User-database interaction can be classified into various abstraction levels. The two higher levels are the conceptual level and the logical level. On the conceptual level, the user talks of things in his world, for example, in terms of entities and relationships. On the logical level, the user talks of things in the database system, for example, in terms of relations and join operations. The user world is something that the user is expected to know, but the data in the database system is something new to the user. Thus, the conceptual level is closer to the user. Therefore, we hypothesized that the conceptual level interaction will be easier for the user. This was tested with the entity-relationship (ER) model for the conceptual level model and the relational model for the logical level model. The results showed that users of the conceptual level had 38% higher accuracy, 16% higher confidence, and took only 35% of the time taken by users of the logical level. The differences were statistically significant with p values better than 0.003. From the user's perspective, the huge differences make it worthwhile to consider the ER model in place of the relational model.
Chan, H. C.; Wei, K. K.; and Siau, K. L., "CONCEPTUAL LEVEL VERSUS LOGICAL LEVEL USER-DATABASE INTERACTION" (1991). ICIS 1991 Proceedings. 45.