Abstract This study aimed to determine the accuracy (and usability) of the Retinomax, a hand-held autorefractor, compared to measurements taken from hand-held retinoscopy (HHR) in a sample of normal 1-year-old children. The study was a method comparison set at four Community Child Health Clinics. Infants (n = 2079) of approximately 1 year of age were identified from birth/immunization records and their caregivers were contacted by mail. A total of 327 infants ranging in age from 46 weeks to 81 weeks (mean 61 weeks) participated in the study. The children underwent a full ophthalmic examination. Under cycloplegia, refraction was measured in each eye by streak retinoscopy (HHR) and then re-measured using the Retinomax autorefractor. Sphere, cylinder, axis of cylinder and spherical equivalent measurements were recorded for HHR and Retinomax instruments, and compared. Across the range of refractive errors measured, there was generally close agreement between the two examination methods, although the Retinomax consistently read around 0.3 D less hyperopic than HHR. Significantly more girls (72 infants, 47.7%), struggled during examination with the Retinomax than boys (52 infants, 29.5%) (P < 0.001). Agreement deteriorated between the two instruments if the patient struggled during the examination (P < 0.001). In general, the Retinomax would appear to be a useful screening instrument in early childhood. However, patient cooperation affects the accuracy of results and is an important consideration in determining whether this screening instrument should be adopted for measuring refractive errors in early infancy. [ABSTRACT FROM AUTHOR]