I'm reading a book called "Writing the Doctoral Dissertation" by Gordon Davis and Clyde Parker. The advice given seems to be for a PhD program that is structured entirely differently than ours. Their model student (they provide a "Goofus and Gallant") plans his research design in his research methods course. On the subject of picking a dissertation topic, they write, "Rarely does a student have a well-defined topic in the beginning. Generally, the initial topic is poorly defined, too general, and too large in scope."
Our methods course occurs in our very first semester. The proposals we created in that class were certainly poorly defined, general, and large. As we refined them over the semester, they improved, but many revealed themselves to have fatal flaws or to be embarrassingly naive. I don't know that anyone came out of there with a workable proposal; by the time we had learned to design an excellent piece of research, it was too late for that particular project. But several semesters later, we are now readying ourselves to choose dissertation topics and defend proposals. Our faculty expect that we are able to propose good research. If a member of my cohort went to them with a topic that was "poorly defined, too general, and too large in scope," I think they would be disappointed.
That doesn't mean the faculty don't have suggestions for improvement; we still only have two years of experience, and they have many more. But we all have what the book supposes only a lucky few will have a minimum of - experience in coming up with research proposals and in working on large research projects.
This book isn't very useful for me, I've decided. It's written for someone with a different personality - someone who takes a lot of convincing that planning is worthwhile. It's also apparently written for someone clueless - someone who needs to be told that when doing a literature review, one should use online databases and dissertation indices.