I thought about this a lot and aside from simple "dont zoom (in)" advices, I was unable to find anything going deeper into the topic so far.
In my opinion, there are many factors contributing to the "to zoom or not to zoom" question, aforemost budget limitations. This can be seen if you compare TV to Cinema: Cinema has more money, so they can do more elaborate things than zooming (tracking/dolly/crane), TV often has to fall-back to zooming as everything else usually invovles more complicated sets, lighting and sound-recording.
The question why there are so many zoom outs and not many zoom in nowadays is also striking to me and I recognized a very important asymetry: If I zoom out, more and more genuinely new information enters the frame. It gives the effect of "aaah, that's the big picture" and done properly (lens choice) feels very natural.
A zoom in doesn't provide new information (except it's really far in) and almost always feels wierd, as the non-uniform change in object sizes (which is easily forgivable in zoom outs) is disturbing and unnatural.
Of course this all depends on the evolution of certain memes and perception habits as they change over time, but nevertheless, zooming in general is hard to get right, so it's seldom done, so viewer don't get used to it (and "bad zooms") so it gets even harder to do, and so on and so forth.
Luckily, we have all the new Indy-movement due to internets and such, so we can always expect an sudden and unexpected change in viewing "fashion".
BTW, compare zooming in video games history. I think it's very enlightening, as it is easier to distinguish the technical necessities from artistic freedom and the memes are much better documented.