An experiment was carried out in which untrained observers estimated the speed of vehicles passing them, the true speeds being measured with a radar meter. The variability of the subjects' estimates was a more important source of error-the standard deviation of estimated minus true speed was about 5 mile/h for all subjects. There was a tendency for observers to underestimate high speeds. The interpretation of the results leads to an intriguing statistical paradox which is explained in this report.
Abstract