The object of this study was to compare the discrimination scores obtained in conventional speech audiometry with the scores recorded when the same speech test lists were presented through individually fitted, ear-level hearing aids. Comparative tests were performed in 50 hearing-impaired students aged 8 to 20 years, all well trained everyday hearing aid users. The test material consisted of Swedish standardized 50 words PB lists, recorded on tape. To keep the test conditions under control in the hearing aid presentation, the recorded speech was transmitted from a neck-loop telecoil to the student's hearing aid (T-mode listening). The result shows that the great majority of children recorded considerably higher discrimination scores when the tests were performed with their individual hearing aids than with the test lists presented through the audiometer and the TDH-49 earphone. The conclusion is that, in children who are using carefully fitted hearing aids and are well adapted to those, the results of traditional speech audiometry tests involve a risk of underrating such children's practical hearing ability. The finding stresses that it is important to maintain one and the same amplification characteristic to which the child has accommodated, and to avoid switching between systems with different characteristics.