than truthfully. These kinds of problems are sometimes refered to as ‘non-sampling errors’, that is, errors due not to the sampling process but to faults or shortcomings of the research design or the manner in which it was implemented. Unfortunately, some of these errors are difficult to foresee and guard against, so that it is possible for surveys to go wrong and produce misleading or false estimates of the characteristics of a population. Ultimately, the most important limitation of survey is that, in order to be successful, it must depend on a tightly structured inflexible questionnaire.
Moreover, howsoever well designed the questionnaire might be, its success depends finally on the nature of the interactions between investigators and respondents, and specially on the goodwill and cooperation of the latter. Interview An interview is basically a guided conversation between the researcher and the respondent. Although it has few technicalities associated with it, the simplicity of the format can be deceptive because it actually takes a lot of practice and skill to become a good interviewer. Interview occupies the ground between a structured questionnaire of the type used in surveys, and the completely open-ended interactions typical of participant observation methods.
Its chief advantage is the extreme flexibility of the format. Questions can be re-phrased or even stated differently; the order of subjects or questions can be changed according to the progress (or lack of progress) in the conversation; subjects that are producing good material can be extended and built upon others that provoke unfavourable reactions can be cut short or postponed to a later occasion, and all this can be done during the course of the interview itself. On the other hand, many of the disadvantages of the interview as a research method are also related to its advantages. The very same flexibility can also make interview vulnerable to changes of mood on the part of respondent, or to lapses of concentration on the part of interviewer.
It is in this sense an unstable and unpredictable format — it works very well when it works, and fails miserably when it doesn’t. There