We examined changes in Minnesota Multiphasic Personality Inventory scores from adolescence to adulthood in a longitudinal study of 540 men who attended college during the Vietnam War. Using change scores that were adjusted for initial values, we compared civilians to veterans who were grouped according to combat exposure: none, peripheral, or direct. In cross-sectional analyses, the groups differed only as adults. Groups were similar in relative stability but differed by multivariate analysis in absolute change on the clinical scales. Only veterans with peripheral exposure differed from civilians in multivariate contrasts, even after controlling for premilitary variables. Effect sizes were small. Results suggest that combat exposure does not produce uniformly negative outcomes and may have positive effects in select populations.