AIMS/BACKGROUND: The intentions of this study were to estimate agreement between two different autorefractors and standard subjective refraction techniques and to evaluate the clinical implications of relying on the autorefractor measurements. METHODS: Subjective refraction was carried out on 448 cycloplegic eyes and compared with cycloplegic readings with the Allergan Humphrey 500 autorefractor (448 eyes) and the Nidek AR-1000 autorefractor (160 eyes). Each refraction was followed by clinical visual acuity measurement. The study population comprised 224 healthy students, 107 men and 117 women, with a mean age of 20.6 (SD 1.1) years. RESULTS: Both the Nidek and Humphrey autorefractors measured more negative or less positive refractive values compared with subjective refraction and these biases were statistically significant (Humphrey right eye -0.23 D, p = 0.0001, left eye -0.20 D, p = 0.0001), (Nidek right eye -0.13 D, p = 0.0001, left eye -0.11 D, p = 0.0002). Comparing the results of autorefraction with subjective refraction, the Nidek was better than the Humphrey autorefractor in several ways: a smaller mean difference, better agreement between spherical equivalent values, narrower limits of agreements, and better visual acuity obtained with the autorefraction. On the other hand, the Humphrey autorefractor agreed better with subjective refraction concerning cylinder axis. CONCLUSION: The results show that both autorefractors represent a valuable complement to subjective refraction, but cannot be used as a replacement.