Educating Conceptual Data Modelers: How to Avoid Misunderstanding Requirements

Abstract

In this dissertation, which is comprised of three related but independent essays, I investigate the application of grounded theory methodology to conceptual data modeling. Although conceptual data modeling, largely focused on entity-relationship diagrams, has been a topic of interest to researchers over the past three decades, the extent to which the existing body of research reflects substantial and cumulative development on how to improve cognitive understanding during modeling activity is not entirely clear. My dissertation tries to address this issue by intensively focusing on application of one of commonly adopted qualitative research methodologies to conceptual data modeling. The first essay aims to clarify why and how the methodological approach from grounded theory research is capable of improving information systems development. Information systems development has been regarded as a crucial investment to achieve business strategic goals. Many development methodologies, project management tactics have been suggested, whereas, failure rates are still beyond rational expectation. The most serious problem is that users come to realize that the system is not reflecting requirements when tangible, workable system applications are finally presented. In this essay, I propose that possibility of misunderstanding user requirements need to be seriously considered. Based on literature reviews on information systems failures and requirement articulation techniques, I found that approaches on analyzing problem domains overly focused on understanding things, and possibility of misunderstanding was not sufficiently considered. As a countermeasure, I reviewed the grounded theory approach, which is one of the famous, conservative qualitative research perspectives to increase understandability and decrease misunderstanding possibility. The essay concludes with a research agenda on how to apply the grounded theory approach to help information system engineers. The second essay evaluates the applicability of the grounded theory approach (i.e., analysis strategies) by conducting small laboratory examination. Conceptual modeling is a critical activity for developing successful business information systems. The objective of this study is to evaluate the possibility of applying the constant comparison method from the grounded theory to conceptual modeling. To achieve the objective, I trained novice modelers and split them into two groups for evaluation. The experimental results show that applying the constant comparison method could increase acceptability from more experienced conceptual modelers. Moreover, while the control group was experienced difficulties when domain knowledge is unfamiliar, the experimental group could handle difficulties more effectively. In addition, applying the constant comparison method also decreased the time to complete analysis for conceptual modeling. The final essay develops a new design concept, concept magnifying, for improving conceptual data modeling process based on education and learning theories. Grounded theory is a social science theory that is rooted in empirical data. It should be noted that analytic method of grounded theory research is not intended to be used in requirement analysis or system development. In addition, training system engineers or data modelers as skillful grounded theory researchers may not be practical. In this essay, two design concepts, fortifying coherence and developing examples, are coined as a design concept – i.e., concept magnifying – based on education and learning theories. Fortifying coherence and developing examples are highly relevant to the analytic strategies of grounded theory research, which is known as constant comparison of coding and analysis. The experiment showed that implementing the suggested design concept was likely to increase performance of conceptual modeling. The finding highlights an important benefit of my approach in that emulating constant comparison of grounded theory yields to positive effects. The result can lead to promising new information technology artifacts to support conceptual data modelers.