Social Media, News Consumption, and Polarization: A Field Experiment (2020) | Hacker News Reader

Social Media, News Consumption, and Polarization: A Field Experiment (2020)

papers.ssrn.com

63 pointsdash213 comments

This is a really interesting article, because it's actually about something more fundamental than filter bubbles, and it turns out that Facebook creating filter bubbles is pretty much a corollary.

The abstract:

Does the consumption of ideologically congruent news on social media exacerbate polarization? I estimate the effects of social media news exposure by conducting a large field experiment randomly offering participants subscriptions to conservative or liberal news outlets on Facebook. I collect data on the causal chain of media effects: subscriptions to outlets, exposure to news on Facebook, visits to online news sites, and sharing of posts, as well as changes in political opinions and attitudes. Four main findings emerge. First, random variation in exposure to news on social media substantially affects the slant of news sites individuals visit. Second, exposure to counter-attitudinal news decreases negative attitudes toward the opposing political party. Third, in contrast to the effect on attitudes, I find no evidence that the political leanings of news outlets affect political opinions. Fourth, Facebook's algorithm is less likely to supply individuals with posts from counter-attitudinal outlets, conditional on individuals subscribing to them. Together, the results suggest that social media algorithms may limit exposure to counter-attitudinal news and thus increase polarization.

Yup, what Facebook does is just one of the results. It also suggests the kind of nudges Facebook could do.

In what sense is conditional used here? E(X | Y) or E(X | Y = y) ?

I don't think the phrase "conditional on individuals subscribing to them" is meant to be a probabilistic conditional. I think they are using "conditional on" solecistically to mean "unless".

In a non-mathematical context it would mean "given (that)", not "unless".

I think you're right, but this is a difficult sentence to parse:

Facebook's algorithm is less likely to supply individuals with posts from counter-attitudinal outlets, conditional on individuals subscribing to them

This is pretty interesting, but I've always thought that the sensationalist side of social media exacerbates this issue. Everyone is driven by clicks, so writing with a sense of urgency (not necessarily extremism) will drive your interaction up. It's part of what makes social media so exhausting these days: a billion users are all jockeying for your attention, and you need to trust that Facebook/Twitter/Reddit/$OTHER_SITE will direct your focus to the things that matter. And when everything "matters", nothing does. We've become desensitized to importance, and I imagine it will have a considerable impact on the future of journalism and how we interact with it.

So much this. My latest theory (I’ve spent more time than I care to admit pacing in my home talking to myself about social media) is that the incentive for content creators/publishers to get clicks combined with the increased behavioral pattern of readers staking claims by way of aligning their social media personas with a visible belief or #hottake (basically an online bumper sticker) — all sprinkled with the instantaneous exchange of information that is the internet — has created a perfect swirl of perpetual polarization without any wedge to throw in to stop it.

I’ve been concluding lately that the only way to slow down or stop continued polarization is to drop off of social media entirely OR remove revenue-earning by way of just clicks.

I don’t see either of these happening.

We started https://rational.app exactly for this reason.

If you’re interested then join the project

I would love to see this done in more realistic dimensions, not just "liberal" & "conservative".

What others would you propose? That is currently the biggest and most divisive one, at least in the US.

The US use of the word 'liberal' is already ambiguous, and is generally orthogonal to 'conservative' (e.g. see https://en.wikipedia.org/wiki/Liberalism ).

Some interesting dimensions might be:

- liberal vs authoritarian

- conservative vs progressive

- nationalist vs globalist

- individual vs collective

I wonder what effect this has on peoples' perceptions and willingness for political parties to negotiate and compromise. My general sense is that polarization makes compromise practically impossible, hindering democracy (in the US) to a point where it looks too ineffective.

---

More personally, most of the political news media I 'consume' comes from outlets that oppose my own perspective. I used to think I was 'hate-listening' to this stuff, but I also realized that listening to outlets I already agree with feels a bit too self-indulgent for me. I'm curious if this is like a separate form of junk food that requires its own kind of self-control to avoid.

Reply on news.ycombinator.com