To identify and quantify sources of error in the refractive outcome of cataract surgery. AMO Groningen BV, Groningen, The Netherlands. Means and standard deviations (SDs) of parameters that influence refractive outcomes were taken or derived from the published literature to the extent available. To evaluate their influence on refraction, thick-lens ray tracing that allowed for asphericity was used. The numerical partial derivative of each parameter with respect to spectacle refraction was calculated. The product of the partial derivative and the SD for a parameter equates to its SD, expressed as spectacle diopters, which squared is the variance. The error contribution of a parameter is its variance relative to the sum of the variances of all parameters. Preoperative estimation of postoperative intraocular lens (IOL) position, postoperative refraction determination, and preoperative axial length (AL) measurement were the largest contributors of error (35%, 27%, and 17%, respectively), with a mean absolute error (MAE) of 0.6 diopter (D) for an eye of average dimensions. Pupil size variation in the population accounted for 8% of the error, and variability in IOL power, 1%. Improvement in refractive outcome requires better methods for predicting the postoperative IOL position. Measuring AL by partial coherence interferometry may be of benefit. Autorefraction increases precision in outcome measurement. Reducing these 3 major error sources with means available today reduces the MAE to 0.4 D. Using IOLs that compensate for the spherical aberration of the cornea would eliminate the influence of pupil size. Further improvement would require measuring the asphericity of the anterior surface and radius of the posterior surface of the cornea.