1.5.3 Data collection

Questionnaires

The term questionnaires can cover a fairly broad range of possible research techniques. For the purposes of this research it means structured surveys delivered remotely and completed without the researcher's input or presence. Such surveys have a number of advantages [Mangione, 1998]:

• relatively low cost
• quick coverage of large numbers of dispersed respondents
• convenience and privacy for respondents
• insulation of respondents from interviewer expectations

These were all seen as advantages for this research topic. Questions of cost and coverage were particularly relevant, but it was seen as important for the attitudes of the researcher not to influence the answers of the respondents.

Surveys are a particularly good choice when:

"(a) you have limited human resources to help you conduct your study, (b) your questions are written in a closed-ended style, (c) your research sample has a moderate to high investment in the topic, and (d) your list of research objectives is modest in length." [Mangione, 1998, pp. 399-400].

All of these conditions obtained for this research. The resources for the study were largely limited to the researcher. The questions (see 7.2: Survey process on page 103) were mostly closed, with a few opportunities for additions or freeform comments. Both research sub-samples were judged to have at least a moderate interest in the topic. The list of research objectics was capable of being fulfilled with a two-page (albeit information-dense) survey instrument.

Questionnaires do have some areas where problems may occur, affecting the quality of the data and therefore the conclusions that can be drawn. The surveys for this thesis sought to minimise such problems by practising "total survey design" [Biemer et al., 1991]. The four broad error types to be avoided are sample selection bias, nonresponse error, item nonresponse error and response error. The ways in which the surveys in this study sought to avoid these errors are discussed in detail in the chapter dealing with the surveys and their results.

Interviews

Interviews were one of the main techniques used in the case-study phase of the research (see 8: Library Case Studies on page 145). These interviews were with senior staff in the library projects selected and followed a semi-structured pattern. Such semi-structured interviews are an appropriate technique for the conduct of descriptive exploratory research in a new area of enquiry [Bickman et al., 1998]. This is because they provide a consistent framework for the questions being asked while still allowing the flexibility to follow particular lines of enquiry if this seems relevant or appropriate. Because of the open-ended nature of the interviews, the respondents need to be considered as informants only, providing verbal reports about their version of events. Where possible, these verbal reports should be checked by triangulation with other informants and by corroboration with other sources of evidence (i.e. documentary evidence) [Yin, 1998]. The specific interview questions are included at the end of this thesis (see 11.3: Library Case-Study Questions on page 191).

Other techniques

Other possible techniques that could have been adopted include direct observation, documents of the past and evaluation research. Direct observation and evaluation research were not used because of the problems of limited human resources and costs. Documentary research was used as part of the case-study phase, primarily to provide background information on the projects and to corroborate information provided by informants.