User Attitudes towards Pattern-Based Enterprise Information Models: A Replicated Experiment with REA Diagrams

This paper presents a study evaluating the REA-patterns-based approach to enterprise information modelling. We replicated an earlier experiment demonstrating that business students perform a comprehension task more accurately when they recognize REA pattern occurrences in entityrelationship diagrams and that they perceive such diagrams as easier to use than informationally equivalent diagrams with no apparent REA pattern occurrences. These findings were confirmed in the replication, but contrary to the original experiment also efficiency gains in terms of faster understanding were demonstrated. Furthermore, the original experiment was extended by testing a more complete model of user attitudes. It was shown that, after completing the comprehension task, students perceived diagrams with REA pattern occurrences as easier to interpret and more useful, and were more satisfied with them. But somewhat surprisingly, students also perceived such diagrams as more semantically expressive, even if they were informationally equivalent to the control diagrams. It was further shown that perceived semantic expressiveness is related to perceived usefulness, but not to perceived ease of interpretation. The effect of perceived semantic expressiveness on user information satisfaction is not direct, but mediated by perceived usefulness.