[ghusto]

"I conducted a meta analysis of 46 studies and concluded that there is _no correlation with social media and teen mental health issues"

What about the studies that that _do_ show a correlation, of which there are more than one. Why were those discounted?

[rich_sasha]

Meta analyses look across published studies and essentially do statistics on these. The claim of the paper is that the studies that show adverse effects are explainable by fluke, and presumably are outnumbered by studies that don't show adverse effects.