No / Purpose To determine the test-retest reproducibility of accommodation measurements gathered in an unselected sample of primary school children. Methods Monocular and binocular amplitudes of accommodation (AA) were collected by five different Testers using the push-up method in an unselected sample of school children (n=137, age: 8.1±2.1-years). Testing was conducted on three occasions (average testing interval: 8-days) in 91.2% of the children. Results The median AA was 19.1D, the variation due to the identity of the Tester was 3.1D (p<0.001) and the within-subject variation (which takes the variation due to Tester identity into account) was 5.2D. Around 75-79% of children exhibited monocular AAs-12D when tested on the first occasion, but more than 90% exhibited an AA-12D when subsequently tested. Around 74-80% of those with an AA<12D on the first occasion had values-12D on subsequent testing even though no treatment had been undertaken. Poorer initial AA measurements were less likely to improve on repeat testing. Conclusions Our results reveal substantial intra-individual variation in AA measurements, raising questions about the usefulness of this test in children aged 4-12-years. We suggest that AA assessment may prove most useful in children in this age range as a pass/fail check for substantially reduced AA, for example, where the AA is <12D. Our sample would suggest that the prevalence of persistently reduced AA may be around 3.2% when tested under binocular conditions and 4-6.4% when tested monocularly.