In December 1999, NASA launched the EOS AM1 satellite. This platform carries 5 instruments whose purpose is to measure important properties of the Earth climate system. One of these instruments is MISR, the Multi- Angle Spectro-Radiometer. This instrument measures the light reflected from the Earth at high resolution (~275 m) at four wavelengths (three in the visible and one in the near IR) and at nine different viewing angles that vary from +70 to -70 degrees along the direction of flight of the satellite. This multi-angle data has the potential to provide information on aerosols, surface and cloud characteristics which compliments traditional single- view-direction satellite measurements. Before this potential can be realized, the accuracy of the satellite radiance measurements must be carefully assessed and the implications of the radiometric accuracy on the remote sensing algorithms evaluated. In this presentation, we show a comparison of the MISR multi-angle measurements and 2D radiative transfer simulations from an inhomogeneous cloud scene. The inputs to the radiative transfer code are based entirely on independently gathered data (ground-based radar, lidar, microwave radiometer, in situ aircraft data, etc.). It is found that the 2D radiative transfer solution compares favorably in the forward scattering directions, but is off by as much as 10% in the backscattering directions. Using 3D radiative transfer modeling, we show that this difference is due to the cloud 3D structure. Comparison of the simulations to the MISR measurements, after accounting for 3D effects, show residual differences which are less than 4% at all angles.