Reflection on Development Techniques Using the Psychology Literature: Over Two Decades of Bias and Conceptual Blocks

URL:

Abstract:

Analysis and development techniques have played an important role in information systems, providing support for developers in structuring and directing tasks. They also provide cognitive support in collecting, collating, analyzing, and representing information about system requirements and attributes. However, by developing previous work further, in particular by classifying techniques into six generic types and transcribing these onto problem/solution space diagrams, this paper argues that by directing tasks and dictating what and how information is collected and represented, techniques can bias developers’ understanding of system requirements and attributes. The 1984 IFIPWG8.2 conference showed how IS can be informed by literature in our foundation disciplines. By drawing on the psychology literature to develop a classification of techniques, this paper shows some of the potential biases inherent in techniques. The classification is applied to many techniques which have contributed to development activity. Through an understanding of the conceptual blocks embedded in them, the paper hopes to inform practice about the selection and mixing of development techniques. More generally, the paper suggests a reexamination of our assumptions when undertaking IS development. The techniques that limit problem and solution scope most and also provide the most cognitive and conceptual biases are the more formal, objective ones, and we recommend that less formal techniques are also used in practice. Parallels can be drawn with the movement encouraging the use of qualitative research approaches in IS research inspired by the 1984 conference. Research methods and techniques also provide support in the form of directed tasks, activities, guidance on data collection, analysis, and representation. Drawing on this parallel, it could be argued that quantitative, formal research approaches may introduce more conceptual biases than less formal qualitative, approaches, and that the latter should be used instead of, or at least alongside, quantitative approaches.

Attachment	Size
Adams.pdf	302.69 KB