Shake table experiments are conducted to support the selection of performance criteria and to verify the inelastic modeling approach for developing the fragility functions of reinforced concrete buildings. Two frames representing the lateral force-resisting system of a preseismic code building are tested under the effect of an earthquake record with increasing severity. Shear failure is detected in columns at a PGA of 1.28g before other failure modes, which was effectively predicted by the fiber-based numerical model, performance criteria, and shear supply approaches adopted for vulnerability assessment. Five buildings, ranging from 2 to 40 stories, are then assessed under the effect of far-field and near-source earthquake records, considering the experimentally verified modeling approach and shear failure prediction models that account for flexural ductility and shear-axial force interaction. The impact of considering shear response on the vulnerability assessment results is considerable, particularly for the lower-height wall structures when subjected to the near-source earthquake scenario. Higher modes dominate the behavior of wall structures, principally under the latter seismic scenario, and shift their response to shear-controlled. Therefore, seismic scenario-structure-based performance criteria are adopted for developing a range of analytically derived, experimentally verified fragility functions for the earthquake loss estimation of buildings with different characteristics.