If a user is entering text at a keyboard, then is forced to use their mouse for something (e.g. clicking a button in a modal dialog), what are the time and cognitive costs of that? Similarly, when users are not entering text, do they tend to interact in physical poses that are costly to switch out from?
Reading literature on GOMS analysis, I see a benchmark of a time cost around 0.4 seconds for switching, though I don't know where this figure comes from. It seems rather conservative, and doesn't seem to take into account the physical exertion and awkwardness I have experienced myself on switching.
Has any HCI or ergonomics research touched this topic?