I would tend to say "no" as the interview and observational data would have been synthesised/integrated at some point, thus making it difficult to extract "just" one element from the overall findings.
However, it may be possible if the two sets of data were analysed and findings properly reported separately.
I'm also wondering why you would want to separate the results, rather than taking them as a whole as intended by the researchers? The point and strength of mixed methods is convergence of data; separating out strands of data thus weakens the overall strength of the approach.
Not so straight forward as the accumulation of logic which occurs in systematic reviews. I agree with Aled's observations over the strength of mixed methods. We recently had similar issues and opted for including quantitative results into a systematic review, and qualitative into metasynthesis?
Thank you Aled Jones and Nigel Rees for your answers. I am agree with you about the mixed methods results. Unfortunately, I couldn't find enough qualitative papers (minimum=5papers) for doing my qualitative systematic review.