Self-administered questionnaires

Here we are dealing with the limitations of the self-completion questionnaire. It is true that ‘eliminating’ the interviewer eliminates interviewer bias but it is also true that no follow-up questions can be asked. It is equally true that respondents are very likely to read through the whole questionnaire before answering the first question (although web-based questionnaires can reduce this effect), which might tend to produce other responses than might have been given spontaneously. Partly because of this pre-reading, respondents may not answer all the questions, causing a coding problem later. However, far from being a disadvantage, researchers are encouraged to set closed-ended questions in this type of questionnaire because open-ended questions may cause problems of interpretation as well as proving too tedious to complete.
Reference: Bryman: Social Research Methods: 5th Edition Page(s) 223,224

Lack of response can seriously hurt the representativeness of a randomly drawn sample, so any device that will tend to improve the response rate, like all the possibilities shown in this question, are to be encouraged. Professional researchers may go so far as to offer a small cash incentive, for example, to stimulate higher response rates. Not all samples are drawn randomly, however, so the actual response rate doesn’t have the same significance, if indeed it has any at all. Bryman concludes “Tips and skills” on page 225 with the advice that “you should not be put off using (these) techniques because of the prospect of a low response rate”.
Reference: Bryman: Social Research Methods: 5th Edition Page(s) 225

Because of the manner in which diarists are approached, they are more likely to be more highly motivated to keep to the task. It must be said, though, that attrition is a common problem in research. For structured diary use in quantitative research there may not be the same problem of attrition as for free-text diaries used in qualitative research. Using a diary method for quantitative research will gather data about frequency and duration of behaviour only, so answer (b) might only be partially correct. Answer (c) relates to free-text diaries and so has no bearing on this question concerning quantitative research. Probably the diary scores better than a questionnaire for measurement validity of amounts of time spent on particular activities, their frequency, and sequencing.
Reference: Bryman: Social Research Methods: 5th Edition Page(s) 237-239

Naturally if you place the possible responses vertically they take up more lines on the page, not less. “Tips and skills” on page 227 demonstrates this quite clearly. On the other hand, setting out questions in this way may make the overall questionnaire appear longer than it really is, which, apart from wasting paper (in a postal survey) might discourage the respondent from completing the questionnaire. Closed-ended questions should not seem to offer more than one answer and if the vertical layout is used, this should be more obvious to the respondent.
Reference: Bryman: Social Research Methods: 5th Edition Page(s) 227

One of the reasons for using self-completion questionnaires is to eliminate interviewer effects, so answer (b) must be wrong. It is the questions themselves that will influence measurement validity, rather than the instrument used to group the questions. Consequently, structured interviews are as likely to have measurement validity as a mailed questionnaire. Finally, since this type of questionnaire is typically coded in advance, it is hard to see how inter-coder variation might occur, whereas separate interviewers might possibly code respondent data differently to each other.
Reference: Bryman: Social Research Methods: 5th Edition Page(s) 221,222

However, bearing the previous question’s discussion in mind, when we present longer lists of items (like attitude statements, perhaps), we should set out the range of responses for each sub-element of the question horizontally. Following the rule of making our questions as clear and as unambiguous as we can, we find that Likert-scale questions work better when they are set out horizontally. Instructions must be given for the provision of responses for these, as for other types of question and it does not make a great deal of sense to have the questions section separated from the answers section for self-completion questionnaires.
Reference: Bryman: Social Research Methods: 5th Edition Page(s) 227,228

A Likert scale shows degrees of agreement or disagreement with statements and usually calls for responses along the range. When setting out the questions we might always have “strong agreement” appearing on the extreme left, with “strong disagreement” appearing on the extreme right. Some respondents develop a method of response such that they always show agreement, for example, labelled a ‘response set’. This forces us to phrase some questions negatively, so that “strong agreement” actually means “strong disagreement” and to code the questions accordingly, which we call ‘reverse scoring’. This means we must not do what answer (a) suggests and “undesirable” social behaviours can be included as freely as those more “desirable”. Explicit instructions must be given to respondents to help them provide the data we seek. We might use phrases like “there are no right or wrong answers, we are simply looking for your opinion” but exhortations about moral behaviour are best left out.
Reference: Bryman: Social Research Methods: 5th Edition Page(s) 228

The diary method can be used to gather qualitative data as well as quantitative data but it is not usually a good idea to mix these in a single diary. ‘Time-use’ is quantitative data, and the purpose of this type of structured diary is to find out how much time people spend on different activities day-by-day. (See “Plate 10.2” on page 240, in this regard.) Consequently answers (b) and (c) are inappropriate, although free-text diaries could be used to gather that kind of qualitative data. We don’t have enough time or space to comment on answer (d).
Reference: Bryman: Social Research Methods: 5th Edition Page(s) 238-240

“Research in focus 10.1”, on page 238, shows how a diary study can be used to gather data for social research. Both quantitative and qualitative data can be gathered using this method but it is not a form of questionnaire, so there are no “answers”, whether open or closed. Rather, people are selected (perhaps randomly) for a sample and asked to keep a diary of their activities over a period of time. Depending on how the concept has been operationalized, diaries are as likely to have measurement validity as any other data-gathering instrument. Corti (1993) thinks it worthwhile to see diaries as being either ‘structured’ or ‘free-text’ (p238). Furthermore, diarists should be given explicit instructions on how to keep the diaries and shown a model of a completed diary section (Corti, 1993, cited on page 239).
Reference: Bryman: Social Research Methods: 5th Edition Page(s) 238,239