Purpose: To describe the prevalence of numerical confusion errors in Ishihara testing in a representative sample of Australian children.
Design: Cross-sectional, population-based study.
Methods: The Sydney Myopia Study examined a representative stratified random cluster sample of 1,741 children (aged 6 years) attending 34 schools in Sydney, Australia (response rate, 78.9%), including color vision tests in 1,735 children. Those with any color vision defects (n = 33; 1.9%; 97% boys) were excluded. Responses for each Ishihara plate were recorded.
Results: Numerical confusion errors were made by 75.8% of children with normal color vision; there was no gender difference. Plates 3 and 7 were particularly prone to numerical confusion errors; 48.4% and 40.8%, respectively, of children misread these plates.
Conclusion: Numerical confusion errors in the Ishihara test were relatively common in children with normal color vision. The frequency of such errors could indicate an inherent deficiency in this commonly used test.