The overall aims of the study were to determine optimal methods and stimuli for eliciting mismatch negativity (MMN), extracting MMN from the deviant and standard waveforms, and identifying the response in children and adults. Several stimulus types were compared (pure tones, chords, and natural speech tokens) to determine which optimally elicit MMN. Deviant-alone and flip-flop MMN extraction methods that control for stimulus effects on MMN were compared for the speech stimuli (/da/ and /ga/). Visual identification, an area criterion, and integral- distribution techniques were used to identify MMN. Eight adults (20 to 28 years) and eight children (8 to 12 years) participated in the study. The deviant-alone method elicited bigger MMN area and duration than the flip-flop method for the speech stimuli. An area criterion of 110 μV × msec identified 90% of visually identified MMN compared to 62% identified using the integral-distribution technique. For both children and adults, speech stimuli and one of the chords most consistently elicited MMN.