It is not always necessary to completely separating them. The issue is objectivity and correctness of the data and the representative data sample. If the one who gathers can objectively prove the validity of the sample and the analysis and show the analysis that brings the thesis to the table that is fine. However, if an antithesis is proposed the one who gathered and the one who analyzed the data must be able to look at the antithesis objectively, without bias. That might be hard for one who wants so deeply the result to conform to their understanding of the "facts".

So I agree that it can be that the two are the same person, but with restrictions.