost et al., 2003).I thus expect (H2a) that Authoritarianism is positively linked to strong positive partisan identities since they provide a sense of belonging, group norms to comply with, as well as a simplified understanding of who is a friend or foe in a complex political world (see also Luttig, 2017).Negative partisan identities, on the other hand, do not satisfy the need for inclusion as easily as positive partisan identities do (for a similar argument, see Zhong, Galinsky, & Unzueta, 2008;Zhong, Phillips, et al., 2008).Instead, NPID turns the exclusion from a group-the "not being one of them"-into a meaningful social identity while it provides little affirmational guidance on who we are.Put differently, NPID leaves more uncertainty and imposes less cognitive order than PPID, which is why I expect Authoritarianism to be negatively related to strong NPID (H2b).Like NfC, Authoritarianism is strongly related to ideological conservatism (Federico & Reifen Tagar, 2014) and right-wing policy preferences (Hetherington & Suhay, 2011).Thus, Authoritarianism might be more predictive of PPID on the ideological Right (H2c).
While I treat Authoritarianism as a determinant of PPID and NPID, some prior work has challenged this causal order.In the example of the US, Luttig (2021, p. 786) notes that: As the GOP became more conservative on social issues, embraced the religious right, advocated being tough on crime…they communicated that their party sees the world as a dangerous place and that they value obedience, respect, good manners, and good behavior.Inferring the associations of the parties with these values, people change either their psychological worldview or the way that they answer survey questions about these topics to reduce cognitive dissonance.
While it is unclear to what extent this nuance also applies to other personality traits, I acknowledge it and avoid any strict causal claims in the results section.

Social Dominance Orientation
SDO is another individual-difference variable that is particularly relevant in the study of prejudice.People on the low end of SDO tend to endorse group equality and oppose societal hierarchies, while people on the high end seek power and high status for their group as well as dominance over others (Pratto et al., 1994).SDO thus draws people towards political parties and policies that rationalize and bolster group-based inequalities (Duckitt & Sibley, 2009).Consistent with these expectations, prior research has shown SDO to be strongly related to conservative or right-wing policies and party preferences in the US as well as beyond (e.g., Van Assche et al., 2019).I thus expect that SDO is more predictive of strong PPID on the right both in the US as well as in Sweden (H3a).At the same time, I expect to find a strong connection between SDO and NPID towards the ideological left since many left-wing policies aim to eradicate intergroup inequalities (e.g., affirmative action, access to social services, and universal healthcare) and promote awareness of systemic discrimination and privilege (H3b).It is uncertain how SDO relates to PPID and NPID overall.It is possible that effects cancel each other out once partisans on the left and on the right are jointly examined.I thus remain agnostic regarding their connection to SDO.

Big Five
The Big Five traits are a well-known and established framework for studying personality which specifies a small set of core traits, including Extraversion, Conscientiousness, Agreeableness, Emotional Stability, and Openness to Experience (see McCrae & Costa, 2008).Prior scholarship has demonstrated the impact of these traits on party preferences on the ideological left and right, albeit with somewhat mixed results.The most consistent finding is the relationship between Openness to Experience and liberalism on the one hand and between Conscientiousness and conservatism on the other (e.g., Alford & Hibbing, 2007;Mondak, 2010;Mondak & Halperin, 2008).There is also some evidence that Emotional Stability is linked to support for conservative candidates and parties and that Agreeableness is connected to support for liberal candidates and parties (Barbaranelli et al., 2007;Caprara et al., 1999;Mondak, 2010).This abundance of prior scholarship provides the foundation for a few concrete hypotheses regarding the relationship between the Big Five and partisan identities on the Left and on the Right.First, and in alignment with prior results, Openness to Experience should be related to a strong PPID on the left (H4a), while Conscientiousness should be related to a strong PPID on the right (H4b).Since Extraversion is connected to social and outgoing behavior, I also expect this trait to predict strong party attachments-regardless of their ideological direction (H4c).Since the evidence on Agreeableness and Emotional Stability is much more mixed, I remain agnostic about their impact.
Prior scholarship is less plentiful regarding the psychological origins of NPID though there is some evidence that Extraversion, Agreeableness, and Emotional Stability are negatively related to strong NPID (Abramowitz & Webster, 2018).Indeed, Extraversion and Agreeableness describe a person who is willing to hear the other side in a polite and trusting manner, while Emotional Stability reduces the chance of experiencing strong negative emotions such as anger and disdain in the first place.Webster (2018) further distinguishes between being a Negative Partisan and the intensity of NPID.The author shows that higher levels of Extraversion are associated with a lower probability of being a Negative Partisan (Webster, 2018).This finding has high face validity since, as Webster (2018) notes, extraverted individuals are more likely to be exposed to a vast array of different political viewpoints.This diverse exposure moderates their negativity towards the out-party and its members.Webster (2018) also demonstrates that higher levels of Agreeableness lessen the degree to which an individual exhibits negative affect toward the out-party and its members since the trait is associated with friendliness, fairness, and decency-even towards the out-party.From this perspective, these three traits-Extraversion, Agreeableness, and Emotional Stability-should be negatively related to strong NPID (H4d).While Abramowitz and Webster (2018) and Webster (2018) use feeling thermometer scales to gauge NPID, I utilize a multi-item scale that measures NPID explicitly as an identity, which I will elaborate on in the next section.

US Sample
For the analysis of partisanship in the US, I utilize original survey data that was collected in 2022 by a survey firm called Bovitz Inc., which provides an online panel of approximately one million respondents who participate in multiple surveys over time and receive compensation for their participation.While the sample is not nationally representative, it does reflect the US population on key demographics (see Table A2 in the Supplementary File).The sample includes 1,007 respondents, 882 of them completed the PPID scale, while 876 of them completed the NPID scale.Respondents who identified as a Democrat (or Republican) received the PPID for the Democratic (or Republican) Party and the NPID scale for the Republican (or Democratic) Party.The sample included 456 Republicans and 447 Democrats.

Sweden Sample
For the analysis of partisanship in Sweden, I utilize original survey data that was collected in 2021 by Bovitz Inc.While the sample is not nationally representative either, it does reflect key demographics of the Swedish population (see Table A3 in the Supplementary File).The sample includes 1,208 Swedish respondents, 968 of them completed the PPID scale, while 975 completed the NPID scale.Respondents received the PPID scale if there was a party that they considered "best" or if they indicated feeling closer to a particular party.Most commonly, that applied to the Social Democrats (28%), the Sweden Democrats (29%), and the Moderate Party (13%).NPID scale was administered based on the question of whether there is a political party that the respondent would never vote for.If so, this party was the target of the NPID scale, which most frequently applied to the Left Party (N = 265), the Green Party (N = 298), the Sweden Democrats (N = 447), and the Feminist Initiative (N = 334).If there were multiple political parties that respondents would never vote for, as is possible in multi-party systems, then respondents were asked to rate these parties on a feeling thermometer scale from 0 to 100.The party with the lowest rating was then selected as the target for the NPID scale.

Positive Partisanship
I conceptualize and measure PPID as a social identity (see Bankert et al., 2017).The importance of the in-party to an individual's self-concept, as well as the emotional significance of the membership in that party, is reflected in items such as "When people criticize this party, it feels like a personal insult" and "When I speak about this party, I usually say 'we' instead of 'they.' " Combining these eight items into one scale yields a continuum that can account for fine gradations in partisan identity strengthan advantageous feature since multi-item partisan identity scales have proven to be more effective than the traditional single-item in predicting political outcomes such as vote choice and political participation in the US and in European (Huddy et al., 2015) multi-party systems (Bankert et al., 2017).

Negative Partisanship
Social identity theory has also been insightful for the development of a negative partisan identity and its measurement.According to social identity theory, identities cannot only form as a function of common characteristics among in-group members but also in opposition to groups to which we do not belong.Thus, the identity is negative in the sense that it centers on the rejection of the out-group's characteristic.The negative partisan identity scale closely resembles the positive partisan identity scale and captures the emotional significance respondents associate with their rejection of the outparty with items such as "When I meet somebody who supports this party, I feel disconnected" and "I get angry when people praise this party" (see Bankert, 2021).
Descriptive statistics for all key variables and their measurements can be found in Tables A4, A5, and A6 of the Supplementary File, while the distribution of the PPID and NPID scale items can be found in Tables A7 to A10 in the Supplementary File.Pairwise correlations of all key variables are included in Tables A11 and A12.

Analyses
In the following analyses, I aim to investigate the personality determinants of PPID and NPID.The correlations between the two in the US are much higher than in Sweden (0.65 versus 0.36), indicating their overlapping nature in the two-party system.From this perspective, simply regressing the personality predictors onto the PPID and NPID values would make it challenging to disentangle the distinct psychological origins of these two types of partisanship.I thus create four different types of partisans based on their values on the NPID and PPID scales.For analytical purposes, "low" is defined as below the sample's mean value on the PPID/NPID scale, while "high" is defined as above the sample's mean value.This strategy preserves sample size while also providing a clear cut-off point.
The percentage shares for each type of partisan are included in Table 1.Both in the US as well as in Sweden, the overwhelming share of partisans fall into the categories of Closed Partisans and Apathetic Partisans; 42% of all American and 43% of Swedish partisans in the sample score highly on both the PPID and NPID scale, while 37% and 24% of American and Swedish partisans respectively are characterized by low scores on both the PPID and NPID scale.Only 10% of American partisans and 16% of Swedish partisans score highly on the PPID scale in conjunction with low values on the NPID scale.Similarly, 11% of American partisans and 17% of Swedish partisans fall on the high end of the NPID scale while also scoring low on the PPID scale.These comparisons reveal an interesting asymmetry: While NPID and PPID can certainly occur independently, the two types of partisanship much more commonly tend to occur together.
In the next part of the analysis, I examine whether these partisan types are related to distinct personality traits (see Figure 2).For this purpose, I regress each dichotomous partisan type onto the personality traits as well as a set of standard control variables (see Table A6 of the Supplementary File).Starting with Positive Partisans in the US, SDO, Agreeableness, and Openness to Experience emerge as strong and positive predictors.The effects of SDO and Openness are quite sizable.Across the range of SDO, PPID increases from 0.04 to 0.17 while keeping all other variables at their mean.There is a similarly steep increase in PPID from 0.02 to 0.1 as Openness increases from 0 to 1.At the same time, NfC is negatively related to being a positive partisan.As NfC increases, the probability of being a positive partisan significantly decreases from 0.12 to 0.03.
Moving on to Negative Partisans, NfC and Emotional Stability emerge as positive predictors of being a Negative Partisan, with similar increases in its predicted probability from 0.05 to 0.17 across the range of these two personality traits.Conscientiousness and Extraversion are uniquely and negatively related to NPID.In combination, these findings suggest that PPID and NPID do have distinct personality origins, in support of the notion that these two are independent constructs.Notes: Coefficients were estimated using a logistic regression model; all variables are scaled to range from 0 to 1; the corresponding table can be found in Table A13 in the Supplementary File; *** p < 0.001, ** p < 0.01, * p < 0.05, + p < 0.1.
Indeed, only NfC appears as a significant predictor in both analyses of Positive and Negative Partisans but with oppositional effects.Closed Partisans are characterized by high levels of SDO as well as Extraversion.The coefficient for NfC is positive and quite substantial.The predicted probability of being a Closed Partisan increases from 0.26 to 0.67 along the range of SDO-an effect that is similar to Extraversion, which is associated with a growth from 0.28 to 0.51 while holding all other personality variables constant.Last, Apathetic Partisans are characterized by higher levels of Conscientiousness and lower levels of NfC, SDO, and Extraversion.The effects are particularly strong for SDO.Across its range, the predicted probability of being an Apathetic Partisan shrinks from 0.56 to 0.12.Taken together, these analyses suggest that all four types of partisanship have distinct personality profiles.Yet three personality traits-NfC, SDO, and Extraversion-emerge frequently as significant predictors.When adjusted for multiple comparisons using Bonferroni correction, several relationships persist, such as the positive relationship between Negative Partisans and Extraversion, Closed Partisans and SDO and Extraversion, as well as Apathetic Partisans and SDO (see Tables A19 to A22 in the Supplementary File).
These results can only speak to American partisans, which limits their generalizability given the idiosyncratic nature of the US political system.Thus, I replicate the preceding analyses with a sample of Swedish partisans, which illuminates the nature of partisanship in multi-party systems (see Figure 3).For the prediction of Positive Partisans, none of the included personality variables appear to exert an impact which is an interesting departure from the US model.Moving on to Negative Partisans, only Authoritarianism emerges as a negative and significant predictor-which, once again, stands in sharp contrast to the results from the US sample.Indeed, as Authoritarianism increases from 0 to 1, the probability of being a Negative Partisan in Sweden decreases from 0.22 to 0.13.Among Closed Partisans, SDO (like in the US) and Authoritarianism exert significant effects.As these two traits increase from 0 to 1, Closed Partisanship's likelihood grows from 0.28 to 0.51.Remarkably, these two traits are negative predictors of being an Apathetic Partisan, with a decline in its predicted probability from 0.31 to 0.17 and 0.35 to 0.14 across the range of Authoritarianism and SDO, respectively.Interestingly, NfC and Conscientiousness are positively associated with being an Apathetic Partisan.The positive effects of Conscientiousness, as well as the negative effects of SDO, also surfaced among Apathetic Partisans in the US.These results are robust to alternative model specifications such as a multinomial logistic regression (see Tables A17 and A18 in the Supplementary File).When using Bonferroni-adjusted p-values, the relationship between Closed Partisans and Authoritarianism remains as well as the effect of SDO and Conscientiousness on Apathetic Partisans (see Tables A22 to A26 in the Supplementary File).Overall, these results provide two novel insights: First, the four types of partisanship are related to distinct personality profiles both in the US and in Sweden.Second, the four types of partisanship in the US and in Sweden are related to different personality profiles.This variation might speak to the role of institutional features such as the number of political parties (two-party versus multi-party system), the electoral rules (proportional versus majoritarian), the ideological space of the political system, as well as the country's political culture.

Ideological Differences Among Positive and Negative Partisans
The preceding analyses revealed distinct personality profiles for each of the four partisan types.Yet it is possible that there are personality differences between Positive and Negative Partisans on the left and right of the ideological spectrum.
To assess this possibility, I first examine the strength of PPID on the right in combination with low levels of NPID towards the left.In the US sample, this involves respondents who identify with the Republican Party but display low levels of NPID towards the Democratic Party.In the Swedish sample, this includes respondents who feel closer to the Moderate Party, the Sweden Democrats, or the Christian Democrats with weak NPID towards the left.Starting with the US (Figure 4), strong positive Republican partisanship is positively related to multiple personality traits, including SDO, Agreeableness, and Extraversion.The significant effects for SDO and Agreeableness remain even when using Bonferroni-adjusted p-values (see Table A27 in the Supplementary File).In contrast, Positive Partisans on the right in Sweden (Figure 5) feature lower levels of Agreeableness while also, similarly to Republicans in the US, scoring more highly on Extraversion.
For PPID on the left, I examine Democrats in the US.In Sweden, I include respondents who feel closer to the Left Party, the Green Party, the Feminist Initiative, or the Social Democrats.In both cases (see Figures 4 and 5), strong PPID on the left is not related to any personality traits.Only religiosity is a positive determinant in both countries, which is a noteworthy similarity.
Last, I replicate the same analyses for Negative Partisans who disdain certain political parties on the left or the right while being only weakly attached to a political party.In the US sample, this approach includes respondents with NPID towards the Democratic (left) and Republican Party (right), respectively.In Sweden, as exemplars of NPID towards the left, I include respondents who would never vote for the Left Party, the Green Party, the Feminist Initiative, or the Social Democrats.For NPID towards the right, I examine respondents who report never voting for the Moderate Party, the Sweden Democrats, or the Christian Democrats.
In the US (Figure 4), NPID towards the Republican Party is positively related to NfC and Emotional Stability but negatively related to Authoritarianism.In contrast, NPID towards the right in Sweden (Figure 5) is associated  In Sweden (Figure 5), NPID towards the left is characterized by high levels of SDO (like in the US).The effects are particularly remarkable for SDO, which is associated with an increase in NPID towards the left from 0.34 to 0.70.This effect also remains when Bonferroniadjusted p-values are used (see Table A34 in the Supplementary File).
Overall, these analyses reveal a few interesting patterns: NfC is positively related to NPID in the US but not in Sweden (H1).Authoritarianism is negatively related to NPID towards the right in the US but not in Sweden (H2).Among Negative Partisans towards a left-wing party, SDO emerges as a common predictor in both the US as well as in Sweden; consistent with H3, a preference for intergroup hierarchies is associated with higher levels of NPID towards the left.Extraversion was positively related to strong PPID on the right in both the US as well as Sweden (partially in alignment with H4c), while Agreeableness was a negative determinant in Sweden but a positive one in the US.This disparity might be reflective of the different political cultures in these two countries, whereby conservatism is much more the leitkultur in the US than in Sweden.Last, in both countries, PPID on the left was not associated with any of the included personality traits.

Conclusion
This article has examined the personality profiles of four distinct partisan types in both the US and Sweden-two vastly different political systems and cultures.The analyses revealed only a few similarities, such as the role of Extraversion among Positive Partisans on the right as well as the relationship between SDO and NPID towards the left.Overall, however, personality profiles differ across partisan types and across countries.These dissimilarities provide two important insights: First, PPID and NPID are two separate entities that can operate together but also independently of each other.Second, personality predispositions naturally interact with their environment, which might explain the inter-country differences in partisan-personality associations.Indeed, there are important nuances to consider.For example, in a multi-party system that does not foster an "us versus them" mindset, a different type of personality is required to develop NPID in the first place.In other words, the bar might be higher for Swedish partisans to acquire NPID than for their American counterparts who have an instinctive out-party within their two-party system.This also has methodological implications.In this study, American partisans automatically received the NPID scale for the opposition party of their in-party, while Swedish partisans received the NPID scale only if they identified a political party that they would never vote for.This extra step might weaken the comparability of the NPID scales across samples since there might be American partisans who identify with one political party but might still be open to voting for the other.Future research might examine these contextual variations and their implications for measuring PPID and NPID in a comparative setting.
The partisan typology also has implications for political behavior.As the survey data shows in both samples, Positive Partisans are significantly more likely to vote than Negative Partisans, while Closed Partisans are significantly more likely to vote than Apathetic Partisans.At the same time, Closed Partisans are also significantly more likely to agree that "violence might sometimes be necessary to fight against parties and candidates that are bad for this country" and to believe that their "party's opponents are not just worse for politics-They are downright evil."From this perspective, it is especially vital to recognize these different types of partisans and to understand their different psychological origins.
Finally, this manuscript also sheds light on the scope of each partisan type in the electorate; only about 40% of partisans in both samples constitute Closed Partisans, while less than 20% comprise Negative Partisans.That still leaves about 40% of all partisans who are either Apathetic or purely Positive.This should spur more research into how to expand the share of Positive Partisans by, for example, turning Apathetic Partisans into Positive ones.

Figure 2 .
Figure2.Personality predictors of partisan types: US sample.Notes: Coefficients were estimated using a logistic regression model; all variables are scaled to range from 0 to 1; the corresponding table can be found in TableA13in the Supplementary File; *** p < 0.001, ** p < 0.01, * p < 0.05, + p < 0.1.

Figure 3 .
Figure3.Personality predictors of partisan types: Sweden sample.Notes: Coefficients were estimated using a logistic regression model; all variables are scaled to range from 0 to 1; the corresponding table can be found in TableA14in the Supplementary Files; *** p < 0.001, ** p < 0.01, * p < 0.05, + p < 0.1.

Figure 4 .Figure 5 .
Figure 4. Personality predictors of PPID and NPID among Democrats and Republicans: US sample.Notes: Coefficients were estimated using an OLS regression model; all variables are scaled to range from 0 to 1; the corresponding table can be found in Table A15 in the Supplementary File; *** p < 0.001, ** p < 0.01, * p < 0.05, + p < 0.1.

Table 1 .
Percentage shares of partisan types.
lower levels of Emotional Stability.Thus, the impact of Emotional Stability is radically different across the US and Sweden.NPID towards the left-i.e., the Democratic Party (Figure4)-is positively related to NfC, SDO, as well as Openness to Experience, and negatively related to Conscientiousness.The effects are particularly powerful across the range of NfC, whereby strong NPID towards Democrats intensifies from 0.13 to 0.40.The impact of NfC and Conscientiousness is robust to Bonferroni adjustments (see TableA30in the Supplementary File).