Is validity a methodological issue or a conceptual problem?

I have selected "Making the Most of Your Video Collection: Trends in Patron Access and Resource Sharing" by Bergman (2010) assignment#3. The peer review on this article has directed me back to the issues of validity, reliability and generalization issues in qualitative researches that we have discussed a couple weeks ago. In his article, Bergman has proposed university and college academic libraries to review their existing loan policies and procedures on video collection. Her research has included detailed literature reviews to present research findings from university libraries, focus group (university student as library patrons) data, 2 surveys conducted by herself, and a case study on her personal experience with the Minnesota State University library. From the perspective of research method, I cannot deny that her research is well designed with considerable efforts. However, in regard to my background as a student library assistant in two different UofT libraries for more than one year (UTSC library as circulation desk SLA 2008-2009; East Asian Library as GSLA at present), some of what Bergman believes as the causes on the academic library policies and procedures are almost the opposite of what I have learned from my own work experience. Even though Bergman's MSU case can be regarded as a successful demonstration, it cannot apply to my case because her assumption seems to base on a general and standardized collection and funding situation of all academic libraries. Besides, the media format conversion issues are heavily restricted by copyright causes. However, Bergman has clearly stated at the end of her article that copyright issue is out of concern here. In my eyes, a research theory with ignorance of the fundamental contexts on the research subject is hardly reliable. The problem that I find on Bergman's article is mainly a validity issue particularly on the concept of generalization. The Research methods listed in her article seem to be data manipulation tools. In terms of this, I wonder is "validity" a methodological issue or a conceptual issue? If the researcher's primary research theory is already biased, the research methods as it supplemental reinforcement will be invalid. Will it be too late to check validity when the research has already been in the middle of its process? Which one shall be criticized first, the research assumptions or the research methods?