When a student is asked to conduct meta-analysis, what is he being asked to do? He is being asked to take the results of different studies, compare them, and see what patterns he can derive from his comparison.
One of the interesting things about meta-analysis is that the research has already been completed. It has been done by someone else. But now further research is being done on the existing results to see what more can be gleaned from the data.
One of the first things that the investigator has to do is to decide what the scope of his study is going to be. Clearly it cannot be anything general. That information has probably been attained already, and the original research is probably just about that. The investigator has to go into the study with a theory or a hypothesis already in his head, and when he is busy digging into the data, he is really looking for a way to prove or disprove his theory.
Meta Analysis in Statistics
Statistics uses meta-analysis a lot for proving a variety of concepts even though some of the earliest known meta-analysis came in the field of medicine and philosophy. One of the reasons why meta-analysis can be done is to get more accurate results and frame more precise concepts. One of the important steps that need to be taken by a meta-analyst is that the search for literature needs to be extensive and thorough.
The analyst cannot miss out on any source lest he works on problems that have already been dealt with. Since he is going to be working on multiple sources of existing research, he needs to make sure that his sources are complete in every sense. He might even want to put a team together to make sure that all possible results of previous research are collated and compiled effectively.