A clinical study was done to determine the bias and accuracy of retinoscopic examinations. Ten subjects were examined by five ophthalmologists on two separate occasions. Retinoscopic measurements of sphere, cylinder, and axis were compared to measurements made by subjective techniques. Statistical analysis of the results uncovered several types of bias. Retinoscopy underestimated myopia and overestimated hyperopia. It also overestimated the size of cylindrical errors. The overall accuracy of retinoscopy was quite good, better for cylinders than for spheres. There were no significant right eye versus left eye differences in regard to accuracy or bias. There were significant differences in accuracy among the ophthalmologists, but they were not clearly related to the previously noted differences in precision.
Hyams L, Safir A, Philpot J. Studies in RefractionII. Bias and Accuracy of Retinoscopy. Arch Ophthalmol. 1971;85(1):33-41. doi:10.1001/archopht.1971.00990050035006