The ARM Enhanced Shortwave Experiment (ARESE) II was conducted in spring 2000 to address unresolved issues about the absorption of solar radiation in the atmosphere in the presence of clouds. In this study, apparent atmospheric absorption derived from surface and aircraft measurements are compared to 3-D radiative transfer model computations. Three cloud fields are examined with the heaviest overcast condition of 29 March being the most reliable from the standpoint of cloud structure and sampling. For this cloud field the model underestimates absorption by 18 W m-2 (9%) when compared to the Kipp and Zonen CM-22s and by 35 W m-2 (18%) compared to the Scripps Institution for Oceanography Radiation Measurement System. By including aerosols with modest absorbing properties into the model computations, the discrepancy is isolated to the near-infrared spectral region (0.68--3.9 ¿m). A comprehensive sensitivity analysis demonstrates that to match simultaneously the observed absorptance, transmittance, and cloud albedo with theory, cloud droplets must have optical properties similar to droplets with sizes three to four times as large as inferred in this study's cloud retrievals.