To clarify a bit further as I can no longer edit my post: The points about the methodology stand, and I don't believe one should update ones beliefs in either

To clarify a bit further as I can no longer edit my post: The points about the methodology stand, and I don't believe one should update ones beliefs in either way based on this paper. The "killer" finding is not necessarily a killer towards the reported results (which are dubious anyway), but towards the way the paper is presenting these results. The missing discussion of this interaction effect between phone dependence and phone location is what makes me very much doubt the quality of the paper.