This study explores the differences between conceptual data models designed by expert and novice data modelling practitioners. The data models are evaluated using a number of quality factors synthesised from previous empirical studies and frameworks for quality in conceptual modelling. This study extends previous studies by using practitioners as participants and using a number of different quality factors in the evaluation. The study found that data models produced by expert data modellers are more correct, complete, innovative and flexible than those produced by novices. The results suggest that further research into the aspects of expertise that lead to such differences and how training courses can narrow the gap between expert and novice performance is required.