The aim of the study was to evaluate two essentially different methods of assessing differences in children's taste preferences with regard to five different paediatric penicillin formulations. The study was performed with a parallel group design comparing five groups. A taste evaluation was recorded in 103 children with upper respiratory tract infections after a single therapeutic test dose. First the patient's own spontaneous verbal judgement was recorded then a judgement was arrived at using a hedonic scale of facial expressions. Both judgements were passed immediately and also 3-4 minutes after the test dose. In the children who were 6 years of age or younger, a better discrimination of taste differences between formulations was achieved by using the patients' own spontaneous verbal judgements instead of the facial hedonic method. Both methods seem appropriate in older children, but the hedonic scale is preferable since its use implies a more standardised procedure.