Most of these results are quite weak, especially those around the number of connections ("ties") a given worker has with those in other parts of the organizatio

Most of these results are quite weak, especially those around the number of connections ("ties") a given worker has with those in other parts of the organization. The only significant result I see is a decrease in time spent collaborating with others. There are two major confounds I see, neither of which is adequately addressed: 1. The study relies entirely on numbers of electronic communications (IM, email, voice and video calls), taking February as a baseline and measuring differences from that in following months. No apparent effort is made to account for in-person collaboration, or for the reduction to ~zero of messages relating specifically to the needs of in-person work. When everyone is remote, no one needs to ask who wants lunch or where everyone wants to go. 2. As the authors themselves acknowledge, they have no way to account for any impact, beyond the switch to WFH, of the COVID-19 pandemic. They attempt to fudge this by applying some multipliers whose derivation seems uncertain, and excuse it by saying that other studies using similar econometric methods do likewise. For management interested in pushing against all sense to force workers back to the office status quo ante, this might make a useful argument assuming that no one they're attempting to persuade looks too closely at the paper. For anyone trying to understand accurately the impacts on work culture and output of a broad switch to WFH, I think there's little of value here.