It's far easier online to find and follow people who have the same opinion as you, and block out people with views that you disagree with. And before you know it, you're sucked into an echo chamber with only news that confirms your views.
This echo chamber concept is cited time and time again in mainstream discussions, so much so that communication researchers argue that it has created “its own discursive reality”[1]. In fact, there is very little empirical support for the filter bubble and echo chamber concepts as a population-wide phenomenon. Social media users have been found to encounter more diverse news than non-users. And even when we think about people holding extremist views, findings suggest them to be even more engaged with the mainstream news compared to other news users—after all, they are interested to know what “the enemy” might be up to.
If you’re interested, the article below gives a great overview on the origin of this concept and different disciplines’ view on the subject (it’s open-access as well).
This is amusingly visible in lots of discussion groups (including HN!), where you will find large numbers of participants arguing that the group itself is an echo chamber—but each of them arguing that it is an echo chamber with a very different viewpoint, usually diametrically opposed to the person making the argument.
I don't think that social media users encounter less diverse news, it's that they encounter news that's specifically engineered for distribution via social media. This is not always written from a particular point on any political spectrum, but is almost always written with strong appeal to emotion, a deficiency of nuance, and a deeply dehumanizing view of the "other."
But filter bubble (FB) and echo chamber (EC) are two different concepts. They are not synonymous. The former is passive, and you have no choice. The latter is human nature. Birds of a feather and such. It's hard to imagine human nature reversing eons of evolution.
> "..indeed, search and social media users generally appear to encounter a highly centrist media diet that is, if anything, more diverse than that of non-users..."
"Appear to encounter" is a vague. Thst aside, encounter doesn't mean you're open to those ideas. Imagine seeing something in your feed that is close to a polar opposite of your view. The majority of the time that is interpreted as "That's BS. See. I knew I was right." That is, few have the time to pause and ask "what if?" or "did I miss something?" The exposure doesn't open their mind, it backfires and pulls the blinders even more narrow. I missed where this study addresses that boomerang effect.
Who are non-users? Fox News watchers? Or Rachel Maddow devotes? That's a pretty low bar. More diversity than than isn't saying much? Also, diversity in sources isn't necessarily diversity in ideas and perspectives. Most media sources are so worried about readers/viewers and ad revenue that they dare not venture too far from the prevailing narrative. Add in cancel-culture and an endless list of triggers and the reporting deviates further and further from traditional journalisms. Yes, content is delivers, how often would it qualify as intellectual junkfood?
> "Nechushtai and Lewis reported that Google News searchers in the US were directed overwhelmingly to 'longstanding national newspapers and television networks'"
That might be true, but it's not proof that FB or EC don't exist. I don't want to get off topic but most of the major news sources suffer from mono-vision, narrative preservation, parroting "conventional wisdow", and so on. The lack of depth and breadth and the sources is and continues to be a problem.
Furthermore, getting results in a SERP does not mean there was a click-through to read. Again, "exposure" isn't necessarily leading to more diverse ideas.
> "...fragmentation of national mediaspheres..."
IDK. From multiple sources I've read that limited ownership is leading to consolidation. With what data do they make this assumption? Most seem to agree that consolidation of ownership has led to lens that zoom in and stay tightly on message, as opposed to zoom out and look for possible other insights.
---
I have to read more later. Hopefully it gets better. I'm not so sure this holds enough water for me. Yes, that's my interpretation. It will be interesting to see how this pov is treated here. Ironic?
You raise some points that were also raised by a sibling commenter, but I’ll try to reply to them. You are correct that FB and EC are different concepts that are unfortunately often used interchangeably (even in research). This stems from the fact that there was no robust definition offered when the concept was introduced. Bruns offers a definition to distinguish the concepts.
> The former is passive, and you have no choice. The latter is human nature. Birds of a feather and such.
Both concepts are closely related to the idea of homophily, the preference to connect or communicate with like-minded individuals (birds of a feather, as you say). In Bruns’ definition, echo chambers mean the preference to connect, while filter bubbles mean the preference to communicate. Both of them are, as you say, human nature at their core. The concepts go further than homophily, though, in that they also exclude individuals or information outside of the chamber or bubble.
> "Appear to encounter" is a vague.
This is from the research abstract summarising recent empirical findings. I think some critical distance is warranted.
> Thst aside, encounter doesn't mean you're open to those ideas. (…) The exposure doesn't open their mind, it backfires and pulls the blinders even more narrow. I missed where this study addresses that boomerang effect.
This effect is addressed as a “oppositional reading stance” in the section “Social and Political Relevance and Impact”.
> Again, "exposure" isn't necessarily leading to more diverse ideas.
You are correct. However, if individuals are encountering information contrary to their bubble’s or chamber’s belief, we would have to conclude that there is no FB or EC at work as it has failed to exclude this information. If the bubble or chamber was as robust as claimed by Pariser and others, this content would not have reached the individual.
> From multiple sources I've read that limited ownership is leading to consolidation.
You are referring to ownership structures. “Fragmentation of national mediaspheres” here refers to the audience, specifically the concern that the increasing number of media offers, especially with the advent of the internet, would fragment the national audience and have adverse effects on discourse and democracy.
> That might be true, but it's not proof that FB or EC don't exist.
It is not epistemologically possible to prove that these effects do not exist. There is limited evidence that they exist in specific contexts, but barely any evidence for a population-wide effect.
> Social media users have been found to encounter more diverse news than non-users. And even when we think about people holding extremist views, findings suggest them to be even more engaged with the mainstream news compared to other news users—after all, they are interested to know what “the enemy” might be up to.
But that's perfectly compatible with echo chambers. They will interact with the outside world, but their interpretation of things that happen comes from within the echo chamber. Anti-Vaxxers know that there are people believing in vaccinations, they read about them in the papers, but then they might just post a comment "it's what Bill Gates wants you to believe so he can sell more 5G towers to Huawei". The echo chamber does echo their world view and interpretation of events, and just exposing them to "diverse news" won't change that, or we really wouldn't have a large variety of fringe groups.
You raise very valid points. The case you describe would be characterised as “oppositional reading” of mainstream media by those believing in a conspiracy.
However, this is unrelated to the common understanding echo chambers. A big problem with the concept is that there is no robust definition, and the fact that it is used widely in mainstream discussion does not make it easier. The tendency to preferentially connect or communicate with like-minded people (homophily) is nothing new, we encounter it everywhere—think hobby groups, special-interest forums like HN etc. Social media’s affordances especially promote homophily. What the filter bubble/echo chamber concepts add to this preference is an exclusion of outsiders. The above commenter summarised a purported consequence as “you’re sucked into an echo chamber with only news that confirms your views”. In this sense, using diverse news would be incompatible with an echo chamber.
What is especially interesting, then, is that ideological polarisation is very much alive on the internet, despite (or indeed because of) the absence of echo chambers.
I thoroughly recommend the article I linked above. It touches upon these concepts and cites current research that allows you to dig deeper.
It's difficult with academic researchers using terms with significantly different meaning from the public at large. I wonder whether it's more harm- than useful, as it'll typically lead to the public either misunderstanding what they say or dismiss them completely because they're "obviously talking nonsense".
The scientific research into echo chambers seems almost straw-man-y, e.g. "echo chambers are perfectly sealed to the outside, no contact whatsoever is possible", so they check and find "that's not really happening" because people might talk about politics with one group exclusively, but will also talk about sports so they're not in an echo chamber. Does that provide value, when they are in a political echo chamber?
The problem with scientific terminology is widely recognised in communication science. The issue is that while terms used by disciplines such as physics, law or medicine are sufficiently esoteric that someone “on the outside” wouldn’t use their terms casually, so they are somewhat protected from “watering down” their concepts in public discourse—or not, as we are seeing with anti-vax ideology, for instance. Terms in communication science also carry specific meaning to those who know them, but can just as well be interpreted as everyday language by outsiders.
I also get what you mean by “straw-man-y”. I think the issue is that definitions have to be robust and hypotheses falsifiable in order to make them researchable, so they must be kind of stringent. Exactly as you say, someone might be in a political echo chamber, but that is not their entire life. They have peers, family, friends, colleagues that also expose them to political information. By focusing so much on the online, the echo chamber assumption was very limited from the start.
I think it's worse than that. What Twitter especially does is make it easy to find people who are the worst possible version of those you disagree with. Every "nobody REALLY thinks that" is easy to disprove (although still generally accurate!)... You find ACTUAL neoNazis. ACTUAL tankies. ACTUAL fill in the blanks.
That's what radicalizes people. That, and the easy ability to mob folks who don't toe the line on whatever issue. The most disgusting takes on both sides are amplified, not blocked. And even when blocked, still screen-capped. You see the most unreasonable perspectives amplified but the broad middle folk, or those who have legitimate points, either driven off the platform or who just aren't amplified.
EDIT:I know this is usually frowned upon on HN, but I must ask: Why the downvotes?
I'm not sure echo chambers are bad - I like the "warrens and plazas" idea. If ideas really are a market - or more relevantly, policy is a market - we'd never buy anything if we only had half-baked prototypes to choose from.
That said, twitter really makes me appreciate traditional politics.
Party states in their manifesto they value x, y, z
Party representative does something that is consistent, or not, with that (individuals get to decide this, even if the media get to present "an angle").
Some "undesirable group" (Neo-nazis, tankies, etc) may vote for that party, but they don't control the message. Or if they do, the party leader has failed.
Twitter politics, by contrast, is built on the fundamental attribution error. If you don't take the most charitable interpretation of my side, and the worst interpretation of theirs, you're my enemy.
Both sides think they are arguing with the reasonable people on the other side of the debate,but in actuality the extremists are posting a lot more content than most users and they are get amplified as they get more "engagement". Engagement is a terrible measure for a single tweet in the nieve way because it amplifies really awful things that produce anger.
What reddit, twitter even youtube get wrong is that mixed up and down votes might mean very engaging but it does not mean good wholesome healthy content, infact it definitely means the opposite. If you just amplify engagement as they do today it might bring more clicks but it hurts society and builds extremists.
