This is a review of the effect of congenital ptosis surgery on refractive error. Seventy-one eyes of 56 patients were reviewed. The unoperated eye in unilateral cases was used as the control group. The mean age at surgery was 5.1 years. Forty-eight eyes had frontalis slings, 20 eyes levator advancements including 11 with tarsal resection, and three Muller’s muscle resection. The mean time between surgery and final refraction was 34 months. Prior to surgery 32 of the 56 patients had strabismus, amblyopia or anisometropia. The mean spherical equivalent was 1.64D in the operated eyes and 1.0D in the control eyes, with average astigmatic errors of 0.77D and 0.37D respectively. Postoperatively the spherical equivalent decreased in operated eyes by 0.82D, but the cylinder increased by 0.4D, compared to a decrease of 0.41D in spherical equivalent and an increase in 0.3D in cylinder in the control eyes. There was no statistically significant difference in change in sphere or cylinder as a result of surgery comparing operated and control eyes. Neither was there an association between amount of ptosis and the refractive change. Interestingly there was a significant increase overall in the amount of astigmatism in the operated eyes, but no difference in the number of patients with amblyopia or anisometropia. This is a large and useful study, with sufficient data for the authors to conclude that contrary to previous recommendations, ptosis surgery should not be performed solely to reduce the risk of anisometropic amblyopia. As far as vision is concerned, with the exception of deprivational amblyopia, congenital ptosis surgery can be safely delayed until the patient is older.