I think this needs a (2017), here is a link to a PDF - https://rady.ucsd.edu/docs/faculty/aGneezy/Published%20Paper...
>Five hundred forty-eight undergraduates participated for course credit. [...] Our final sample consisted of 520 smartphone users.
520 isn't as bad as I was expecting, but it's still an order of magnitude away from being meaningful. And, obviously, a study population consisting only of students attending the same class is maybe going to be slightly biased.