So how do I know when I move from surveying existing research and doing research myself? The answer to this question is very involved. For example, when do you know you have finished your survey of existing information about your family? Let me give you a hypothetical situation:
Researcher makes a cursory survey and overlooks a huge completely documented and sourced user submitted family tree of his/her ancestors. As a consequence Researcher starts plugging away at collecting U.S. Census Records and a few vital records from online sites. When did the Research begin doing research? Is an individual's research independent of the existing body of knowledge?
What if I change my hypothetical so that the user submitted family tree is entirely complete but not documented at all? Does this change when a person completes a survey and begins doing research? Yes, these questions are similar to the old standby, if a tree falls in the forest and there is no one to hear was there any sound? (By the way, the tree question is a lot easier to answer, Sound is generated regardless of whether or not someone is present to hear it, the question is usually framed in terms of whether or not there is a "noise" which is intended to differentiate between a sound and a perceived noise. Even though there is physically no real distinction).
The difference between the question about the research and the question about the sound are only superficially similar. Since completing a survey is usually considered part of the research process, just because you don't do an adequate job of surveying what has been done on your ancestral lines, does not mean that you are doing research. In another example, when a PhD candidate makes a proposal for the subject of a doctoral dissertation, it is necessary to show what research has been done already before the candidate is allowed to proceed to do "new" research.
Now I readily admit that we aren't doctoral candidates and most of us never will be, but there is a fundamental principle here. The principle is that research, by its very nature, implies ploughing new ground. I speak from experience since it took me more than ten years to find out all of the information that was already known about my family. Even now, many years later, I do not feel secure that what I am doing is really adding anything to what is already known and recorded about my family unless I find something I know that no one has discovered previously.
You may find yourself in different circumstances, with little or no information about your family and you may complete the survey process in a matter of hours or days. But the important point is, even if you entirely ignore existing information about your family, your lack of knowledge of what has already been done does not mean that you are finding "new" information and thereby doing "research."
So, returning to one of the initial questions, when do you know you have finished your survey of existing information about your family? Looking to the Genealogical Proof Standard, it is when you have made a reasonably exhaustive search of a wide range of high quality sources. In my case, that took a long time and to some extent is still going on. Unfortunately, I often see people diving into original source records before making any effort whatsoever to ascertain what has already been done.