COMPARING QUALITY OF DATA REPRESENTATIONS PRODUCED BY NONEXPERTS USING LOGICAL DATA STRUCTURE AND RELATIONAL DATA MODELS (DATABASE, APPLICATION, INFORMATION SYSTEMS, REQUIREMENTS, DETERMINATION)

The purpose of this study was to investigate differences in the quality of data representations produced by nonexperts using Logical Data Structure and Relational Data Model formalisms.;The study was an empirical investigation involving MIS MBA students. The subjects were asked to read a case, ask questions, and generate application data models which were then evaluated using an instrument developed by the researcher.;The main proposition of the study was that the dominant component of a data modeling formalism would lead subjects into the component-driven type of data modeling. This proposition was supported by the research results at the syntactical but not at the semantical level.;The main conclusion of the investigation is that data modeling formalism concepts alone are not enough to evoke the appropriate types of questions necessary in data modeling for producing higher quality of application data models. The research results suggest that for novices explicit and detailed methods of data modeling, including an indication of the types of questions asked during the modeling process, may be more important than the data modeling formalisms used.