Abstract
Research on echo chambers has produced mixed support for the “echo chamber hypothesis,” partly because this complex phenomenon cannot be reduced to a single testable claim. To move the field forward, it is necessary to examine its internal dynamics. Building on a teleological reading of prior research, this article theorizes an integrative framework that distinguishes between endogenous media aspects and exogenous meddling in social media’s information flows. This analytical move sharpens the boundaries between echo chambers and the social dynamics of online communities, and accounts for the amplification of polarizing political narratives. The outcome is a socio-technical framework divided into three stages: ignition, where provocative content triggers an insular community; amplification, where algorithmic and networked distribution accelerates reach; stabilization, where institutions normalize the narrative in a broader audience. It concludes with premises for future research linking echo chambers to hybrid threats, disinformation, and monetization dynamics in the creator economy.
Keywords
Introduction
“We are waging a 21st-century information warfare campaign against the left, and they’re using tactics from the 1990s,” Watters said, “What you’re seeing on the right is asymmetrical. It is like grassroots guerrilla warfare. Someone says something on social media, Musk retweets it, Rogan podcasts it, Fox broadcasts it, and by the time it reaches everybody, millions of people have seen it. It is free money.”
The echo chamber concept has different meanings in policy circles and in the academic literature. In academic research, it refers to the selective exposure to information that confirms preconceived views (Garrett, 2009), usually operationalized through algorithmic filters (Berman and Katona, 2020; Pariser, 2011). In policy circles, however, echo chambers refer to how reactionary, polarizing narratives gain traction on social media and enter mainstream discourse. For example, the quote above presents a situation in which social media does not just serve information people already believe, at least not initially; instead, the argument is about the amplification, not repetition, of fringe polarizing ideas, and about how they become accepted by the public. These two interpretations differ significantly, limiting our capacity to generate effective public policy.
Empirical research on echo chambers remains contested, with some researchers even labeling it “a distraction” (Garrett, 2017: 370), or questioning whether echo chambers exist (Dubois and Blank, 2018; Flaxman et al., 2016; Nechushtai and Lewis, 2019). Part of the reason is that previous research has condensed the phenomenon into a single “echo chamber hypothesis” (Hartmann et al., 2025; Mahmoudi et al., 2024; Terren and Borge-Bravo, 2021), which means rejecting it dismisses the entire phenomenon.
The literature is fragmented into multiple streams. One draws from acoustic echo chambers: a delayed return of sound perceived as a repetition. Under this analogy, people consume (and reverberate) ideas they already believe (Vaccari, 2013). This view entails three partially distinct mechanisms: first, people like to gather with like-minded people (homophily, McPherson et al., 2001). Second, they tend to select information that confirms preconceived views (selective exposure, Garrett, 2009). Third, the Internet curates what users consume, excluding some voices (filter bubbles, Pariser, 2011). Although some studies explore each mechanism in isolation, researchers tend to aggregate these mechanisms without accounting for their dynamics. See Table 1 for a glossary.
Glossary.
The acoustic analogy has critics. For Nguyen (2020), rather than a problem with information flow, the distinctive feature of echo chambers is the creation of trust asymmetries with outsiders. Drawing from partisan media, a second stream conceptualizes echo chambers through the efforts to discredit an out-group as enemies (Jamieson and Cappella, 2008). According to Cinelli et al. (2021), the key effect is the increased polarization of society (c.f., Boulianne et al., 2020; Grusauskaite et al., 2024; Karlsen et al., 2017).
A third stream situates echo chambers in reactionary politics and the growing circulation of fringe anti-democratic sentiment (Lewis, 2018, 2020). Table 2 situates research in the context of how supremacist (Törnberg and Törnberg, 2024) and anti-scientific views (Diaz Ruiz and Nilsson, 2023) enter mainstream discourse (Massanari, 2017). The theoretical gap is that neither homophily nor selective exposure accounts for the growing acceptance of fringe and extremist political positions.
Examples of echo chambers in fringe extremist politics.
Recent systematic literature reviews (SLRs) have examined echo chambers (Table 3), and they all find mixed support for the echo chamber hypothesis, which claims that communication environments featuring selective exposure, homophily, and algorithmic curation disproportionately circulate attitude-consistent information that amplifies perceived validity, altogether intensifying societal polarization. To advance the field, we need to understand why research on echo chambers yields inconsistent findings.
Recent review articles on echo chambers.
The gap is that prior SLR studies did not interrogate conceptually whether the literature is studying a phenomenon that can be captured in a single hypothesis. Instead, previous research has treated “echo chambers” as a monolithic entity, bundling distinct mechanisms into a unitary construct that lacks internal dynamics. The ensuing compression may help explain why studies yield mixed results: they may be testing different dynamics lumped under the same label.
The purpose of this article is to develop a set of fundamental premises that reconcile the contradictions and tensions in the literature on social media echo chambers, thereby generating integrative guidelines for future research. This study has two research questions:
RQ1. What is the nature of the echo chamber effect and the underlying teleological assumptions adopted in the literature?
RQ2. What are the applicable common elements of echo chambers that can work together in a conceptual framework?
This article advances the literature by treating echo chambers as a complex phenomenon with internal dynamics operationalized in five socio-technical dimensions. First, social silos refer to how people create insular social epistemes. Second, selective exposure explores the tendency to seek confirming sources and avoid disconfirming ones. Third, filter bubbles refer to how algorithmic recommender systems restrict information flows. Fourth, strategic provocation involves creating trust asymmetries. Fifth, synthetic manipulation explores exogenous meddling with the informational flow of social media through inauthentic accounts and/or AI.
Research approach
This investigation mobilizes a “meta-narrative” review (Greenhalgh et al., 2005) of the fragmented literature in three phases: (1) an SLR identifies the relevant body of work across multidisciplinary fields: social sciences, computer sciences, humanities, psychology, business and management, and economics; (2) organizing according to their teleological assumptions; and (3) developing an integrative framework.
Phase 1: mapping out relevant research on echo chambers
The study uses the SLR method, which is a systematic approach to reviewing a large corpus of literature with a transparent protocol (PRISMA) for identifying and screening prior studies (Page et al., 2021). Figure 1 presents the standard flowchart, but some decisions require explanation. The first is that the Scopus database was selected because it is a reliable source for multidisciplinary queries (Gusenbauer and Haddaway, 2020: 208). The second is setting the starting date in 2005 to time it with algorithmic media. A third decision was setting up a minimum number of citations (30). The reason is that, over the 2020–2025 period, a large number of articles have been published (n = 1219), but the term “echo chamber” is often used in passing. A second query included other relevant papers published between 2024 and 2025 (n = 132). Supplemental Annex 1 has more details on the method, and Supplemental Annex 2 lists the 288 research papers ultimately included in the corpus.
SLR flow diagram.
Phase 2: a teleological analysis
SLRs were developed in medicine and healthcare, often featuring descriptive and quantitative synthesis of clinical trials (Page et al., 2021). However, word counts are unhelpful here because prior research uses a single label to represent multiple understandings of what echo chambers do: filtering out information, gathering like-minded people, or driving social polarization. It also includes multiple mechanisms of how they work: selectively circulating content that fits pre-existing worldviews, fomenting trust asymmetries, or filtering out discordant opinions. In the social sciences, a “meta-theoretical” approach is a way to make sense of mixed findings, helping researchers to theorize about theories (Becker and Jaakkola, 2020; Greenhalgh et al., 2005).
SLRs use a keyword search to create a relevant corpus, which then can be analyzed using multiple techniques. Here, the author conducted a teleological analysis by reading the relevant passages from the 288 articles, meaning analyzing phenomena in terms of their purpose and the means by which they pursue it.
Teleology derives from the Greek telos (purpose or end goal) and logos (explanation or reason). From its etymology, teleology means an explanation that inquires about an entity’s purpose: what it is for (its purpose). It is “characterized by the use of the words ‘function’, ‘purpose’, and ‘goal’, as well as by statements that something exists or is done ‘in order to’” (Mayr, 1974: 134). As an analytical tool, Wright (1976) proposed the formal study of means-ends relations: the function X produces the outcome Y, and producing Y helps explain why X persists. In other words, X (function) is there because it reliably produces Y (outcome).
In practical terms, when encountering a relevant passage, the author examined the expected outcome (Y) and the means or functions (X) that reliably produce it.
For example, Marks et al. (2019: 74) write that “people prefer to consult and learn from those whose political views are similar to their own, thus creating a risk of echo chambers or information cocoons.” (. . .) “This is partly because people assume that like-minded people are more likely to be correct– a phenomenon that can lead to echo chambers.” In this passage, the outcome (information cocoons/epistemic bubbles) is reliably produced by a function (preference for like-minded people), and this function likely explains the outcome that echo chambers generate (social silos).
In another example, Müller and Schwarz (2021: 2131) write, “This paper investigates the link between social media and hate crime. (. . .) Consistent with a role for ‘echo chambers,’ we find that right-wing social media posts contain narrower and more loaded content than news reports.” This passage frames an outcome (hate crime) that is reliably produced by the means (loaded/incendiary posts), explaining what echo chambers produce (polarization). Table 4 summarizes the distinctive teleological themes, and Table 8 (in Supplemental Annex 1) provides additional examples.
Teleological categorization of the echo chamber literature.
Wright (1976): X is there because it produced Y, and producing Y helps explain why X persists.
The categories in Table 4 suggest that the means and purposes are always aligned, even though this was not always the case. In some passages, there are misalignments between functions and outcomes. Del Vicario et al. (2017: 6) write, Selective exposure and confirmation bias, indeed, have been shown to play a pivotal role in content consumption and information spreading. Users tend to select information that adheres to (and reinforces) their worldview and to ignore dissenting information. This pattern elicits the formation of polarized groups – i.e., echo chambers – where the interaction with like-minded people might even reinforce polarization.
This passage contains a teleological contradiction: consuming information that aligns with pre-existing views (selective exposure) does not necessarily lead to an adversarial attitude toward another group (societal polarization). In some cases, these misalignments can be reconciled by considering echo chambers not as a single phenomenon but as a staged process: first, groups form self-reinforcing views, and then, in a following step, polarization occurs.
To overcome the limitations of a single analyst, the author organized a half-day research workshop with another researcher to refine the teleological categories against exemplary papers. Also, the author assigned these papers as coursework readings, asking students to identify themes independently and discuss them in class. See Supplemental Annex 1 for more details on the teleological analysis.
Phase 3: theorizing an integrative conceptual framework
The outcome of the teleological analysis is hard to quantify because there are occasions in which authors write a passage, citing a well-established definition (e.g., selective exposure, Garrett, 2009) even when the argument is about something else, such as societal polarization (Cinelli et al., 2021) or social media manipulation tactics (Ferrara et al., 2020). To give a sense of quantity, in the first batch of influential papers (citations n > 30), the most common themes are that echo chambers are media environments that selectively serve content that fits with pre-existing views (73%) and/or social environments in which people socialize with like-minded people (64%). In addition, researchers mention algorithmic recommendations or filter bubbles (51%). Researchers often pair two themes: selective exposure paired with either homophily or filter bubbles.
In the second batch of papers published in 2024–2025, passages are linked to polarization (53%) and digital media platforms (42%), and exogenous manipulation (e.g., misinformation) (17%). What changed is that whereas previous researchers discussed echo chambers as a neutral media phenomenon, more recent papers adopt an affective stance, viewing them as negative for society.
After identifying high-order themes through the teleological analysis, the outcome is a visual model and a research agenda.
Socio-technical dimensions of the echo chamber concept
Based on the teleological analysis, Table 5 summarizes echo chambers through five high-order thematic clusters: social silos, selective exposure, filter bubbles, strategic provocation, and synthetic manipulation.
Thematic classification of echo chambers.
Theme 1: social silos
Echo chambers reflect a human tendency to gather with similar people. In the early 2000s, Cass Sunstein predicted that the increasingly digital nature of the public sphere would lead individuals to naturally gravitate toward people with similar values and preferences, while ignoring voices that challenge their worldview (Sunstein, 2001). Over time, people converge on a social episteme, which is a shared, historically situated set of assumptions about what counts as knowledge within a group (Godler et al., 2020), and the shared informational environment that determines what can be known and by whom.
Social epistemes can harden into social silos when repetition and reward structures filter out critiques (Schwarzenegger, 2020). In other words, a social episteme can become insular when its norms for assessing the validity of knowledge actively exclude dissenting voices.
Teleological assumptions and explanatory limitations
An echo chamber can be understood as a self-selected social episteme that hardens into a social silo. However, Nguyen (2020) argued that a self-reinforcing informational environment is not an echo chamber, but instead refers to a normal human tendency to avoid alternative viewpoints through omission. Therefore, notions such as homophily explain echo chambers by focusing on the social environment in which ideas circulate, but do not explain how they harden into insularity.
Moreover, insularity does not necessarily lead to aggressiveness toward outsiders (Hobolt et al., 2023). If we were dealing primarily with groups of like-minded people, we would expect a preference for insiders and perhaps ignorance of outsiders. However, arguments on social media often feature animosity and controversy, explicitly directed toward the out-group, to the point that discussions resemble “trench warfare” rather than dialogue (Karlsen et al., 2017).
Theme 2: selective exposure
This view frames echo chambers as environments in which users disproportionately choose information that reaffirms their preconceived views, sustaining prior commitments, and producing a self-reinforcing sense that one’s position is epistemically dominant (Dvir-Gvirsman et al., 2016; Nelson and Webster, 2017). The selective exposure lens emphasizes individual choice in directing attention, and thus, people seek out and attend to information that aligns with their prior attitudes while avoiding dissonant content, especially on political issues. On social media, users repeatedly choose like-minded sources and social links, while avoiding disconfirming views (Parmelee and Roman, 2020).
Teleological assumptions and explanatory limitations
The selective exposure view has mixed empirical support (Dubois and Blank, 2018; Flaxman et al., 2016). Although people do seem to prefer reading confirming news sources and avoid disconfirming ones (Parmelee and Roman, 2020), it is unclear whether individual preferences drive the echo chamber phenomenon. Moreover, the theme has merged with the algorithmic recommender systems of social media, which can be conceptualized as a filter bubble.
Theme 3: filter bubbles
The restrictions on information flow produced and maintained by the digital infrastructures of the Internet result in echo chambers. This view is an extension of the selective exposure argument that focuses on a specific culprit: algorithmic recommender systems (Fletcher et al., 2023), the digital infrastructure of search engines and social media that curates user feeds according to the content users are most likely to engage with. The resulting “filter bubbles” hide content from users (Pariser, 2011).
Researchers have analyzed whether the Internet actually filters or restricts information flows (Bruns, 2019; Haim et al., 2018; Kitchens et al., 2020), with mixed results (Garrett, 2017). For instance, researchers have investigated whether algorithms serve biased news sources that emphasize only one side of the political spectrum (Nelson and Webster, 2017). Overall, the filter bubble argument appears overstated (Dubois and Blank, 2018): search engines and recommender systems appear to expose users to a variety of news media sources (Flaxman et al., 2016; Fletcher and Nielsen, 2018; Fletcher et al., 2023), suggesting that information flow is not restricted in the way that Pariser (2011) predicted.
Teleological assumptions and explanatory limitations
This stream assumes that the primary effect or function of echo chambers is to restrict the flow of information (Nguyen, 2020), limiting exposure to differing perspectives (Rhodes, 2022), via a process curated by digital platforms (Bruns, 2019; Pariser, 2011). The issue is that, if the evidence for filter bubbles is scarce, the alternative explanation is that social media reflects pre-existing social dynamics of an already polarized society.
One limitation is that the filter bubbles argument does not explain how social media creates conditions for the amplification of fringe beliefs, from niche groups to mainstream audiences, as in the cases of Table 2. If gatekeepers restrict information flow and circulate content that fits pre-existing views, this should lead to homogenization, not fragmentation. Instead, we see the growing circulation of fringe views and the growing animosity toward differing perspectives.
Theme 4: strategic provocation
Understanding echo chambers solely as an information-flow problem does not allow us to examine how they are designed to discredit opponents and silence critics. For Nguyen (2020: 2), “echo chambers are structures of strategic discrediting, rather than bad informational connectivity.” Moreover, Diaz Ruiz and Nilsson (2023) argued that echo chambers can emerge through the communication tactics designed to weaponize and amplify controversies against perceived adversaries, thereby creating epistemic distance from the out-group. For Karlsen et al. (2017), echo chambers do not merely reflect pre-existing positions; they offer a space for overtly and aggressively challenging people considered to be in the out-group (Grusauskaite et al., 2024).
This interpretation of echo chambers predates the Internet era and dates back to partisan talk radio from the 1990s (Jamieson and Cappella, 2008). In American conservative politics, talk radio started framing liberals as an existential threat: it follows that if conservatives are patriots and liberals are globalists, then liberals must be traitors. The notion of echo chambers describes a situation in which a media figure stirred culture wars, fostering an “us against them” worldview to keep the audience hooked. This formula has now been adopted for social media by reactionary politics (Lewis, 2018).
Provocation in the creator economy is a business model that fuels controversy and outrage into attention; hence, profit. For Ulver (2022), a “conflict market” is when marketers deliberately orchestrate and escalate identity-led conflicts among consumers to capture attention and thus monetize polarization by exploiting the platform’s recommender systems that amplify viral content. Whereas echo chambers existed before social media, the creator economy monetizes them (Diaz Ruiz, 2025b).
Teleological assumptions and explanatory limitations
Studying echo chambers through a provocation lens reveals that they are built and maintained to create a self-reinforcing attention machine: it earns visibility to attract engagement, which can be monetized, incentivizing even more provocation. Participants circulate claims with a dual purpose: to create the impression of a community under siege and to mobilize their audience against perceived opponents. Provocation is strategic when it is designed to exploit platforms’ algorithms, which use high-arousal reactions as proxies to identify content that the algorithms reward with extra reach.
However, one limitation is that provocation is neither new nor specific to social media. Moreover, it over-attributes agency to the provocateur, overlooking the role of institutional actors (e.g., political parties) whose participation normalizes fringe views, extending them to comparatively moderate offline communities.
Theme 5: synthetic manipulation
Whereas pre-algorithmic media have long been subject to various forms of manipulation, the combination of decentralized communication and monetized user-generated content creates opportunities for exogenous influence operations that target the particularities of social media. In the discussion concerning security and democracy, echo chambers have been described as a medium through which malicious actors spread disinformation (Krafft and Donovan, 2020) and conduct hybrid warfare (European Commission, 2025).
These campaigns exploit some of the vulnerabilities of social media, especially its overreliance on engagement metrics to recommend content (Diaz Ruiz, 2025b), for example, by using inauthentic social media accounts to hype content (Mazza et al., 2022). The term coordinated inauthentic behavior (CIB) describes networks of accounts that conceal shared control while pursuing the same messaging (Cinelli et al., 2022). These accounts are now powered by synthetic media (Vaccari and Chadwick, 2020), AI-generated content, including voice, text, and video. CIB can be used for astroturfing (Keller et al., 2020), the practice of disguising a public relations campaign as if it emerged from spontaneous “grassroots” public support. Even when users successfully identify fake accounts, the sheer volume of posts floods the space, drowning out dissent (Porup, 2015).
Teleological assumptions and explanatory limitations
We do not know enough about the exogenous attempt to meddle with social media information flows. Social media users can encounter countless fake accounts that parrot similar messages, often featuring polarizing narratives (Arceneaux et al., 2026; O’Shaughnessy, 2020). Future researchers can study how echo chambers amplify reactionary topics through synthetic means, including deepfakes and conversational AI. The limitation is that, even though CIB drives social engagement by nudging algorithms to boost the desired narrative, this is insufficient to explain social polarization without accounting for the legitimizing role of institutional actors.
Echo chamber dynamics: an integrative framework
A remaining issue is that each dimension remains disconnected: Filter bubbles addresses how algorithms restrict information flows, but does not explain aggressiveness among communities. Social silos explain insular thinking, but not polarizing attitudes. Strategic provocation emphasizes trust asymmetries in the creator economy, but does not explain its normalization. The next step is to place the themes on the same plane to determine whether the elements that appear contradictory are compatible when studied within a dynamic process (Table 6).
Echo chamber dynamics on social media: A three-stage process.
What appear to be inconsistencies can be viewed as part of a dynamic process that overcomes the limitations of each approach in isolation, placing each dimension into the “big picture.” Building upon the model for echo chamber amplification in Diaz Ruiz and Nilsson (2023: 30), Figure 2 proposes a three-stage dynamics process.
Echo chamber dynamics on social media.
The dynamic aspect of the model lies in how four premises enable echo chambers to move from one stage to another. P1 (insular thinking and epistemic isolation) holds that isolated communities within close-knit social epistemes are more susceptible to becoming echo chambers that circulate both confirming and disconfirming content. P2 (strategic provocation in the attention economy) explores the type of provocative (incendiary) content that stirs echo chamber dynamics, and whether it is designed and formatted to go viral through algorithmic recommendations in the creator economy of social media. P3 (synthetic and organic amplification) examines whether amplification is driven by people’s spontaneous interest and/or influence campaigns that exploit the vulnerabilities of social media via CIB or conversational AI tools. P4 (institutional normalization) holds that online echo chambers spill over into offline moderate communities through institutional actors who normalize fringe or reactionary views into the mainstream.
Stage 1: ignition
The first unresolved tension in the literature is whether social media echo chambers are primarily products of platform algorithms or extend pre-existing social dynamics. To resolve this confusion, this article proposes that the echo chamber captures both the human tendency toward selective affiliation and also that the ensuing insularity is susceptible to provocation via engagement-optimized algorithmic recommendations.
Both aspects converge in the ignition stage of an echo chamber, when a relatively insular community encounters a triggering narrative packaged with moral-emotional or provocative framing. Like kindling in a fire, provocation ignites the early bursts of engagement that algorithms upscale, increasing visibility. The ignition analogy fits the starting point of an echo chamber, since provocative content that goes viral is often called “incendiary,” and it is “angry by design” (de Roos et al., 2024; Munn, 2020). Understanding this stage requires knowledge of how pre-existing social silos operate and also how strategic provocation elicits reactions from their participants.
Insular thinking and epistemic isolation (P1)
One conceptual confusion in the literature relates to the nature of the content that echo chambers circulate. Building on the analogy of acoustic echo chambers, the selective exposure stream has assumed that echo chambers serve congruent content that aligns with pre-existing assumptions (Flaxman et al., 2016; Garrett, 2017; Haim et al., 2018). However, an alternative view argues that echo chambers circulate disconfirming content to stir up controversy (Diaz Ruiz and Nilsson, 2023; Karlsen et al., 2017).
The ignition stage is a point of interaction for both types of content. On the one hand, as predicted by Sunstein (2001), people gravitate toward communities around shared interests, which tend to circulate congruent content. On social media, this takes the form of posts and memes that express locally accepted ideas, which can lead to insular thinking. Sunstein (2017) later argued that social media’s personalization tools lead to a fragmented democracy and that democracy needs diverse perspectives.
The issue is that encountering diverse perspectives on social media is rarely neutral or objective; instead, it is often colored by provocation. We know that the content that thrives on social media often stirs anxiety, prompting users to click on links, watch videos, and post comments, increasing the financial value of user engagement (Berger and Milkman, 2012). Therefore, we expect that insular social epistemes also circulate disconfirming content designed to elicit an emotionally charged response (Buder and Said, 2025).
Implications of Premises 1a and 1b for future research
If insular epistemic communities circulate both congruent and disconfirming content (1a) yet are susceptible to provocative disconfirmation (1b), future research should not limit itself to the selective exposure to confirming news sources (congruent content). Instead, it can focus on both understanding the pre-existing dynamics of a particular social episteme and identifying the types of content that elicit engagement (disconfirming content). The literature often operationalizes echo chambers through news sources. And yet, users get their news on social media, where influencers comment on the news and users circulate memes. Moreover, disconfirming content can take many forms on social media (Buder and Said, 2025), from objective corrections (fact-checking) to provocation (trolling or ragebait).
Strategic provocation in the attention economy (P2)
Since algorithmic growth hinges on rapid engagement, it is crucial to understand how creators monetize controversy and reward identity-led conflicts (Husemann et al., 2015; Ulver, 2022). Future research should contextualize the echo chamber phenomenon within the creator economy of social media, including its commercial incentives to overemphasize conflict and controversy.
Implications of Premises 2a and 2b for future research
Premise 2 guides future researchers to study echo chambers as part and parcel of the business model of social media and the creator economy. Without a creator-economy lens, the ignition stage relies on spontaneous outbursts, when, instead, it is likely a product of a system designed to compound attention. Cross-fertilization between media researchers and business scholars could broaden the research agenda by drawing on studies of monetization in the creator economy and testing whether revenue structures reward incendiary narratives.
Stage 2: amplification
Extant research on echo chambers typically studies relatively stable systems. However, we do not know much about the main issue discussed within policy circles, namely, the rapid circulation of fringe, reactionary beliefs. The gap is about how echo chambers expand their reach and attract new participants.
Some studies have explored the type of cultural conflicts that attract participants to debate (de Roos et al., 2024; Karlsen et al., 2017; Ulver, 2022), and thus, we know that echo chambers spread “adversarial narratives” designed to stir “identity-driven controversies” (Diaz Ruiz and Nilsson, 2023: 18), monetizing conflict (Ulver, 2022). In turn, creators routinize controversy as their growth engine. And yet, provocateurs do not act alone; they rely on a network of linked accounts to hype content. In some cases, these accounts are operated by humans, but not always. Fans and partisan outlets may have highly committed users, but fake accounts can simulate organic interactions, inflating the metrics that platforms use to identify engaging content. The feedback loop that amplifies reach can feature a mix of synthetic and organic hype.
Implications of Premises 3a, 3b, and 3c for future research
Researchers should study the socio-technical dynamics that hype content and amplify the reach of echo chambers on social media. Some of these accounts represent highly engaged users (e.g., activists), but sometimes, networks of fake accounts are enabling a new form of astroturfing in which social bots simulate human conversations online, but when in reality it means “bots talking to bots” hyping content (Diaz Ruiz, 2025a). One avenue for future research is to identify the exogenous factors that accelerate the reach of echo chambers and spill over to offline audiences.
Stage 3: stabilization
The stabilization stage begins when the echo chamber no longer requires constant provocation, as it is once again a social episteme, though one that spills over into moderate offline communities. More research is required to address the disconnect between fringe narratives circulating online and their normalization among moderate offline audiences.
Building upon the literature on conspiracy theories (Venturini, 2022), institutional distrust is likely the binding theme. Conspiracy theories usually portray democratic institutions as adversaries, casting them as part of a cabal that includes journalists and academics, pre-emptively discarding disconfirming voices (our sources are authentic; their sources are corrupt). Institutional actors can help normalize the enclave by circulating their beliefs to more moderate communities: political parties and partisan organizations extend the echo chamber, conferring durable visibility and credibility.
Implications of Premises 4a and 4b for future research
The normalization of discourse is a common theme in media research, but this research is not actively connected to echo chamber dynamics. Potential future research can examine the role of institutional actors in laundering fringe narratives, giving them mainstream visibility through coverage and even refutation.
Conclusion
This article distinguishes between (1) the internal dynamics of echo chambers as a media phenomenon and (2) inorganic meddling into the flows of information on social media. Doing so sharpens the conceptual boundaries between echo chambers and other social dynamics, such as online communities. It also distinguishes between a neutral social phenomenon and the exogenous manipulation of social media to amplify fringe extremist political narratives. Table 7 presents guidelines for research.
Premises of the echo chamber effect on social media: an integrative framework.
Echo chambers have internal dynamics
A three-stage process represents the internal dynamics of echo chambers (Figure 2). An insular community can be triggered with provocative content, (1) igniting echo chamber dynamics; exploiting algorithmic recommender systems and the financial incentives of the creator economy to (2) amplify and accelerate its distribution; and how its institutional normalization (3) stabilizes it into an enlarged social silo echo, extending its reach into offline communities. By studying its dynamics, researchers can refine parts of the model without dismissing the phenomenon entirely.
Inorganic amplification of information flows
The concept of echo chambers has emerged as a site for research into influence operations and disinformation campaigns that meddle with social media’s information flows, a growing threat to democratic societies (European Centre of Excellence for Countering Hybrid Threats, 2023). Returning to Watters (2025) quote on information warfare, more research is needed to understand the spillover from fringe online groups to mainstream attention. Future research should examine how malicious actors stir debate, flooding social media with AI-generated content.
Finally, the commercial nature of social media means that scholars should also study the business model of echo chambers. One sub-theme in this study is about how strategic provocation elicits responses, abusing the creator economy (Ulver, 2022). One remaining gap is that research on echo chambers rarely addresses how influencers make money (monetization). Future research should explore echo chambers at the intersection of influence operations, disinformation, and the creator economy.
Supplemental Material
sj-docx-1-nms-10.1177_14614448261441872 – Supplemental material for Echo chamber dynamics on social media
Supplemental material, sj-docx-1-nms-10.1177_14614448261441872 for Echo chamber dynamics on social media by Carlos Diaz Ruiz in New Media & Society
Supplemental Material
sj-docx-2-nms-10.1177_14614448261441872 – Supplemental material for Echo chamber dynamics on social media
Supplemental material, sj-docx-2-nms-10.1177_14614448261441872 for Echo chamber dynamics on social media by Carlos Diaz Ruiz in New Media & Society
Footnotes
Acknowledgements
Early versions of this manuscript received feedback from participants in the workshop “Addressing the Complexities of Hybrid Threats” hosted by CENS (Center of Excellence on National Security) at Nanyang Technological University, Singapore. I want to thank my colleagues in the CCT community (Consumer Culture Theory) and at CERS (Center for Relationship Marketing and Service Management) for their constructive feedback during research seminars and presentations.
Ethical considerations
This conceptual work does not require special ethical approval.
Informed consent
This conceptual work does not require informed consent.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
Declaration of conflicting interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data availability statement
No empirical data were generated for this project.
Supplemental material
Supplemental material for this article is available online.
Author biography
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
