←back to thread

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

(www.nber.org)
506 points imakwana | 1 comments | 21 Apr 25 04:24 UTC | HN request time: 0.228s | source
Show context
nomilk ◴[21 Apr 25 05:58 UTC] No.43748834[source]▶
The surprise here is how little of an effect it has. Deactivating facebook makes you only 1/16th of one standard deviation happier. And instagram even less. And this was measured during elections, when the effect is likely to be greatest.

Kinda crazy that the magnitude is so small! (my next [admittedly rather cynical] thought is "who funded this?")

replies(13): >>43748884 #>>43748905 #>>43748936 #>>43748945 #>>43748990 #>>43749080 #>>43749338 #>>43749527 #>>43750904 #>>43751816 #>>43752078 #>>43752906 #>>43754629 #
highwaylights ◴[21 Apr 25 06:09 UTC] No.43748884[source]▶
I’d be interested in the results of a study that cuts out all social media, but the problem I can already see with that is self-selection bias (the people that would volunteer for it are probably already eager to get away from social media so the results would likely be skewed).

Personally I’ve been mentally in a better place since getting rid of my social media accounts during COVID, but it does cause problems because Facebook has become a utility as well (schools and real-life social groups use it for co-ordinating activities).

replies(1): >>43749512 #
1. photonthug ◴[21 Apr 25 08:20 UTC] No.43749512[source]▶
The perceived utility of social media seems pretty variable, not just across people, but with the same person in different circumstances. With covid, social media might scare people out who were regular users previously, and yet for other occasional or reluctant users it's suddenly seen as the only option for human contact and they use it constantly. After lock-downs are over, people flip to the polar opposite of their previous preference. With recessions, social media might be the only affordable entertainment but during better times, many would rather do something else. In general I bet it's insanely hard to run good experiments for behavioural economics in volatile times, even if you're really trying to be careful about methodology.