Do not start with a computer search (eg
Med-Line). You may find that you get overwhelmed with references and
are not yet be able to judge their relevance.
Instead start with a few recent papers related to your question. If your
study relates to previous studies in the lab you are in, start with a few recent
from the lab.
In these papers, look up each of the references that relate
to your question. Repeat the process in each of these
papers until the returns (new references) start to diminish.
Using citation index look forward, to see what
new papers have referenced this paper. Look up the references in these papers.
One such search engine is the Web of
Science from ISI
Only now use a search engine likePubMed, to check if there is
anything that you have missed. Only now do you have enough background to
clearly define a focused search
Look up meeting abstracts. These tell you what
papers will be coming out in the next year or so.
Look up the references in the related chapters of a general text to make
sure you have not left out some classic paper.
The hard thing is to remain focused but at the same time not miss that
one critical paper.