Novice errors in conceptual database design

Abstract

Conceptual and logical database modelling are difficult tasks for designers, and the potential for committing and correcting errors is significant. This paper reports on two laboratory experiments that investigated the underlying causes of errors committed by novice designers engaged in conceptual database modelling tasks. These causes can be traced to combinatorial complexity of the task, biases resulting from misapplication of heuristics, and incomplete knowledge about database design. The most common error was that subjects translated their initial understanding of the application into final database structures and did not consider alternative hypotheses and solutions. The paper includes recommendations to reduce the occurrence of errors.