More than Two-Party Divides? Social Media,Ideological Extremity,and Affective Polarization in the Multi-party System

Abstract

How social media fuels affective polarization, characterized by favorability toward in-party members and apathy toward out-party members, has emerged as a crucial topic in political communication. However, few studies examine the relationship between social media and affective polarization in multi-party contexts outside American politics. Applying Wagner’s methodology for measuring affective polarization in a multi-party system, we calculate affective polarization in three ways: the traditional method as the absolute difference between liberal and conservative politicians, mean distance as the average distance from a person’s favorite politician, and spread as the average distance from an individual’s mean feelings toward politicians. We investigate how social media use is related to affective polarization using these three measures based on a survey of 1,159 respondents from Colombia in 2022. The findings indicate that the three measures of affective polarization intensify with the use of social media, mediated by ideological extremity. The results suggest that the role of social media and ideology may be consistent regardless of the way of measuring affective polarization when a country has multiple parties and political elites, further suggesting implications for affective polarization in political systems with more than two parties.

Keywords

social media affective polarization multi-party systems Colombia ideological extremity

Introduction

Social media has now become a main source for gaining news and information on a daily basis. In 2023, 68% of Americans use social media platforms, including Facebook, X, and Instagram (Gottfried, 2024). Latin America also has a growing social media audience, and journalists’ presence on social media has become salient (Harlow et al., 2023). Especially, Colombia has experienced a sharp penetration of social media, as 80% of urban adults reported receiving news from online or social media (García-Perdomo, 2023).

Concerning the widespread nature of social media, its role in facilitating an “us versus them” mentality along party lines has been of interest to scholars, as social media works as a platform for building information ecosystems with filter bubbles and saturating influence on affective polarization. In Latin America, popular social media platforms like WhatsApp and Facebook have been blamed for spreading misinformation and promoting hate crimes online (Harlow et al., 2023). More specifically, in Colombia, which has experienced guerrillas, mass protests, and distrust in election integrity in recent years, increased polarization has become an issue across politicians with different ideologies. Given the lack of focus on how social media might shape affective polarization in the Colombian context (Comellas & Torcal, 2023), it is worth exploring the relationship.

When studying affective polarization among multiple political figures from different parties, it is worthwhile to visit recent approaches. While the majority of early research on affective polarization has studied two-party systems, mainly focusing on American politics, there is a growing awareness demanding methodological and theoretical approaches to examine countries with more than two parties, as most countries live with multi-party systems (Magyar, 2022). The definitions of in-group and out-group in multi-party systems and the mechanisms of understanding polarization may be much more complex than those of bipartisanship; thus, nuanced approaches to measurement and analysis of affective polarization are required (Reiljan, 2020).

By employing the methodology proposed by Wagner (2021) for calculating affective polarization where the party system is not dichotomous, this article aims to extend the understanding of affective polarization to the multi-party system and the role of social media use in Colombia. Through a comparative analysis of the outcome variable, affective polarization, using the traditional bipolar method and Wagner’s (2021) method for multi-party settings, this article aims to map the subtle variation of partisan affect and its relationship with social media in the Colombian political landscape. In the Colombian case, a country with turbulent party politics characterized by the interplay between existing party systems and rising political forces, measuring Leader Affective Polarization (LAP) is more appropriate than measuring party polarization. This is because in Latin America, where party systems are weak (Lupu, 2016), Leader Affective Polarization captures an affective gap more effectively (see Areal, 2022; Harteveld, 2021b; Moraes & Béjar, 2024). The party system in Colombia has experienced significant dealignment and become volatile (Carreras et al., 2015) as many politicians create parties or alliances for a particular election, so the attachment is more to the leader of the party, and the dislike is more to the leader of another party, than to a particular political party. In recent years, two political figures, Álvaro Uribe and Gustavo Petro, have captured more effectively Colombian political divides as they play a strong role in fueling negative emotions toward opponents (Garzón-Velandia et al., 2024).

Moreover, the mediating role of ideological extremity cannot be disregarded. As digital media landscapes become more tailored and political discourses become more divided, ideological positions may become more extreme, reinforcing affective polarization (Sunstein, 1999). Thus, this study will examine the potential mediating role of ideological extremity on the relationship between social media consumption and affective polarization. Using a survey in Colombia collected in 2022, we performed a series of Hayes’ PROCESS macro model 4, to test the path from social media to ideological extremity and to affective polarization. Prior research has also employed similar models to capture the mediation from social media to political behavior or polarization (see Barnidge, 2015; Macafee et al., 2019; Starke et al., 2020). The results revealed a meaningful mediation among social media use, ideological extremity, and affective polarization, showing subtle differences across the different measures of polarization.

Literature Review

Social Media and Affective Polarization

In the early stages of social media’s growth, a promising outlook regarding its role in promoting democracy was prevalent, as seen in commentary around the Arab Spring. However, more recently, social media has been blamed for democratic backsliding, entrenching extreme viewpoints, limiting interactions with diverse networks, and exacerbating polarization (Harel et al., 2020; Tucker et al., 2017). Affective polarization, described as in-group attachment and out-group hostility, has become a focus of academic research in political communication as it captures partisan divides with social identity theories (Iyengar et al., 2019). Burgeoning social media platforms prompt questions of whether their use can shape affective polarization.

Sunstein (2018) has argued that people tend to selectively engage with like-minded information, and social media can be the online spaces that offer that type of information. Through algorithmic personalization and recommendation systems, these platforms often result in filter bubbles and echo chambers, providing information and messages that align with users’ pre-existing beliefs (Allcott et al., 2020; Bakshy et al., 2015). Empirical research has provided some support for this argument. For example, on social media platforms like Facebook and X, homophilic networks accounted for the majority of online interactions among users (Cinelli et al., 2021). In addition, deactivating Facebook leads to more interpersonal discussion and less political polarization (Allcott et al., 2020). Moreover, individuals tend to avoid counter-attitudinal information on social media even when they have a variety of choices (Bakshy et al., 2015). This can be explained by the pursuit of directional goals in motivated reasoning, seeking out sources that reinforce their pre-existing attitudes (Kunda, 1990). This confirmation bias leads to selective exposure to content on social media, further facilitating highly polarized subgroups on social media (Brugnoli et al., 2019).

Given that affective polarization is rooted in an “us versus them” mindset (Iyengar et al., 2019), exposure to attitude-consistent information on social media platforms can highlight users’ partisan identity, potentially resulting in intensified in-group attachment (Levendusky, 2013). Moreover, participating in political discussions on social media reinforces one’s pre-existing opinions (Cho et al., 2018). Through engaging in homogeneous enclaves on social media, users’ hostility toward out-group members becomes heightened (Harel et al., 2020).

Another negative aspect of dwelling on social media is that it is often rife with uncivil rhetoric and hate speech (Auxier, 2020). Social media serves as a platform for incidentally encountering political content (Nelson & Webster, 2017) and online extremism (Risius et al., 2024). Especially in Latin America, social media is a place for violence and threats against journalists (Harlow et al., 2023). Encountering negative comments about political adversaries and dehumanizing expressions has been linked to increased affective polarization (Harel et al., 2020; Suhay et al., 2018). Similarly, exposure to incivility on social media posts could foster negative feelings toward members of opposing parties (Brundidge & Garrett, 2024; Neyazi et al., 2023).

The widespread misinformation on social media poses another significant threat to democracy. A large body of research on misinformation has confirmed that fake news is prevalent on social media (Tucker et al., 2018). In a highly polarized environment, misinformation is more likely to go viral and spread online (Jenke, 2024). Furthermore, people favor news, including fake news, aligning with their pre-existing attitudes (Zimmer et al., 2019). They tend to believe misinformation that matches their political views compared to misinformation that supports opposing views or a mixture of both (Rhodes, 2022). What’s more, misinformation sharing on social media is often fueled by out-group hatred (Osmundsen et al., 2021), potentially creating a “vicious cycle” harmful to democracy.

The polarizing role of social media news has been found across diverse platforms and countries. News on social media is integrated with interactions with other users and additional information from them (Schulz et al., 2022). This is because, in terms of news content shared on social media, it is likely that people follow and share pro-attitudinal news and discuss it as a way of intensifying their existing political opinions (Levy, 2021; Lottridge & Bentley, 2018). Even though someone doesn’t follow a particular news source, algorithms on social media are likely to recommend news that aligns with their attitudes and ideologies (Swart, 2021). Moreover, news on social media often covers extreme exemplars of politics, eliciting strong feelings toward parties (Gill, 2022; Hasell & Weeks, 2016). For example, using WhatsApp and Snapchat for news intensifies affective polarization in America, and using Line for news increases affective polarization in Japan (Lee et al., 2022). A study in Finland found that Facebook news users have come to feel more affective polarization during the course of elections compared to consuming other news outlets in a multi-party context (van Erkel & Van Aelst, 2024). Furthermore, a study involving Colombian adults found that frequent social media users are more likely to read news on social media, and this engagement with political news on social media can also lead to increased political disagreement (Barnidge, 2015).

However, counter-arguments still exist that social media does not necessarily exacerbate affective polarization. This line of argument suggests that social media attenuates affective polarization because it helps develop social networks with cross-cutting opinions (McLeod & Lee, 2012). This type of diverse information facilitates learning about opposite viewpoints, which should enhance understanding between groups and reduce negative biases and hostile feelings toward members of an outgroup (Guess et al., 2023; Mutz, 2006). Kubin and von Sikorski’s (2021) meta-analysis on social media and affective polarization suggests that not all research conclusively supports either side of the argument. For example, several empirical studies have posited that the use of social media can exacerbate affective polarization (Cho et al., 2018; Iandoli et al., 2021), others find its effect to be negligible (Johnson et al., 2017), statistically insignificant (Boxell et al., 2017; Nordbrandt, 2021), or may even serve to mitigate affective polarization (Becker et al., 2019). Studies on the line of a minimal-effect approach show that exposure to news on social media or online, especially partisan news, does not predict increased polarization (Brundidge & Garrett, 2024; Guess et al., 2021).

Therefore, we are postulating a research question based on these mixed findings.

RQ1. Will social media use be related to affective polarization?

Ideological Extremity as a Mediator

Ideological segmentation has also been the interest of research in polarization as the dissemination of the Internet, cable, and news has been accompanied by ideological divides (Davis & Dunaway, 2016; Levendusky, 2009), and algorithms in social media allow users to be exposed to attitude-consistent content and further increase ideological polarization (Levy, 2021). Social media, by creating echo chambers and facilitating interactions among individuals with similar attitudes and ideologies, can allow exposure to extreme viewpoints and further endorse them (Flaxman et al., 2016; Hohmann et al., 2023; McPherson et al., 2001; Qureshi et al., 2020; Sunstein, 2018). Moreover, even repeated exposure to out-party members on social media accelerated partisans’ ideological polarization through backfire effects (Bail et al., 2018). Kubin and von Sikorski’s (2021) meta-analysis also concluded that all experimental studies confirmed that social media intensifies the left and right divides among people.

Such ideology, fueled by social media, could further lead to affective polarization, as ideologies often distinguished as left to right can signal people’s in-group and out-group status. A group-based approach and tribalism in party politics explain that ideological distance significantly shapes feelings toward opposite-party candidates and the broader party spectrum (Webster & Abramowitz, 2017). Social media use itself can indirectly relate to affective polarization through responses provoked by content and networks on social media, leading to out-group hostility and in-group favoritism (Gill, 2022). Thus, ideology derived from political identity may widen the affective polarization among people who express extreme ideologies (Comellas & Torcal, 2023).

Therefore, considering empirical evidence, it is plausible that entrenched ideological beliefs may act as a mediator between the relationship between social media and affective polarization. The relationships are visualized in Figure 1.

RQ2. Will ideological extremity mediate the relationship between social media use and affective polarization?

Figure 1.

Hypothesized model for social media, ideological extremity, and affective polarization.

Affective Polarization in Multi-Party Systems

In recent years, a growing body of literature has turned its attention to comparative contexts, suggesting that affective polarization could be a social phenomenon explaining democratic backsliding, public disbelief in governance, and social cleavages worldwide, as well as in America (Gidron et al., 2020; Reiljan, 2020). For example, Boxell et al. (2022) examined trends in affective polarization across 12 OECD countries and found that the United States experienced the sharpest rise. Their analysis indicated that five nations, including Canada, Denmark, France, New Zealand, and Switzerland, witnessed a slight increase in affective polarization, while the other countries, such as Australia, Britain, Germany, Japan, Norway, and Sweden, experienced a decline in when measuring the gap (Boxell et al., 2022). However, the status and trend of affective polarization might differ in countries other than the United States, given that political systems in those countries consist of many competing parties (Ferreira da Silva & Garzia, 2024; Gidron et al., 2020).

The operationalization of affective polarization is straightforward in a two-party system like the United States; we can measure affective polarization by subtracting in-party sentiments from those of a single out-party and taking the absolute value of it. On the other hand, it is much more complex for multi-party systems because of the number of parties, the definition of the out-party (Gidron et al., 2020), and the in-party (Reiljan, 2020). Addressing these difficulties of measuring affective polarization in multi-party systems, Wagner (2021) provides an advanced methodological approach to calculating affective polarization beyond bipolar-party systems. The major theoretical inquiry when assessing affective polarization in multi-party contexts is whether multiple parties can be perceived as in-groups since partisan attachments aren’t mutually exclusive or if only one party can play a role as an in-group since affection toward a party is incompatible (Gidron et al., 2020; Wagner, 2021). In light of this, Wagner (2021) developed two measures of calculating affective polarization in multi-party settings: (1) mean distance, an average gap from the most preferred party, assuming that the primary affection is toward a single party, and (2) spread, “the average absolute party like-dislike difference relative to each respondent’s average party like-dislike score” (pp. 10–11), supposing the possibility of favorable attachments for multiple parties.

Several recent studies have aligned with Wagner’s (2021) approach to gauging affective polarization in multi-party contexts. For instance, Torcal and Comellas (2022) performed a longitudinal study in Spain and found a rise in affective polarization among both leaders and voters. Further studies have also noted a heightening affective polarization between in-party affiliations and out-party animosity in countries like Norway (Knudsen, 2021), the Netherlands (Harteveld, 2021a), and Sweden (Kekkonen et al., 2022). Nevertheless, these studies largely focus on Europe and North America, leaving many countries on other continents wrestling with party divides and party proliferation.

Study Context: Colombian Case

Colombia maintains a formal democracy with regular elections. The country’s multi-party political system grew out of the conservative-liberal party divide in the 20th century. On the right side, parties like the Conservative Party and Centro Democrático prioritize free-market policies and a strong military. Centrist groups like Partido Liberal and Partido Verde focus on social reforms. On the other side, left-wing parties such as Colombia Humana and Partido Comunes propose a more expansive government intervention in protecting Colombian industry and land redistribution. The Colombian press largely operates as a market-based system that is linked to corporate interests and suffer from tends to be closely tied to big business interests and characterized as “weak legacy of media pluralism” (Waisbord, 2008, p. 3).

In 2021, a proposed tax reform led to massive social protests that quickly expanded from equality to other issues, such as education, inclusion, and peace. In this context, a coalition of left and center-left parties won the presidency in 2022 with Gustavo Petro. The 2022 elections included over 16 parties disputing congressional seats and five large coalitions of parties participating in the initial round for president. The personalization of politics in Colombia can be best understood with a popular slogan, “por el que diga Uribe,” meaning people in the right willing to vote for whoever the former president Uribe would say they should vote for. Center candidates, such as Sergio Fajardo ran under the auspices of the Alianza Social Independiente (ASI), a party he did not belong to, while the left formed an alliance under the rubric Pacto Historico, which came into existence in 2021, just months before the election.

In the Colombian context, then, while multiple parties have flourished in the XXI century, most people feel identified with a particular leader in a center-left continuum, rather than with a political party. To take into account the peculiarities of the case under study, we conceptualize affective polarization as out-group animosity toward particular politicians, rather than the members of a particular party, which is more common in other contexts. In addition, for the 2022 election, the most well-known right-wing candidates did not make it to the second round of the election, which was ultimately between a consolidated leader on the left, Petro, and a newcomer, Rodolfo Hernandez, who captured voters from the right.

Regarding affective polarization in this context, there is a lack of relevant studies. Recently, Harteveld (2021b) found that affective polarization tends to be lower in Latin America compared to North America. Comellas and Torcal (2023) specifically focused on affective polarization in Argentina, Chile, Italy, Portugal, and Spain—countries with multi-party systems—and found that ideological identity accelerates affective polarization in these systems. One of the unique characteristics of politics in Latin America is that parties have weak ideological alignment and structure (Harteveld, 2021b; Lupu, 2016; Moraes & Béjar, 2024). Considering this, it is more appropriate to measure Leader Affective Polarization (LAP) in Latin America, where weak ideology allows developing affective polarization through out-party leaders and candidates as suggested in the line of literature (e.g., Areal, 2022; Moraes & Béjar, 2024; Reiljan et al., 2023).

However, there are only a handful of studies exploring affective polarization in Colombia. Scherman et al. (2022) investigated general WhatsApp use and affective polarization in Colombia and found no social media effects on affective polarization, which was computed by the absolute difference between ratings given to the government’s leader and the opposition party’s leader, divided by the number of leaders. Despite the lack of effects in Scherman et al. (2022), social media can leverage affective polarization when comparing operationalizations of it.

In the current article, we will examine three measures of LAP for comparison based on Wagner (2021): (a) the traditional bipolar method, which calculates the bipolar absolute difference between the two major candidates or leaders, (b) mean distance, based on the assumption that an individual has one most favored politician, and (c) spread, working on the presumption that people can identify with multiple politicians. We will then assess the relationships between social media and these three measures within the Colombian context.

Methods

Data

The research used national survey data collected between July 5 and August 3, 2022, across 10 Colombian cities. This was a collaborative project between two universities to study political behaviors and communication in Colombia post-election. The task included the Registraduria Nacional del Estado Civil de Colombia for data collection. The survey was designed to reflect Colombia’s urban adult demographic, given that 76.7% of Colombia’s total population of 49.8 million reside in urban settings (DANE, 2018).

The data was collected using multi-stage stratified random sampling by selecting households according to city size and census information. Within each city, city blocks were chosen based on housing sectors and strata. Specific households within these blocks were randomly selected. To decide on the individual respondent from each household, the “adult in the household who most recently celebrated a birthday” approach was used. To ensure survey participation, up to three visits were made to each house if necessary. Data collection was administered by Deproyectos Limitada, a local expert survey agency in Colombia, resulting in 1,159 in-person responses, a 47% response rate.

Measures

Ideological Extremity

Respondents were asked to place their political ideologies on a scale from 0 (left) to 10 (right). To measure their ideological extremity, we recoded the item ranging from 0 (those who identified as centrist) to 5 (those who ranked themselves at either extreme end, 0 or 10). Higher scores on this scale represent stronger ideologies (M = 1.64, SD = 1.86).

Social Media Use

Respondents were asked to report their frequency of social media usage, ranging from 0 (Never) to 5 (Frequently): (1) WhatsApp, (2) Facebook, (3) Twitter (Now X), and (4) TikTok. We also asked respondents to report their frequency of social media news use from 0 (Never) to 5 (Frequently) among people who said they used social media more than 0 (Never) in the previous question. We took the mean of general social media use and social media news use items (M = 2.61, SD = 1.11).¹

Affective Polarization (LAP)

We asked respondents to rate their feelings toward political leaders of a party from 0 (not much favorable) to 10 (very favorable), like the previous literature using likes/dislikes of political leaders to measure LAP (Iyengar et al., 2019; Klar et al., 2018; Knudsen, 2021). The political figures included in our survey were Ivan Duque (standing president from the Democratic Center), Rodolfo Hernández (League of Anti-Corruption Governors), Gustavo Petro (elected president from Humane Colombia), Álvaro Uribe (Centro Democrático), César Gaviria (Liberal), and Antanas Mockus (Alianza Verde). While Uribe and Duque were not on the ballot for this election, they represented the right wing’s most salient political leaders at the time.

These three measures were employed to investigate the research questions. First, the traditional bipolar method involved computing the absolute difference in feelings toward Rodolfo Hernández and Gustavo Petro, both leading candidates in the 2022 presidential election in Colombia (M = 4.31, SD = 3.28). The mean distance was calculated by averaging the discrepancy from the most favored politician, ranging from 0 to 10 (M = 4.26, SD = 2.46). For instance, if an individual’s preferred politician was Duque, the mean absolute distance from other politicians, relative to feelings for Duque, was calculated. Spread was computed by averaging the distance from an individual’s mean feelings toward the listed six politicians, ranging from 0 to 5 (M = 2.12, SD = 1.17).

Control Variables

Demographics including age (M = 43.19, SD = 15.54), gender (male = 45.66%), race (White = 10.47%), education measured with an 11-point scale from 1 (none) to 11 (graduate school) (M = 6.82, SD = 2.50), and monthly household income measured with an 8-point scale from 1 (less than $1,000,000) to 8 (More than $10,000,000) (M = 2.40, SD = 1.46) were controlled. Political interest ranges from 0 (Not at all interested) to 3 (Very interested) (M = 1.27, SD = 0.95), and party identification, asking whether the respondent identified with any party (0 = No; 1 = Yes; Yes = 26.93%), was also controlled. In addition, political news use (M = 2.02, SD = 1.83), a mean index of TV, radio, and newspapers consumption (0 = Never; 7 = seven days per week), was also used as one of the control variables. We also controlled for ideological polarization by using ideological extremity in the model.

Analytic Procedures

To test the relationships between social media, ideological extremity, and affective polarization, we first ran OLS regressions on the three measures of affective polarization: bipolar, mean distance, and spread. Then, we performed PROCESS macro model 4 of Hayes (2013) to investigate the proposed model in Figure 1.

Results

First, we performed OLS regressions on three measures of affective polarization. As shown in Table 1, the model fits as 10% of the binary measure, 12% of the mean distance, and 14% of the spread of affective polarization were explained. Addressing RQ1, social media use was not significantly related to none of the affective polarization measures (bipolar: β = .05, p > .05; mean distance: β = .04, p > .05; spread: β = .05, p > .05). Ideological extremity was significantly related to affective polarization, measured as bipolar (β = .15, p < .001), followed by mean distance (β = .11, p < .01) and spread (β = .12, p < .001).

Table 1.

OLS Regressions on Affective Polarization.

	Model 1	Model 2	Model 3
	Bipolar	Mean distance	Spread
Age	-0.08. (0.04)	-0.04 (0.04)	-0.02 (0.04)
Gender	0.03 (0.04)	-0.03 (0.03)	-0.04 (0.03)
Education	-0.07 (0.04)	0.05 (0.04)	0.04 (0.04)
Income	0.03 (0.04)	-0.03 (0.04)	-0.03 (0.04)
Race	0.02 (0.04)	-0.05 (0.03)	0.004 (0.03)
Political interest	0.04 (0.04)	0.02 (0.04)	0.07. (0.04)
Partisanship	0.18*** (0.04)	0.24*** (0.04)	0.22*** (0.04)
News consumption	0.01 (0.04)	0.05 (0.04)	0.08* (0.04)
Ideological extremity	0.15*** (0.04)	0.11** (0.04)	0.12*** (0.04)
Social media use	0.05 (0.04)	0.04 (0.04)	0.05 (0.04)
N	754	794	795
R ²	.10	.12	.14

Note. All coefficients are standardized.

p < .05, **p < .01, ***p < .001.

To test the indirect effects as hypothesized in Figure 1, we performed Hayes’ PROCESS macro model 4. The model used 5,000 bootstrap samples to construct 95% confidence intervals (CIs). Figure 2 illustrates the general findings of regression coefficients and significance across the three measures of affective polarization. Table 2 presents both indirect pathways, testing the proposed mediation model in Figure 1.

Figure 2.

Summary of regression coefficients for affective polarization.

Table 2.

Coefficients for the Mediation Model of Affective Polarization.

	Estimate	SE	LLCI, ULCI
DV1: Affective Polarization (Bipolar)
Social media use → Ideological Extremity	0.027	0.010	0.011, 0.049
DV2: Affective Polarization (Mean distance)
Social media use → Ideological Extremity	0.021	0.008	0.007, 0.039
DV3: Affective Polarization (Spread)
Social media use → Ideological Extremity	0.022	0.008	0.008, 0.040

Note. This table presents standardized coefficients for the mediation models of Hayes’ PROCESS macro model 4.

For RQ2, findings indicated that ideological extremity mediates the relationship between social media use and affective polarization. The size of coefficients was most pronounced for the traditional bipolar measure (β = .027, SE = 0.010, 95% CI: [.011, .049]), compared to the coefficients of mean distance (β = .021, SE = 0.008, 95% CI: [.007, .039]) and spread (β = .022, SE = 0.008, 95% CI: [.008, .040]). Therefore, RQ2 receives empirical support for the path from social media use, ideological extremity, and affective polarization, regardless of how we measure it. This indicates that social media use does not have a direct relationship with affective polarization; however, it is indirectly related to increased affective polarization through heightened ideological extremity.

Discussion

The findings in this study provide a nuanced understanding of the role of social media in shaping political affective polarization in Colombia. In summary, social media use was not directly related to affective polarization. This result seems to diverge from the argument that social media use is a strong predictor of polarization (see Bail et al., 2018). Instead, the results suggested that social media use mediated by ideological extremity plays a significant role in aggravating affective polarization.

The mediation analysis indicated that extreme ideologies, which can be amplified by social media, contribute to heightened affective polarization despite the measurement strategy employed. This finding contributes to the discourse on the relationship between online platforms and polarization, highlighting the importance of underlying ideological dynamics rather than attributing polarization solely to social media use. This echoes the previous studies suggesting that the social media environment promotes political polarization (see Sunstein, 2018). In addition, recent research indicates that partisan media outlets, significant influencers on social media platforms, often stick to motivated reasoning and encourage users to seek out information that aligns with their existing beliefs (see Lee, 2021; Lee et al., 2022), and algorithms on social media filter counter-attitudinal content based on user’s social networks and online activities (Flaxman et al., 2016; Fletcher & Nielsen, 2018). This further indicates that using social media is likely to create an insular environment where users predominantly receive news and information from like-minded sources, additionally enhancing extreme ideologies and affective polarization.

Returning to the methodological attempt of this study to compare different measures of affective polarization in a multi-party context, the results offered consistent and clear patterns: regardless of how we measure the concept, social media use was indirectly associated with affective polarization. The size of the coefficients also suggests that frequent social media use is consistently related to heightened affective polarization, and the relationship is most pronounced when measured as the traditional absolute difference between two major politicians compared to a multi-party measure. This counterintuitive result might stem from the fact that in Colombia, despite the existence of multiple parties, two clear poles emerge on a left-right continuum that might resemble a two-party system. It could also be that the indirect effects of social media use may trigger party cues for well-known and famous politicians, eliciting stronger likes or dislikes toward those presidential candidates, while the effect is persistent but not as large for less well-known politicians. In a country like Colombia, without a strong party system, but with politicians clearly aligned around two poles, the affective gap between the two major politicians might emerge more dramatically. Future studies should test this relationship in countries with multi-party systems and more consolidated parties.

The results call for more attention to the specific context. The 2022 presidential election, which brought Colombia its first left-wing president, Gustavo Petro, highlights the shifts and fluctuations in polarization within the country. Social media might serve as a platform for receiving partisan cues, shaping ideological leanings, and fostering leader-based affective polarization. While it is acknowledged that Latin America generally exhibits less ideological alignment along party lines (Harteveld, 2021b), some countries, such as Peru and Uruguay, stand out for their extremely high levels of polarization (Wagner, 2021).

The limitations of this study should be noted, such as its focus on urban populations, which may not be generalizable to rural settings where social dynamics and social media consumption patterns may differ. Future research could explore these phenomena in different contexts. Another limitation could be the cross-sectional nature of the survey. Although we tested the mediation models, it is difficult to argue causal relationships. Future studies should address this issue by conducting experiments. In addition, we focused on a single Latin American country. Multi-party systems, particularly in different regions, can present a distinct political environment, characterized by phenomena like the rise of populism, technocracy, or strongman leadership. Therefore, future research should expand its geographical scope to corroborate whether social media’s indirect role in affective polarization remains consistent across diverse settings.

Conclusion

This study contributes to the literature by addressing the pathways through social media, ideological extremity, and affective polarization in a post-election scenario in Colombia. The findings elucidate that while the use of social media is not directly related to affective polarization except for the bipolar measure, it does have an indirect effect by exacerbating ideological extremities that contribute to affective polarization.

This study could be further extended to different political contexts with multi-party settings. Even within multi-party settings, party systems are varied in terms of their size, number of parties, and history of the political system (Magyar, 2022). Coalition, competition, negotiations among parties, and intra-party interactions might be potential factors shaping the social media effects on polarization (Gross & Sigelman, 1984; Knutsen, 2017). Future studies might account for those systemic differences across countries when measuring polarization with multiple parties. Furthermore, the social media landscape is also diverged across countries; for example, some countries have local social media that dominate their market (Sun et al., 2025); thus, future studies should also consider these contextual differences.

Footnotes

ORCID iD

Heysung Lee

Ethical considerations

The study was approved by the IRB at the University of Wisconsin-Madison.

Consent to participate

Informed consent to participate in this study was obtained from all participants. Participants were provided with a clear explanation of the research objectives, procedures, potential risks, and benefits in a manner ensuring comprehension. Consent was obtained prior to participation in the online survey, with participants voluntarily agreeing to take part in the study after their questions, if any, were addressed. They were also able to opt out of the survey if they disagreed.

Consent for publication

This study utilized an online survey that did not collect any identifiable personal information from participants. All responses were anonymized, and no data can be traced back to individual participants. The research complies with ethical standards for anonymity and data protection.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the University of Wisconsin Foundation (AAA2223).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

The data used in this study are publicly available through the Comparative National Election Project (CNEP) website ().

Notes

Author biographies

Heysung Lee is a postdoctoral scholar at the Institute of Comparative Governance at Korea University. She received a PhD from the School of Journalism and Mass Communication at the University of Wisconsin-Madison. Her research interests broadly include three major lines of inquiry: (a) the information-seeking behaviors in mass and interpersonal communication and their relationship with political attitudes; (b) polarization stirred by news use and perception of media and public opinion; and (c) cross-cultural variations in media consumption and democratic engagement, considering cultural and historical contexts.

Hernando Rojas is Professor of Journalism and Mass Communication, Helen Firstbrook Franklin Chair in the School of Journalism and Mass Communication at the University of Wisconsin-Madison. His scholarship focuses on political communication, in particular examining: (a) the deployment of new communication technologies for social mobilization in a variety of contexts; (b) the influence of audience perceptions of media (and audience perceptions of media effects) on both public opinion and the structure of the public sphere; and (c) the conditions under which media support democratic governance.

References

Allcott

Braghieri

Eichmeyer

Gentzkow

(2020). The welfare effects of social media. American Economic Review, 110(3), 629–676. https://doi.org/10.1257/aer.2019060658

Areal

(2022). ‘Them’ without ‘us’: Negative identities and affective polarization in Brazil. Political Research Exchange, 4(1), 2117635. https://doi.org/10.1080/2474736X.2022.2117635

Auxier

. (2020, October 15). 64% of Americans say social media have a mostly negative effect on the way things are going in the U.S. today. Pew Research Center. https://www.pewresearch.org/short-reads/2020/10/15/64-of-americans-say-social-media-have-a-mostly-negative-effect-on-the-way-things-are-going-in-the-u-s-today/

Bail

C. A.

Argyle

L. P.

Brown

T. W.

Bumpus

J. P.

Chen

Hunzaker

M. F.

. . .Volfovsky

(2018). Exposure to opposing views on social media can increase political polarization. Proceedings of the National Academy of Sciences, 115(37), 9216–9221. https://doi.org/10.1073/pnas.1804840115

Bakshy

Messing

Adamic

L. A.

(2015). Exposure to ideologically diverse news and opinion on Facebook. Science, 348(6239), 1130–1132. https://doi.org/10.1126/science.aaa1160

Barnidge

(2015). The role of news in promoting political disagreement on social media. Computers in Human Behavior, 52, 211–218. https://doi.org/10.1016/j.chb.2015.06.011

Becker

Porter

Centola

(2019). The wisdom of partisan crowds. Proceedings of the National Academy of Sciences, 116(22), 10717–10722. https://doi.org/10.1073/pnas.1817195116

Boxell

Gentzkow

Shapiro

J. M.

(2017). Greater Internet use is not associated with faster growth in political polarization among US demographic groups. Proceedings of the National Academy of Sciences, 114(40), 10612–10617. https://doi.org/10.1073/pnas.1706588114

Boxell

Gentzkow

Shapiro

J. M.

(2022). Cross-country trends in affective polarization. Review of Economics and Statistics, 106(2), 557–565. https://doi.org/10.1162/rest_a_01160

10.

Brugnoli

Cinelli

Quattrociocchi

Scala

(2019). Recursive patterns in online echo chambers. Scientific Reports, 9(1), 20118. https://doi.org/10.1038/s41598-021-83758-0

11.

Brundidge

Garrett

R. K.

(2024). The “clinching effect” and affective polarization: Exposure to incivility via social media in the presence of online news. International Journal of Public Opinion Research, 36(3), edae042. https://doi.org/10.1093/ijpor/edae042

12.

Carreras

Morgenstern

Y. P.

(2015). Refining the theory of partisan alignments: Evidence from Latin America. Party Politics, 21(5), 671–685. https://doi.org/10.1177/1354068813491538

13.

Cho

Ahmed

Kerum

Choi

Y. J.

Lee

J. H.

(2018). Influencing myself: Self-reinforcement through online political expression. Communication Research, 45(1), 83–111. https://doi.org/10.1177/0093650216644020

14.

Cinelli

De Francisci Morales

Galeazzi

Quattrociocchi

Starnini

(2021). The echo chamber effect on social media. Proceedings of the National Academy of Sciences, 118(9), Article e2023301118. https://doi.org/10.1073/pnas.2023301118

15.

Comellas

J. M.

Torcal

(2023). Ideological identity, issue-based ideology and bipolar affective polarization in multiparty systems: The cases of Argentina, Chile, Italy, Portugal and Spain. Electoral Studies, 83, 102615. https://doi.org/10.1016/j.electstud.2023.102615

16.

DANE. (2018). Colombia’s national department of statistics-DANE. www.dane.gov.co

17.

Davis

N. T.

Dunaway

J. L.

(2016). Party polarization, media choice, and mass partisan-ideological sorting. Public Opinion Quarterly, 80(Suppl. 1), 272–297.

18.

Ferreira da Silva

Garzia

(2024). Affective polarization towards parties and leaders, and electoral participation in 13 parliamentary democracies, 1980–2019. Public Opinion Quarterly, 88(4), 1234–1248. https://doi.org/10.1093/poq/nfae053

19.

Flaxman

Goel

Rao

J. M.

(2016). Filter bubbles, echo chambers, and online news consumption. Public Opinion Quarterly, 80(Suppl. 1), 298–320. https://doi.org/10.1093/poq/nfw006

20.

Fletcher

Nielsen

R. K.

(2018). Are people incidentally exposed to news on social media? A comparative analysis. New Media & Society, 20(7), 2450–2468. https://doi.org/10.1177/1461444817724170

21.

García-Perdomo

(2023, June 14). Digital news report 2023: Colombia. Reuters Institute for the Study of Journalism, University of Oxford. https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2023/colombia

22.

Garzón-Velandia

D. C.

Barreto-Galeano

M. I.

Sabucedo-Cameselle

J. M.

(2024). When political elites talk, citizens reply. Affective polarization through temporal orientation and intergroup emotions. Analyses of Social Issues and Public Policy, 24(3), 621–644. https://doi.org/10.1111/asap.12416

23.

Gidron

Adams

Horne

(2020). American affective polarization in comparative perspective. Cambridge University Press.

24.

Gill

(2022). Testing the effect of cross-cutting exposure to cable TV news on affective polarization: Evidence from the 2020 US presidential election. Journal of Broadcasting & Electronic Media, 66(2), 320–339. https://doi.org/10.1080/08838151.2022.2087653

25.

Gottfried

(2024, January 31). Americans’ social media use. Pew Research Center. https://www.pewresearch.org/internet/2024/01/31/americans-social-media-use/

26.

Gross

D. A.

Sigelman

(1984). Comparing party systems: A multidimensional approach. Comparative Politics, 16(4), 463–479. https://doi.org/10.2307/421950

27.

Guess

A. M.

Barberá

Munzert

Yang

(2021). The consequences of online partisan media. Proceedings of the National Academy of Sciences, 118(14), Article e2013464118. https://doi.org/10.1073/pnas.2013464118

28.

Guess

A. M.

Malhotra

Pan

Barberá

Allcott

Brown

. . .Tucker

J. A.

(2023). How do social media feed algorithms affect attitudes and behavior in an election campaign? Science, 381(6656), 398–404. https://doi.org/10.1126/science.abp9364

29.

Harel

T. O.

Jameson

J. K.

Maoz

(2020). The normalization of hatred: Identity, affective polarization, and dehumanization on Facebook in the context of intractable political conflict. Social Media + Society, 6(2), 2056305120913983. https://doi.org/10.1177/2056305120913983

30.

Harlow

Wallace

Cueva Chacon

(2023). Digital (in) security in Latin America: The dimensions of social media violence against the press and journalists’ coping strategies. Digital Journalism, 11(10), 1829–1847. https://doi.org/10.1080/21670811.2022.2128390

31.

Harteveld

(2021a). Fragmented foes: Affective polarization in the multiparty context of the Netherlands. Electoral Studies, 71, 102332. https://doi.org/10.1016/j.electstud.2021.102332

32.

Harteveld

(2021b). Ticking all the boxes? A comparative study of social sorting and affective polarization. Electoral Studies, 72, 102337. https://doi.org/10.1016/j.electstud.2021.102337

33.

Hasell

Weeks

B. E.

(2016). Partisan provocation: The role of partisan news use and emotional responses in political information sharing in social media. Human Communication Research, 42(4), 641–661. https://doi.org/10.1111/hcre.12092

34.

Hayes

A. F.

(2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford.

35.

Hohmann

Devriendt

Coscia

(2023). Quantifying ideological polarization on a network using generalized Euclidean distance. Science Advances, 9(9), eabq2044. https://doi.org/10.1126/sciadv.abq2044

36.

Iandoli

Primario

Zollo

(2021). The impact of group polarization on the quality of online debate in social media: A systematic literature review. Technological Forecasting and Social Change, 170, 120924. https://doi.org/10.1016/j.techfore.2021.120924

37.

Iyengar

Lelkes

Levendusky

Malhotra

Westwood

S. J.

(2019). The origins and consequences of affective polarization in the United States. Annual Review of Political Science, 22, 129–146. https://doi.org/10.1146/annurev-polisci-051117-073034

38.

Jenke

(2024). Affective polarization and misinformation belief. Political Behavior, 46(2), 825–884. https://doi.org/10.1007/s11109-022-09851-w

39.

Johnson

T. J.

Kaye

B. K.

Lee

A. M.

(2017). Blinded by the spite? Path model of political attitudes, selectivity, and social media. Atlantic Journal of Communication, 25(3), 181–196. https://doi.org/10.1080/15456870.2017.1324454

40.

Kekkonen

Suuronen

Kawecki

Strandberg

(2022). Puzzles in affective polarization research: Party attitudes, partisan social distance, and multiple party identification. Frontiers in Political Science, 4, Article 920567. https://doi.org/10.3389/fpos.2022.920567

41.

Klar

Krupnikov

Ryan

J. B.

(2018). Affective polarization or partisan disdain? Untangling a dislike for the opposing party from a dislike of partisanship. Public Opinion Quarterly, 82(2), 379–390. https://doi.org/10.1093/poq/nfy014

42.

Knudsen

(2021). Affective polarization in multiparty systems? Comparing affective polarization towards voters and parties in Norway and the United States. Scandinavian Political Studies, 44(1), 34–44. https://doi.org/10.1111/1467-9477.12186

43.

Knutsen

(2017). Social structure, value orientations and party choice in Western Europe. Springer.

44.

Kubin

von Sikorski

(2021). The role of (social) media in political polarization: A systematic review. Annals of the International Communication Association, 45(3), 188–206. https://doi.org/10.1080/23808985.2021.1976070

45.

Kunda

(1990). The case for motivated reasoning. Psychological Bulletin, 108(3), 480.

46.

Lee

(2021). “Easy” and “hard” issues: Attitude extremity and a role of the need to evaluate. Social Science Quarterly, 102(6), 2930–2941. https://doi.org/10.1111/ssqu.13094

47.

Lee

Rojas

Yamamoto

(2022). Social media, messaging apps, and affective polarization in the United States and Japan. Mass Communication and Society, 25(5), 673–697. https://doi.org/10.1080/15205436.2021.1953534

48.

Levendusky

(2009). The partisan sort: How liberals became Democrats and conservatives became Republicans. University of Chicago Press.

49.

Levendusky

M. S.

(2013). Why do partisan media polarize viewers? American Journal of Political Science, 57(3), 611–623. https://doi.org/10.1111/ajps.12008

50.

Levy

R. E.

(2021). Social media, news consumption, and polarization: Evidence from a field experiment. American Economic Review, 111(3), 831–870. https://doi.org/10.1257/aer.20191777

51.

Lottridge

Bentley

F. R.

(2018). Let’s hate together: How people share news in messaging, social, and public networks. In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems (pp. 1–13). Association for Computing Machinery. https://doi.org/10.1145/3173574.3173634

52.

Lupu

(2016). Party brands in crisis: Partisanship, brand dilution, and the breakdown of political parties in Latin America. Cambridge University Press.

53.

Macafee

McLaughlin

Rodriguez

N. S.

(2019). Winning on social media: Candidate social-mediated communication and voting during the 2016 US presidential election. Social Media + Society, 5(1), 2056305119826130. https://doi.org/10.1177/205630511982613

54.

Magyar

Z. B.

(2022). What makes party systems different? A principal component analysis of 17 advanced democracies 1970–2013. Political Analysis, 30(2), 250–268. https://doi.org/10.1017/pan.2021.21

55.

McLeod

J. M.

Lee

N. J.

(2012). Social networks, public discussion, and civic engagement: A socialization perspective. In Semetko

Scammell

(Eds.), The Sage handbook of political communication (pp. 197–208). Sage.

56.

McPherson

Smith-Lovin

Cook

J. M.

(2001). Birds of a feather: Homophily in social networks. Annual Review of Sociology, 27(1), 415–444. https://doi.org/10.1146/annurev.soc.27.1.415

57.

Moraes

Béjar

(2024). Affective polarization in Latin America. In Torcal

Harteveld

(Eds.), Handbook of affective polarization (pp. 133–147). Edward Elgar.

58.

Mutz

D. C.

(2006). Hearing the other side: Deliberative versus participatory democracy. Cambridge University Press.

59.

Nelson

Webster

(2017). The myth of partisan selective exposure: A portrait of the online political news audience. Social Media + Society, 3(3), 205630511772931. https://doi.org/10.1177/2056305117729314

60.

Neyazi

T. A.

Kuru

Mukerjee

(2023). Political campaign ads on Facebook: Investigating the effects of incivility in videos and user comments on affective polarization and mobilization. International Journal of Communication, 17, 24.

61.

Nordbrandt

(2021). Affective polarization in the digital age: Testing the direction of the relationship between social media and users’ feelings for out-group parties. New Media & Society, 25(12), 3392–3411. https://doi.org/10.1177/14614448211044393

62.

Osmundsen

Bor

Vahlstrup

P. B.

Bechmann

Petersen

M. B.

(2021). Partisan polarization is the primary psychological motivation behind political fake news sharing on Twitter. American Political Science Review, 115(3), 999–1015i. https://doi.org/10.1017/S0003055421000290

63.

Qureshi

Bhatt

Gupta

Tiwari

A. A.

(2020). Causes, symptoms and consequences of social media induced polarization (SMIP). Information Systems Journal, 11.

64.

Reiljan

(2020). ‘Fear and loathing across party lines’(also) in Europe: Affective polarisation in European party systems. European Journal of Political Research, 59(2), 376–396. https://doi.org/10.1111/1475-6765.12351

65.

Rhodes

S. C.

(2022). Filter bubbles, echo chambers, and fake news: How social media conditions individuals to be less critical of political misinformation. Political Communication, 39(1), 1–22. https://doi.org/10.1080/10584609.2021.1910887

66.

Risius

Blasiak

K. M.

Wibisono

Louis

W. R.

(2024). The digital augmentation of extremism: Reviewing and guiding online extremism research from a sociotechnical perspective. Information Systems Journal, 34(3), 931–963. https://doi.org/10.1111/isj.12454

67.

Scherman

Etchegaray

Browne

Mazorra

Rojas

(2022). WhatsApp, polarization, and non-conventional political participation: Chile and Colombia before the social outbursts of 2019. Media and Communication, 10(4), 77–93. https://doi.org/10.17645/mac.v10i4.5817

68.

Schulz

Fletcher

Nielsen

R. K.

(2022). The role of news media knowledge for how people use social media for news in five countries. New Media & Society, 26(7), 4056–4077. https://doi.org/10.1177/14614448221108957

69.

Starke

Marcinkowski

Wintterlin

(2020). Social networking sites, personalization, and trust in government: Empirical evidence for a mediation model. Social Media + Society, 6(2), 2056305120913885. https://doi.org/10.1177/2056305120913885

70.

Suhay

Bello-Pardo

Maurer

(2018). The polarizing effects of online partisan criticism: Evidence from two experiments. The International Journal of Press/Politics, 23(1), 95–115. https://doi.org/10.1177/1940161217740697

71.

Sun

Liao

Rojas

(2025). Regionality of messaging applications: Everyday socialization as a gateway to civic and political participation in China, Japan, South Korea, and the United States. Social Media + Society, 11(2), 20563051251344076. https://doi.org/10.1177/20563051251344076

72.

Sunstein

(2018). # Republic: Divided democracy in the age of social media. Princeton University Press.

73.

Sunstein

C. R.

(1999). The law of group polarization [John M. Olin Law & Economics Working Paper(91)]. University of Chicago Law School.

74.

Swart

(2021). Experiencing algorithms: How young people understand, feel about, and engage with algorithmic news selection on social media. Social media + Society, 7(2), 20563051211008828. https://doi.org/10.1177/20563051211008828\

75.

Torcal

Comellas

J. M.

(2022). Affective polarisation in times of political instability and conflict. Spain from a comparative perspective. South European Society and Politics, 27(1), 1–26. https://doi.org/10.1080/13608746.2022.2044236

76.

Tucker

J. A.

Guess

Barberá

Vaccari

Siegel

Sanovich

Stukal

Nyhan

(2018). Social media, political polarization, and political disinformation: A review of the scientific literature. Hewlett Foundation. https://doi.org/10.2139/ssrn.3144139

77.

Tucker

J. A.

Theocharis

Roberts

M. E.

Barberá

(2017). From liberation to turmoil: Social media and democracy. Journal of Democracy, 28(4), 46–59. https://doi.org/10.1353/jod.2017.0064

78.

van Erkel

P. F. A

Van Aelst

. (2024). Social media and affective polarization: does Facebook news use fuel political in-and out-group affect in a multi-party context? Acta Politica, 60, 286–306. https://doi.org/10.1057/s41269-023-00324-4

79.

Wagner

(2021). Affective polarization in multiparty systems. Electoral Studies, 69, 102199. https://doi.org/10.1016/j.electstud.2020.102199

80.

Waisbord

(2008, May 9). Press and the public sphere in contemporary Latin America [Conference session]. Paper presented at the Harvard-World Bank Workshop, Cambridge, MA, United States.

81.

Webster

S. W.

Abramowitz

A. I.

(2017). The ideological foundations of affective polarization in the US electorate. American Politics Research, 45(4), 621–647.

82.

Zimmer

Scheibe

Stock

W. G.

(2019). Fake news in social media: Bad algorithms or biased users? Journal of Information Science Theory and Practice, 7(2), 40–53.