You are developing the database schema for a SQL Server Analysis Services (SSAS) BI Semantic Model (BISM). The BISM will be based on the schema displayed in the following graphic. You have the…

A many-to-many dimensional relationship in a multidimensional project uses a bridge measure group to correctly route aggregations across both dimension tables, which is necessary when a student can belong to multiple classes and the average must not double-count. Multidimensional models natively support role-based security at the dimension and cell level, as well as measure group partitions, satisfying all stated requirements with standard, wizard-driven tooling that minimizes development effort.

A reference relationship routes dimension queries through an intermediate dimension rather than through a bridge fact table, so it cannot correctly resolve many-to-many aggregations and would produce inflated or incorrect averages for student age by class.

A tabular project with standard measures does not natively resolve many-to-many join semantics without explicit DAX bidirectional filter configuration, requiring additional development effort to produce correct averages.

Calculated columns compute row-level scalar values during data refresh and do not perform aggregation across relationships, making them unsuitable for computing an average student age aggregated across a many-to-many class enrollment relationship.