Clinical use of the Millon Adolescent Personality Inventory (MAPI) depends on computerized interpretation of the pattern of scale score elevations in the profile beyond certain cutoff scores rather than the elevations of single scales. There has been no reported work on the stability of the scale scores or the stability of the 2-point codes in a delinquent population. The MAPI was administered to 46 incarcerated male juvenile delinquents and was readministered after a period of 2 to 4 weeks. The test-retest correlations of the base-rate scale scores ranged from .33 to .89 with a mean of .74, which compare favorably to Millon's (1982) standardization sample. However, only 41% of the 2-point codes were judged to be congruent between administrations. The poor congruence of the 2-point codes across administrations raises doubts about interpretive statements based on these codes.