Abstract: Reinforced concrete columns with widely spaced and poorly configured transverse reinforcement, also known as shear-critical columns, have been used in buildings constructed in regions of high seismicity. Earthquake reconnaissance and experimental research show that these columns are vulnerable to strength degradation under earthquake motions, and this can lead to shear and axial failures and collapse. An experimental program was designed to study the dynamic behavior of simple reinforced concrete structures having differing degrees of strength degradation associated with column damage. Each structure was configured as a planar frame supported by two one-third scale columns. The columns were either ductile columns, non-ductile columns, or a combination of one ductile column and one non-ductile column. Column axial loads were set at either 0.1 f_c'A_g or 0.24f_c'A_g , where f_c' is the target concrete compressive strength and A_g is the column gross cross-sectional area. A total of twelve of these structures was tested in the laboratory on a shaking table with two types of ground motions (long-duration, low-velocity motion or short-duration, high-velocity motion). Observed behavior was strongly affected by column details, axial loads, and input base motions. Capabilities of detailed models and simplified single-degree-of-freedom models to reproduced measured behavior is studied. The study concludes that dynamic response is strongly sensitive to details of the analytical model, and that some strength-degrading models reproduce results better than others.