That's a good analysis, but I don't find it convincing. He's trying really hard to disprove Haidt's post by poking holes in many of the studies. If you look at 386 studies in the social sciences, of course you'll find issues with the analysis or design of many of them.
The larger trends ("most of the effect is driven by teen who use no social media", etc.) aren't supported by the data he presents (look at the table of "social media time" -> Depression for example).
Are the researchers who look into this problem predisposed to finding a connection? Probably. But I do think the open, community based analysis Haidt led was done well and if you look at what they found digging through 386 studies, it's compelling.
