It is impossible to do without the interview process, given the additional and unexpected insight it has the potential to yield to pure numerical data, yet no matter how potentially therapeutic, it is always using the experience of the subject for purposes that he or she might not directly benefit from. Perhaps the answer to the unanswerable questions that Nunkoosing raises is that the ethical obligations of interviewing are similar to that of much of qualitative medical research: the research should be designed to serve the community used in the interviews, thus making at least the benefits of the research outweigh any possible individual misunderstandings that might take place. Using health data from a low-income community to help a high-income community, for example, would be deemed unethical, regardless of the level of 'truth' of narration. Furthermore, by acknowledging...

But the audience -- the interviewer and the ultimate readers of that self -- must be aware of the language that the interviewee is speaking, as well as his or her own biases. Thus, the interviewer, the first transmitter of the words of the subject, must be well-versed in the culture and assumptions of the person he or she is interviewing, to avoid the 'static' of miscommunication and misplaced, unintended subjective shaping of the interviewee's experience.
References

Nunkoosing, Karl. (2005). "The problem with interviews." Qualitative…