The effect of deactivating Facebook and Instagram on users' emotional state

Original Article ↗ Hacker News Discussion ↗

Study findings & their size

  • Commenters focus on the reported ~0.06 SD improvement in an emotional-state index after FB deactivation and ~0.04 SD for Instagram.
  • Many interpret this as statistically real but very small in practical terms – roughly a 1% change on a 0–100 mood scale, far below common therapy effect sizes (~0.3 SD).
  • Others stress that “small” doesn’t mean “zero”: even a small average shift might hide larger benefits for vulnerable subgroups (e.g. undecided voters, young women) and for outliers who feel dramatically better.
  • Several explanations of standard deviation and percentile-shift try to make the numbers graspable; there is some confusion between statistical significance, effect size, and “percent vs percentile”.

Methodological and scope concerns

  • The experiment lasted only six weeks and coincided with the run-up to the 2020 US election, when background stress and political noise were already high.
  • Participants only quit Facebook/Instagram, not all social media or news, so many could substitute with other feeds (Reddit, news sites, etc.), potentially diluting measured benefits.
  • Less than 1% of invitees completed the experiment; several readers highlight strong self-selection and question generalizability.
  • Some wonder whether funding and Meta’s involvement biased framing, while noting the formal claim of academic independence.

Lived experience vs measured effects

  • A large number of anecdotes report much bigger subjective gains from quitting FB/IG/Twitter/Reddit/LinkedIn/news feeds: less anxiety, fewer political spirals, better focus, improved sleep and work output.
  • People describe feedback loops where algorithms keep surfacing grief, body-image, or political triggers, especially during vulnerable periods.
  • Others find that simply stopping engaging (no replies, no arguments, minimal voting) sharply reduces stress, even without fully quitting.

Feeds, algorithms, and incentives

  • Strong consensus that algorithmic feeds optimize for engagement, not user wellbeing; many see this as structurally incompatible with ad-driven business models.
  • Nostalgia for early Facebook/Instagram: reverse-chronological posts from real-life friends, fewer brands and influencers. Users share URL hacks and buried settings to approximate this.
  • Several propose user-controlled or local ML filters to re-rank or strip “slop,” but doubt platforms will ever prioritize that.

Lock-in and alternatives

  • Many want to quit but feel trapped because schools, clubs, hobbies, businesses, and neighborhoods coordinate via FB, IG, or WhatsApp.
  • Group chats (WhatsApp, iMessage, Signal), niche forums, RSS, Mastodon/Bluesky/Lemmy, and email are praised as healthier replacements, though they lack the reach and convenience of big feeds.
  • Some predict social media will eventually be viewed like smoking or gambling; others liken it to alcohol—harmful for some, manageable in moderation.