Political Polarisation

Polarisation can take a number of forms, encompassing a range of political and social dynamics that differ not only in form but also in how they manifest across contexts. At its most basic level, polarisation refers to the degree to which political actors are divided along salient cleavages, but the form of polarisation and its consequences vary depending on who is polarised and along what dimensions. Table 1 summarises how we conceptualise different forms of polarisation as a function of the type of polarisation and actors involved.

The literature typically distinguishes between three different types of polarisation, each of which can apply to different actors: 1. ideological polarisation, or divergence in general policy preferences; 2. issue polarisation, or divergence on specific political topics; and 3. affective polarisation, or animosity towards opposing political groups. While these dimensions often overlap – ideological and affective polarisation are related phenomena, for example – descriptive evidence suggests that they do not always evolve in tandem (Riera and Madariaga 2023).

Within these categories, polarisation can centre around different actors: 1. individuals may become more extreme or hostile; 2. elites may adopt divisive rhetoric; and 3. media ecosystems may amplify or reflect polarised discourse.

Table 1: Conceptualising Political Polarisation

Note: Polarisation can manifest through ideological worldviews, concrete policy disagreements, or emotional and social hostility. These dimensions often interact but require distinct measurement strategies.

Furthermore, while much of the academic literature on advanced democracies uses the term ‘polarisation’ to describe these political dynamics, similar patterns are widely observed across more fluid political systems – even if labelled differently. The functional equivalents of polarisation can appear as identity-based political distrust, partisan social sorting, and information-processing biases aligned with political cleavages. These dynamics may be framed around ethnicity, religion, or regional divisions rather than ideological or policy positions, but they similarly shape citizen behaviour, elite strategy, and institutional performance. Research in contexts like Ghana, India, Kenya, Lebanon, and the Philippines has documented how group-based appeals, social segmentation, and divisive campaign rhetoric can fragment accountability, reduce responsiveness to citizens' policy requests, and heighten affective mistrust between political camps. While these studies may not always use the language of affective or ideological polarisation, they offer critical insights into how similar dynamics play out in low- and middle-income settings.

Ideological Polarisation

Ideological polarisation refers to the increasing ideological divide among individuals within a society, typically marked by a shift towards more extreme political views and a movement away from centrist positions. In terms of measurement, ideological polarisation is often measured using spatial distance between policy preferences: using survey scales, text analysis of party positions, or ideal point estimates (such as the DW-Nominate scores commonly used to measure polarisation in US Congress, e.g. Poole and Rosenthal 1985, Canen et al. 2020). One way that ideological polarisation is distinguished from affective polarisation is that ideological polarisation can occur without hostility between groups, especially when it centres around principled disagreement over economic or social policies.

Ideological polarisation has been documented not only in countries with more established political parties like the US, but also in countries where partisan competition is newer or more fluid, such as Brazil, Turkey, and parts of South and Southeast Asia, where emerging parties have increasingly positioned themselves in ideological contrast to incumbents or elites.

In addition to the foundational work on ideological polarisation in party settings, recent work by Draca and Schwarz (2024) provides new evidence that ideological polarisation can also be rooted in the political orientations of citizens. Drawing on decades of data from the World Values Survey and the European Values Study, the authors employ unsupervised machine learning to classify individuals into latent ideological types based on their economic preferences, social attitudes, and views about core institutions. This approach reveals not only distinct ideological types, but also that the centrist groups are declining over time.

Issue Polarisation

Issue polarisation refers to the growing divergence in opinions on specific policy issues, where individuals increasingly align their positions with those of their political or ideological group. Rather than just generally holding different opinions, people become more consistently opposed on a particular issue with less room for moderate views. This form of polarisation can often occur regardless of individuals’ broader ideological framework, especially when that issue is strategically politicised by elites or linked to social identity. Measurement of issue polarisation typically uses opinion surveys or voting behaviour in single-issue elections or referendums.

Conceptually, issue polarisation tends to refer to more than a simple difference in ideology or divergence in attitudes. Issue polarisation is often conceptualised as having a strategic basis, particularly when parties or politicians disproportionately amplify certain policy issues precisely to create public cleavages. In De Sio and Weber (2014)’s “issue yield” framework, parties focus on high-impact issues precisely to shift public alignment, effectively fostering polarisation.

This is especially the case when issue polarisation occurs in political contexts without strong ideological cleavages or institutionalised political parties. In these settings, issue polarisation often reflects efforts to exploit issues with group-based political salience. Politicians may exploit these divides to create ‘wedge issues’, generating polarisation even in systems where ideological debate is moderate or largely absent.

Affective Polarisation

Affective polarisation reflects the extent to which individuals feel animosity or distrust towards people, media sources, or political groups that hold opposing political views (Iyengar et al. 2019). Unlike ideological or issue-based polarisation, affective polarisation is less about substantive disagreement and more about how emotional and social cleavages can create an ‘us vs them’ dynamic in politics. At its most extreme, affective polarisation can lead to avoidance, dehumanisation, or even repression of political opponents.

Scholars increasingly conceptualise affective polarisation as a form of partisan social identity, where partisans are seen not only as political rivals, but fundamental threats to their values (Iyengar et al. 2019). Measurement of affective polarisation is typically measured using gaps in feeling thermometers or measures of out-group versus in-group trust. Applied to attitudes about elites, for example, Reiljan et al. (2024) use standardised cross-national indices to measure affective polarisation towards parties and leaders.

Affective polarisation is particularly relevant in settings where partisanship overlaps with social identities – ethnicity, religion, or linguistic groups – and where histories of conflict or repression can reinforce group-based distrust. However, even in contexts without a history of longstanding divides, elites can foster affective polarisation through aggressive rhetoric, elite cues, and strategic framing to foster political hostility and undermine norms of tolerance. Misperceptions about out-group demographics and values can exacerbate these divisions: voters tend to overestimate the extremism of opposing groups, for example, a bias that can contribute to polarisation even when ideological differences are minimal (Ahler and Sood 2018).

Who is Polarised? Polarisation Across Political Actors

Political polarisation can manifest differently depending on who is becoming polarised. While the preceding sections distinguished between ideological, issue-based, and affective forms of polarisation, this section focuses on the political actors through which polarisation operates – namely, individual citizens, political elites, and media institutions.

Understanding who is polarised is not just a matter of classification: it is essential for identifying causal mechanisms and designing interventions. For example, affective polarisation among voters may be rooted in elite rhetoric, but it may also reflect long-standing identity cleavages. Similarly, elite polarisation may reflect deliberate strategic positioning rather than shifting personal beliefs. Media polarisation, meanwhile, can both reflect and intensify these dynamics, shaping how information is framed and how audiences interpret it.

At the individual level, polarisation is not limited to emotional partisanship. While much of the literature has focused on affective polarisation, recent research has expanded to include social sorting, selective trust in information, and asymmetric beliefs about policy consequences. For instance, individuals’ economic expectations have been shown to diverge dramatically depending on which party is in power, with partisans projecting highly optimistic or pessimistic futures depending on electoral outcomes (Coibion et al. 2020). In many contexts, these divides overlap with ethnicity, religion, or region, reinforcing polarisation even in the absence of conflict over citizens' preferred policies.

Among elites, polarisation can be both a reflection of voter preferences and a strategic response to political incentives. Formal theory highlights how candidates may polarise strategically: Callander and Carbajal (2022) develop a dynamic model in which elites move away from the centre to attract voters, who in turn shift their beliefs in response. Empirically, Gentzkow et al. (2019) show that Congressional speech in the US has become increasingly predictive of party affiliation, with rhetorical shifts outpacing changes in policy substance. This evidence suggests that polarisation is often embedded in elite communication styles and institutional incentives – not just ideological drift.

Just like elites, media institutions play a dual role: they both reflect polarisation and help shape its trajectory. Ideologically aligned media outlets and algorithmically curated platforms increase the likelihood of selective exposure to like-minded content. Thus, the more polarised people are, the more these media dynamics polarise them, which reinforces partisan identities and reduces trust in alternative sources. 

These patterns suggest that polarisation is a multi-actor phenomenon, with implications for citizens, elites, and media ecosystems. Disaggregating actor-level polarisation clarifies its origins, how it spreads across institutions and information systems, and which levers are most effective for intervention.

For full reference list see the end of the conclusion chapter.