A number of factors affect the accuracy of aerosol retrievals from satellite imaging radiometers, including algorithm assumptions, the quality of the associated cloud masks, the prescribed aerosol optical and microphysical models, and calibration uncertainties. In this paper, we highlight a concerted effort by the Terra Multi-angle Imaging SpectroRadiometer (MISR) team to evaluate the accuracy and stability of the instrument's radiometric calibration, with the twofold objective of (1) making improvements in the absolute and relative calibration where supported by multiple lines of evidence, and (2) evaluating the effect of those calibration refinements on aerosol retrievals. Aspects of the instrument's on-board calibrator design, including careful pre-flight handling of the Spectralon diffusers and the novel use of detector-based standards, have contributed to excellent long-term radiometric stability. In addition, multiple methodologies, including comparisons with other Terra sensors, in-flight and laboratory tests involving AirMISR (the airborne counterpart to MISR), lunar observations, camera-to-camera radiometric comparisons at specialized viewing geometries, and investigations using surface-based radiometer data over dark water sites have provided a detailed picture of radiometric performance at the low light levels typical of a large fraction of global aerosol observations. We examine the sensitivity of aerosol property retrievals to small band-to-band and camera-to-camera calibration adjustments, and demonstrate the importance of calibration in meeting climate-quality accuracy requirements. Because combining downward-looking (satellite-based) and upward-looking (surface-based) radiometers can constrain the optical properties of an aerosol column to a greater extent than possible from either vantage point by itself, achieving radiometric consistency, or “closure” between them is essential to establishing a long-term aerosol/climate observing system.