Yes, I think that's the point of the analysis: that "screen time" is wildly underspecified and that it's probably correlated to other unconsidered factors.