Purpose To establish standardised protocols for vision screening, testability and comparability of three different vision tests were examined in a population-based, cross-sectional sample of preschool children (Sydney Paediatric Eye Disease Study). Methods Measurement of presenting monocular distance visual acuity (VA) using the Amblyopia Treatment Study (ATS) HOTV protocol, was attempted by all (1774) children agedZ24 months. In addition, in children agedZ60 months (576), VA was also tested using the logMAR retro-illuminated HOTV or Early Treatment Diabetic Retinopathy Study (ETDRS) linear charts (CSV 1000). Children able to have both eyes tested monocularly were considered. Results Testability significantly increased with age for all VA tests. The ATS HOTV with an overall testability of 80% (females: 82%, males: 78%) was the most testable of the VA tests (Po0.0001). In children aged o3 years testability was low (r47%) rising toZ80% in children agedZ3. In childrenZ60 months, testability was higher for the HOTV (94%) than the ETDRS (59%) chart. In those that did two VA tests, mean difference of the ATS HOTV compared with the HOTV(CSV) was À 0.1, and compared with ETDRS was À 0.12 (Po0.0001). Conclusions Children aged o3 years had poor VA testability, whereas those 3 years and above were highly testable using the ATS HOTV. The HOTV (CSV) retroilluminated test was appropriate for children aged 45 years, and may be possible in younger children with early educational exposure. When comparing VA measures using these tests, the higher VA attained using the ATS HOTV, needs to be taken into account.