BACKGROUND: Dichotic listening (DL) to consonant-vowel syllables is frequently used in clinical and experimental studies of brain laterality. However, the paradigm of consonant-vowel syllables has not been thoroughly validated through a comparison with injections of amobarbital sodium (Amytal). OBJECTIVE: To validate the DL test for hemisphere dominance preoperatively vs postoperatively with the results from intracarotid injections of amobarbital (i.e., the Wada test) in epileptic children and adolescents. DESIGN AND MAIN OUTCOME MEASURES: Patients were tested with DL preoperatively and at 6-month follow-up. Correct reports in the DL tests were entered in a stepwise discriminant analysis for calculation of correct classification of hemisphere dominance with the results from the injections of amobarbital as the grouping variable. Correct reports from the right and left ears on the consonant-vowel DL test were compared preoperatively and postoperatively, separated for the subjects with regard to language dominance in the left and right hemispheres. SETTING: The Department of Pediatrics, Ostra Hospital, University of Göteborg, Göteborg, Sweden. PATIENTS: Thirteen children and adolescents between the ages of 10 and 19 years, who were surgically treated for resistant epilepsy, were included in the study. The operated area corresponded with morphological changes and functional dysfunctions according to findings from computed tomography, magnetic resonance imaging, single photon emission computed tomography, and electroencephalography. RESULTS: The results of the Wada tests revealed that 10 subjects had left hemisphere language dominance, with 3 subjects having right hemisphere language dominance. All 3 subjects with right hemisphere language dominance showed a left ear advantage on the DL test preoperatively and postoperatively, with 8 and 7 of the 10 subjects with left hemisphere dominance showing a right ear advantage, preoperatively and postoperatively, respectively. However, according to discriminant analysis, knowledge of the DL performance led to a correct classification according to the Wada test results in 12 (92%) of the 13 subjects. CONCLUSIONS: A quantitative classification procedure like discriminant analysis may be more sensitive when predicting hemisphere speech dominance from DL data than a qualitative procedure based on the ear advantage dichotomy. The ear advantage dichotomy may actually introduce arbitrary left-right categories that do not correspond to the actual clustering of the data.