Comment by doctorpangloss

what it means is that this paper shows probable causality and models a lot of interesting features. it is most definitely not flawed.

i think the tough thing is that 0.6 percentage points gain for the average student is quite small. it's actually less than you gain by studying for 1h for the SAT, which is probably about 0.9 percentage points, depending on how you interpret college board's research (it recommends 20h of studying). that is to say, if students studied one fucking hour for the FAST, they would probably get a bigger benefit on it than all the time they get back not looking at their phones throughout two years of school.

so whatever cell phone use (1) in school (2) causes, it causes a small effect on test scores.

you would have to pick some other objective criteria, for example mental health assessment, for maybe a larger effect, or seek a larger treatment, perhaps a complete ban of cell phones period, to observe a larger effect.