Estimates of patient satisfaction are increasingly used as an outcome measure in the evaluation of health care. However, surveys are based on untested and incorrect assumptions about what a patient is reporting. In this paper some of these general concerns are presented. This is followed by results from a study on satisfaction from 195 patlents attending a low vision service. Results indicate complex patterns of association with four different estimations of satisfaction from both visual and non-visual factors. Different combinations of visual and non-visual factors had different predictive values for the satisfaction scales. The results are in line with more general findings on satisfaction and suggest that satisfaction is a multi-faceted concept which may easily be misinterpreted in a typical satisfaction survey.