PURPOSE: Few data exist regarding the upper limits of refractive error distributions in healthy infants; the data that do exist are biased because they were selected from the records of pediatric ophthalmology practices. We sought to obtain these data to validate examination failure criteria for vision screening. METHODS: We reviewed records from all children age birth to 5 years seen at the Tennessee Lions Eye Center at Vanderbilt Children's Hospital with a billing diagnosis of nasolacrimal duct obstruction and no comorbid ocular diagnoses except for refractive error. This was to avoid referral bias for any condition that could have influenced refractive error. All patients received a complete eye examination and cycloplegic refraction. Cumulative probability distribution (CPD) plots and means for spherical and cylindrical refractive error and anisometropia were prepared. RESULTS: One hundred thirty patients were studied; mean age was 15.5 +/- 9.9 months (range, 2 days to 66 months). The mean refractive error (spherical equivalent) was +1.4 D +/- 1.1 D. CPD plot analysis showed 95% of hyperopia to be < +3.25 D. Two children had myopia +1.00 D in one eye. CPD plot analysis showed 95% of astigmatism to be < +1.50 D and 95% of meridional anisometropia to be < 1.50 D. Six children had anisometropia >/=1.50 D, and 3 children had anisometropia > 3.00 D. CONCLUSIONS: At least 95% of children have hyperopia < +3.25 D, astigmatism < +1.50 D, and anisometropia < 1.50 D. This information will prove useful in identifying the natural history and prevalence of amblyogenic factors identified during preschool vision screening.