3.3 Questionnaires
There are different kinds of questionnaires. They can be classified either by their form (written or verbal) or by the way they are distributed:
Postal and other mass distributed questionnaires are good for retrieving a large amount of answers in a short time. The major problem, however, is that the reliability of self-reported data is questionable. One can never be sure that the answers are really spontaneous and candid as they might be given after some discussion with others. Participants may use a more formal form of a word (see observer's paradox) or have trouble understanding questions. In addition, mitigating the observer's paradox by trying to mask the intention of a questionnaire distributed in this way is very difficult and makes the questionnaire a lot less handy (cf. Milroy and Gordon 2001: 14, 51-52).
Fieldworker distributed questionnaires are a mixture of on the spot transcription/recording, postal questionnaires and interview questionnaires. Here, the advantages are that illiterate people can be reached also, and that both fieldworker and informant can ask questions in regards to clarification of any kind, i.e. in regards to the meaning of a certain question or the pronunciation of a particular word etc. The problems, on the other hand, are that it takes a long time to gather large amounts of data and that masking the reasons of the questions is not easy. Furthermore, in regards to the observer's paradox the same holds true for this approach as it does for postal questionnaires (cf. Milroy and Gordon 2001: 14, 54-56).
To eliminate the observer's paradox one can apply a rapid and anonymous survey. One of the most famous studies and simultaneously the prototype of this method is Labov's (1972a) “fourth floor” study in New York. The underlying principle is to pose a question with a specific response in mind and then write down the answer given by the informant. The problem is that the goal has to be very clear cut and considerable thought has to be given to what to ask and how to pose the question. It yields a very limited scope of information and one cannot make a very good categorization of the informant (cf. Milroy and Gordon 2001: 56-57).
Questionnaires used in interviews are explained further in the chapter interviewing.
Created with the Personal Edition of HelpNDoc: Create HTML Help, DOC, PDF and print manuals from 1 single source