Vision screening was performed on 268 pre-school children: 170 from a private pre-school, 33 from a Caribbean-American parochial pre-school and 65 pre-school children from a clinic serving indigent Spanish farm-workers. Using a multi-station format, a stereoacuity test and two visual acuity tests were performed during a single screening session. The time it took to complete a test was recorded. To pass the screening, children were required to pass one visual acuity test and the stereoacuity test. Children who could not complete the protocol were retested at a later date. Children who failed the screening and every fourth child who passed the screening were referred for a full eye examination. The parents and teachers were masked to the results of the screening as well as the optometrists who performed the eye examination. Sensitivity, specificity and accuracy scores were 100%, 79% and 80%, respectively. Three-year-old children completed the Lea Symbol Chart more often than the HOTV. No differences in time required to complete a visual acuity test were found. The Lea Symbol chart is more likely to be completed by young children. Testability changes significantly with age rather than with the instrument when socio-ethnic factors are held constant. Differences among groups and the sensitivity of the screening are discussed.