An evaluation of the effectiveness of interview techniques in the elicitation of tacit knowledge for requirements engineering in small software projects.

摘要

In the business domain, interviewing is the requirements elicitation technique of choice for practitioners participating in the development of software solutions. Interviewing is used for reasons such as its simplicity to implement, its familiarity with participants, and the ability to conduct more thorough analysis sessions. It is widely acknowledged that during interviews, experts experience difficulty with articulating their tacit knowledge. This can affect the interview technique. This dissertation examines the effectiveness of interview techniques in eliciting user requirements when knowledge is tacit. A framework is proposed to classify interview techniques using two dimensions developed in this study---specificity and structure. The framework is tested using a survey of qualified practitioners responsible for projects with varying levels of complexity that are 400 hours or less of effort. Multivariate statistical tests are performed to assess the effect of interview dimensions (specificity and structure) after accounting for requirements complexity and human factors. Results suggests that the two proposed dimensions, specificity and structure, do not have a strong effect on effectiveness in general. Further analysis is performed after taking into account the low sample sizes and the use of non-validated scales. This results in preliminary evidence that when there is more structure and less specificity, interview techniques trend towards a higher level of effectiveness. Overall, the dimensions of structure and specificity do not appear to significantly impact the effectiveness of interviewing techniques. The findings of this study are sufficient to encourage further work on the research model. Several suggestions are made for future research.