Multi-look processing provides a straightforward method for enhancing the quality of grainy (speckle-filled) synthetic aperture radar (SAR) imagery. This improvement in quality results from a reduction in variability of the individual pixel values brought about by non-coherent averaging of multiple, statistically independent views of the same scene. That is, the variance of the average of independent, identically distributed random variables is inversely proportional to the number of independent looks that are averaged. Of course, the quality of the averaged image depends heavily on the accuracy of the algorithm used to align the individual looks prior to averaging. The Army Research Laboratory (ARL) has recently applied multi-look processing techniques to sets of images from independent views to generate look-averaged images. In the course of this processing, we have identified the need to quantify any degradation in multi-look image quality resulting from mismatches in registration of the individual looks used to form the look-averaged image. Since these degradations can affect the performance of automatic target detection algorithms, we have also identified the need to quantify the effects of look-averaged image degradation on the performance of such algorithms. In this paper, we use X-Band SAR data to determine the extent to which translational and rotational errors in the registration of individual looks degrade the quality of the resulting multi-look averaged image. We then examine how this degradation in image quality impacts the automatic detection of small targets in the final, look-averaged image. Finally, we introduce variations in target-to-clutter ratio within each of the individual looks and analyze how these changes affect the resulting look-averaged image.