•  
  •  
 

Abstract

Research has extensively documented the importance of accurate system requirements in avoiding project delays, cost overruns, and system malfunctions. Requirement elicitation (RE) is a critical step in determining system requirements. While much research on RE has emerged, a deeper understanding of three aspects could help significantly improve RE: 1) insights about the role and impacts of support tools in the RE process, 2) the impact of using support tools in multiple stages of the RE process, and 3) a clear focus on the multiplicity of perspectives in assessing RE outcomes. To understand how using support tools could improve RE, we rely on the theoretical lens of mental models (MM) to develop a dynamic conceptual model and argue that analysts form mental models (MMs) of the system during RE and these MMs impact their outcome performance. We posit that one can enhance analysts’ MMs by using a knowledge-based repository (KBR) of components and services embodying domain knowledge specific to the target application during two key stages of RE, which results in improved RE outcomes. We measured the RE outcomes from user and analyst perspectives. The knowledge-based component repository we used in this research (which we developed in collaboration with a multi-national company) focused on insurance claim processing. The repository served as the support tool in RE in a multi-period lab experiment with multiple teams of analysts. The results supported the conceptualized model and showed the significant impacts of such tools in supporting analysts and their performance outcomes at two stages of RE. This work makes multiple contributions: it offers a theoretical framework for understanding and enhancing the RE process, develops measures for analysts’ mental models and RE performance outcomes, and shows the process by which one can improve analysts’ RE performance through access to a KBR of components at two key stages of the RE process.

Share

COinS