Prior research with children generally supports the two-dimensional structure of Attention-Deficit/Hyperactivity Disorder (ADHD; inattentive and hyperactive/impulsive factors) of the DSM-IV-TR as well as invariance of the two-factor structure across nations and cultures. Research with adults supports either a two-factor or three-factor structure depending on reporting source and breadth of symptoms assessed. However, research with adults is limited and there are few studies addressing cross-national invariance in adults. The purposes of this study were to (1) assess relative fit of two- versus three-factor solutions for self-report of childhood and recent ADHD symptoms in adults; and (2) further establish cross-national invariance of factors. Participants included 271 U.S. and 712 Japanese university students who completed a rating scale assessing the 18 DSM-IV-TR ADHD symptoms. Confirmatory factor analysis using Mplus (Version 6) and the mean and variance-adjusted weighted least squares (WLSMV) procedure showed invariance of two- and three-factor models across U.S. and Japanese samples. The two- and three-factor models showed similar fit indices. Neither a two-factor or three-factor model was clearly superior. The two-factor model was favored, however, because it is more parsimonious and consistent with current theory, and because of high correlations between hyperactive and impulsive factors in the three-factor models. Invariance across nations is consistent with previous studies and supports ADHD as a universally valid syndrome rather than a cultural construct. These results add to the limited knowledge of assessment of ADHD symptoms in Japan. © 2011 Elsevier Ltd.