Critical evaluation of conceptual data models

Abstract Conceptual data models are representations of enterprise databases. Information systems professionals must often critically evaluate conceptual models. After system designers create conceptual models, they or members of their design team corroborate the created models. System auditors also evaluate conceptual models. Little guidance exists to advise system designers and auditors on how to conduct such conceptual model validation. Mistakes in corroborating conceptual models may result in unnecessary changes to valid systems or in false assurance of invalid systems. Evidence from prior studies suggests that evaluators exhibit a cognitive bias that causes them to focus on structural constraints even when there is conflicting information in the surface semantics. Studies also show mixed results for whether relationships involving optional participation of entities are more difficult for users to comprehend than relationships with only mandatory participation. Decompositions of large representations may be easier to evaluate than are the large representations in their entirety. Because conceptual models often are portrayed as large representations that include relationships with optional participation of entities, these concerns must be investigated in the context of system evaluation. This study provides evidence of a debiasing technique for the tendency to focus only on structural constraints, and compares evaluation responses for entities’ optional and mandatory participation in relationships presented in decomposed versus full models.