Just pointing out that your data has a pretty clear confounder. “Wfh” and “wfh during the initial stages of a global pandemic” are very different things.