Hacker News new | comments | ask | show | jobs | submit login
Why Echo Chambers Are Useful [pdf] (ox.ac.uk)
56 points by crunchiebones 73 days ago | hide | past | web | favorite | 19 comments



The problem with echo chambers isn't so much verification (is the information accurate?) but rather validation (is the process of acceptance rational or qualified?).

Echo chambers tend to be hyper-subjective, which means new information is accepted or rejected on the basis of content-oriented rules opposed to whether decisions upon the information are valid, qualified, or balanced. That suggests the injection of precisely targeted information focusing on group acceptance is the only criteria of success, which makes the group dangerously susceptible to manipulation.

Counter-intuitively echo chambers are primarily desirable for a number of irrational reasons not related to subject matter. People tend to find security in conforming group dynamics, generally fear originality, and strongly enjoy commitment even in opposition to strong evidence.

Contrarily, in echo chamber opposed groups the primary qualifier of information acceptance is process of argumentation. The idea or information that survives a brutal, but process-approved, argument is the most worthy. That form of objectivity can be emotionally disruptive for persons not prepared or understanding of the group process.


> Contrarily, in echo chamber opposed groups the primary qualifier of information acceptance is process of argumentation.

Ostensibly, but there's a lot of people notionally opposed to echo chambers who are very happy in their own.


Lots of people are hypocrites or just completely oblivious to the cocoons they’ve built.


I found this an interesting read, and it certainly makes a good point. Echo chambers are definitely going to have a greater net throughput of information compared to a network in which members aren't keen to share with one another due to perceived differences. So overall, members of an echo chamber can end up more informed than if they hadn't been. However, I think it's important to note that modern discourse isn't critical solely of echo chambers, but also of their effects when false information is introduced into one. The same trust that allows more informed members also facilitates misinformation.

That being said, it's probably more effective to use that lack of verification to police misinformation within the echo chamber than to introduce members that could offer contradictory - but also potentially false - information. It would just require some added diligence from members of the echo chamber.


it's probably more effective to use that lack of verification to police misinformation within the echo chamber than to introduce members that could offer contradictory - but also potentially false - information.

Putting aside the verification problem for a minute, doesn't the echo chamber problem lead almost automatically to group think?

I do machine learning, and I see a strong analogy here with overfitting. If all the data is from a subset of of population and it is never validated outside that subset then the model will become strong on that subset data but much weaker outside.

In machine learning we use a variety of regularization techniques to combat this. One thing that actually works pretty well to build robust models is (a small amount of) mislabeled data: the model learns not to depend on any single feature.


What do people actually mean by "groupthink" here and why precisely is it bad? Everyone drives on the same side of the road and (mostly) follows the same traffic rules and we don't call it "groupdrive".

Surely having everyone believe the same true thing isn't actually bad - nobody calls Maxwell's equations "groupthink" unless they're trying to sell you a perpetual motion machine.


Groupthink is a psychological phenomenon that occurs within a group of people in which the desire for harmony or conformity in the group results in an irrational or dysfunctional decision-making outcome. Group members try to minimize conflict and reach a consensus decision without critical evaluation of alternative viewpoints by actively suppressing dissenting viewpoints, and by isolating themselves from outside influences

https://en.m.wikipedia.org/wiki/Groupthink

The most commonly used example is the Bay of Pigs invasion, where no one advising Kennedy raised any possible negative consequences. That’s seen in contrast to the Cuban Missile crisis a few years later where Kennedy actively sort contrary advice.


I think that'd be the line at which it ceases to be an echo chamber though, no?


On a related note, I have recently been wondering if the large scale swings in society from being outgoing to being insular and back, might be evolutionarily beneficial due to the speed of genetic drift varying for different sizes of breeding populations.

Small populations get the fast drift for adaptation, but mixing those with other adaptations from other small populations gets strength in depth, so you have two opposing evolutionary strategies, one which prefers isolation of groups and one that prefers mixing.

Over evolutionary timescales, oscillating between these two behaviours could presumably get baked into the genes, so leading our politics to swing between isolation and gregariousness as part of a natural evolutionary cycle.


Very interesting. I have had this feeling watching political news/gamergate/comicgate/damore etc.

Too much energy gets expended trying to convince/shame/punish the other side. None of it imho has produced outcomes or increased understanding.

When dealing with unknown unknowns/ambiguity etc why not just split the groups, let them go do their thing on two separate islands, like running two parallel jobs based on contradictory assumptions and let the best job win.

How do we split without causing a mess is the big question? We maybe in an overly connected state atm or certain issues require disconnection and echo chambers.

Thought experiment (slightly ridiculous and loaded but use your imagination here) - there is a lot of talk about splitting up Google for example. What if Google is split on an issue like unconscious bias training. Those who support it produce one search engine and those who don't (damore camp) produce another. Don't we get better information over all?

Right now when I read a "this is how you need to think" article on CNN, I shake my head and go look at what Fox has to say. And when Fox is fawning over Trump I go the other way.


> Too much energy gets expended trying to convince/shame/punish the other side. None of it imho has produced outcomes or increased understanding.

I find a better fit to take the (somewhat cynical) approach that each side is trying to increase "vendor lock-in" for their own side. Media coverage of politics makes much more sense when I assume each side is only talking to people who already support them.


Parallel jobs are great (the map step)... but you generally need a reduce step as well.

If you at all try to be a neutral player in a partisan world, you quickly notice that each faction does a curious filter before the reduce. Its a filter that keeps the worst arguments/conclusions of the opposition and discards the best, before they can be compared. And for the in-group, it keeps the best and discards the worst.

Only after that, is there a (metaphorical) reduce. And lo and behold (drum roll...) their own ideology wins! How comforting and reassuring! You were right all along!

But the major league, all-star team is playing against a team from the t-ball league (relatively speaking).


From an advertiser analysis of both wings, it seems like the only thing both liberal-biased and conservative-biased television news consumers can agree on is discreetly mail-ordered boner pills.

I can't stomach too much of either CNN or Fox News. I can take just a bit more of MSNBC. But I really miss unbiased, fact-based journalism, without all the relentless subtext and agenda-pushing, without the editorializing. But I'm also not willing to pay for it, so I really can't blame anyone else for its absence.

That doesn't mean I can't complain about all the blatant, rampant editorial bias. It is polarizing the audience, likely by design, so as to better target the advertisements that pay for it all. Echo chambers are great for advertising niche products to niche audiences.


When dealing with unknown unknowns/ambiguity etc why not just split the groups, let them go do their thing on two separate islands, like running two parallel jobs based on contradictory assumptions and let the best job win.

The problem, currently, is that the winning ideology is enforced on all islands. The stakes wouldn't be so high if the other islands had no influence on your island.

This is part of the rationale for keeping decisions at as local a level as possible (and avoiding one-size-fits-all policies.)


I've had a similar thought experiment before, where basically you let people of different political leanings go to their own countries with their like minded ilk. Two problems arise:

1. People tend to hold complex combinations of views which may not completely align with everyone else. How many partitions do you need in order to make this work?

2. Mobility remains an issue. People simply don't have the resources to get up and go to something they actually support and build it up.


I've also had similar thought experiments here. But take it a step further. It's probably safe to say that the current biggest divide in society is in political ideology. Let's just call this left/right, though that's far from precise. Imagine we placed all 'left' individuals in one US, and all 'right' individuals in another completely identical (in terms of physical resources, preexisting infrastructure, etc) US. And migration between these two nations was freely and instantly allowed, but only if you absolutely abided the initial ideological split. How long would it take before these two nations then had their own internal splits with people dividing themselves among some new subissue?

Of course you hit on this with #1, but I think the answer is that there is no limit to the number of partitions. In other words that division is itself inevitable. So rather than trying to ideologically homogenize ourselves, I think it's more important to for people to embrace diversity of views as opposed to attack everything that doesn't conform to, what to them, seems self evident. Because for 'the other side', their views are likely seen as identically self evident. When sides remain incapable of doing anything except working to antagonize the other, in the end the only way things would be resolved is by force of arms which is certainly something almost nobody wants.


This is an interesting idea (that echo chambers may be objectively useful in certain ways), but to me the bigger question is, even if that is true, how much of a factor is it compared to more mundane causes like plain old cognitive dissonance?


Disappointed to discover this isn't about acoustics.


You shouldn't be surprised, "echo chamber" has taken a new meaning ever since social media became influential.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: