Sensory scientists have adapted several sensory methods to fit children's cognitive abilities. Although several discrimination methods have been reported with children, the A-not-A test has not been studied yet. The aims of this work were to: (i) evaluate the feasibility of using the A-Not-A test with school-aged children, and (ii) compare how the framing of the question (overall differences vs. differences in liking) may influence the results. A total of 126 children were involved in the study. They participated in three sessions, each composed of a familiarization task with a visual stimuli and sample tasting with one of three dairy products (vanilla milk desserts, chocolate-flavored milk, and vanilla-flavored yogurt). Half of the children evaluated the samples in terms of overall differences and the other half in terms of differences in liking. Results from the familiarization step showed that children correctly identified the visual reference in the A-not-A test regardless of how the question was framed, suggesting that they were able to understand the methodology. In the case of tasting samples, children were significantly more likely to correctly identify the reference in two of the three studied dairy products, when the question was based on liking, as compared to the framing based on overall differences.