Meta & Children's Online SafetyChat: if we want product leadership to prioritize wellbeing product changes we have to come up with a killer statistic
Source PDF not yet released. Text is transcribed from the public filing.
Thread View
████ name redacted. Different bars may be the same person or different people.
quoted in an excerpt in the archive.
Friday 25 September 2020
09:23████Reminder that Nino will be presenting today at 1pm in the CDS Tech talk, about 2021 opportunities
09:25████yes, please join!
09:54████for the people in the south bay/peninsula, I just took the flu shot in Crossover in Santa Clara. There's another one in mountain view (shoreline/101). It was pretty fast and easy, I didn't have to get out of the car or touch anything and was done in 3-4 minutes. You can see the exact locations and make an appointment in the links in this wiki: https://www.internalfb.com/intern/people/portal/benefits/health-and-wellness/flu-shots
14:08████My takeaway was that if we want product leadership to prioritize wellbeing product changes we have to come up with a killer statistic. (That seems both simple and hard.)
14:09████and for some problems, probably impossible
14:11████@Justin, I think some of our descriptive stats from the body image survey could potentially be enough to change the conversation.
14:12████was this "girl teens being 32x more likely to get unsolicited sex talk on IG compared to FB" the main example?
14:12████i think so
14:12████seems like everyone is getting all hot about problematic use again, bc of the stupid documentary. could be an opportunity.
14:19████32x, but what's the baseline rate?
14:20████I don't think they mentioned it in the talk, but there was a reference in the slides.
14:23████Yep, glad that we didn't abandon the research.
14:34████the baseline was FB
14:35████whatever they use to detect sex talk with minors on FB, on IG it was 32X. The prevalence is low, of course
14:35████(fortunately)
14:35████Or by baseline you were asking for the specific prevalence?
14:36████Also, Reels seems to be everything they denounce in the stupid documentary and everything we know from our research: passive consumption of an endless feed, without any connection to the content creator. yay
14:37████Exactly. Ugh.
14:37████Passive consumption of content from people we don't care about. It's like the anti-Moira
14:39████I think this was the number of teens who had a DM/Messenger thread with a "sketchy" adult. So on IG there were 32x as many of these threads in the US than on FB. (Don't know how they defined sketchy.)
14:40████I think it's sex-talk between adult and minor, from a prior conversation I had with Nino
14:41████(with all the caveats about the quality of our detection of these things)
14:41████Oh! Glad you clarified.
14:41████IDK if it is restricted to non-followers
14:45████Yeah. So like 32x could still be small and/or the CIs could also be large.
14:46████Yes. I had some convesations with the child safety team when I was working on age affinity. It's low prevalence, but extremely high intensity/risk/damaging
14:47████The investigator I talked to at FB was the saddest person I've ever met. Seriously, it was heart breaking.
14:47████Not that it's not important, but I guess my point is that there's a lot of leeway in how stats are presented to achieve some goal.
14:48████#fakedataforgood?
14:48████(I kid, I kid)
14:48████more than #fake it's like #lyingwithstatistics
14:49████I think the point was that there was an effort at FB, not a lot of attention at IG and when they looked they realized it was much worse on this side of the fence
14:57████Q: Is loneliness decreasing over time?
A1: Yes it is, and the difference is statistically significant.
A2: No, it's not changing. The trend is basically flat.
15:07████████ potential example of #lyingwithstatistics or...?
15:08████A difference of opinion? Haha
15:11████Yeah - it's tough. A1 is stats, A2 is interpretation of those stats/effect sizes as actually meaningful/important or not (which can go either way, and easily influenced by what we want to find/conclude).
15:26████These discussions made me think about Joseph Adler's old note about cheating with deltoid: https://fb.workplace.com/notes/experimentalists/how-to-lie-with-deltoid/281428872493535/
15:50████That is the most amazing note I've read this year.