Quantifying the Impact of EER Modeling on Relational Database Success: An Experimental Investigation

Abstract

Despite the widespread idea in literature that the inclusion of EER modeling in the design process of a relational database is beneficial for the success of that database, almost no quantitative cost-benefit analyses of EER modeling exist today to support this statement. In order to fill this need, an empirical study is performed in which the success of a relational database of which the design process contains an EER modeling phase is compared to the success of a relational database in which only the minimally needed design effort was put. Hereby, database success is treated as originally proposed by the DeLone and McLean Information Systems Success Model, by specifically focusing on the information quality and system quality of both databases. To this end, respectively, the total amount of time that is needed by an end user to complete a set of tasks by using the database, and the total execution cost that is needed by the database system before a correct solution to each task is submitted, is analyzed. Moreover, the work accounts for the possible moderation of the technical competence of an end user in the relationship between EER modeling and the success of the eventual relational database. Preliminary results indicate that the inclusion of EER modeling in relational database design significantly highers the perceived information quality and system quality of that database. Moreover, there is statistical evidence that this result is independent of the competence profile of that user.