Here's a blog post summarizing @manoelribeiro's research on post approvals in online communities! To paraphrase a key takeaway, "post approvals creates communities centered around fewer, higher-quality posts". research.facebook.com/blog/2…

How does social comparison on social media vary by country? research.fb.com/blog/2020/10… #CSCW2020

If you’re interested in working on computational social science problems at FB in 2021, I’m looking for interns! Also happy to chat about any questions you have! More info: facebook.com/careers/jobs/32…

When do misperceptions happen in online conversations? (with Jonathan Chang & Cristian Danescu-Niculescu-Mizil) research.fb.com/blog/2020/04…

Research on @GamesDoneQuick by Stephen Tsing-Han Sher suggests that featuring “the couch” on stream helps create a virtual living room that viewers can be a part of. #CSCW2019 static1.squarespace.com/stat…

A large-scale analysis of dark patterns on shopping websites found that (1) 40% of countdown timers don’t actually matter (2) 9% of uses of social proof (e.g., “N people bought X in the past hour”) are false. Work by @AruneshMathur at #CSCW2019 webtransparency.cs.princeton…

Algorithm-in-the-loop decision making should be accurate, reliable, and fair. But people don’t do the last two: they don’t rely more on an algorithm even if it’s more accurate, and rely on their own judgment more when evaluating some groups of people. @benzevgreen at #cscw2019

Many factors to consider when designing interventions for reducing political polarization (e.g., if an interface seems to question a person’s competency). By @matnel at #cscw2019 tuhat.helsinki.fi/ws/portalf…

On Github and Wikipedia, larger teams produce more work per team member. (Have/can these observations be replicated in an RCT?) Work by @goranmuric, @aabeliuk, @KristinaLerman, @emilio__ferrara #CSCW2019

When given a speech speedometer, native speakers “slowed down” by using shorter sentences, not speaking slower. Unfortunately, the speedometer was distracting, so their speech became less clear to non-native speakers. By Wen Duan. dl.acm.org/citation.cfm?id=3… #CSCW2019

Exposure to opposing groups in online platforms may exacerbate polarization. Subreddit members with intergroup contact tend to use more negative language in their affiliated group. By @darkblue_jason @ChenhaoTan arxiv.org/abs/1908.10870 #CSCW2019

When a team that previously broke apart reconvenes in an anonymous setting (members don’t know they’re working together again), it fractures again on some tasks (debating viewpoints) but not others (generating ideas). By @MarkWhiting, @msbernst hci.stanford.edu/publication… #CSCW2019

Slowing algorithms down can both improve people's judgment and promote task reflection. Work by @joon_s_pk dl.acm.org/citation.cfm?id=3… #CSCW2019

User behavior may improve after comment deletion simply because they were in a worse mood when they made a deleted comment. Controlling for this, comment deletion still reduces future comment deletions, but does not decrease swearing. @ChenhaoTan chenhaot.com/papers/content-… #cscw2019