School achievement attributions for success and failure were examined for American and Polish preadolescents between the ages of 12 and 13 years. The American sample included 115 preadolescents (53 males and 62 females), and the Polish sample included 64 children (34 males and 30 females). The American subjects came from a suburban racially heterogeneous school district near Cincinnati (Ohio), and the Polish subjects were located in the state schools of Gdansk. Data were obtained from both subject groups during the 1987-88 school year. The construct validity for a revised version of the Multidimensional Multiattribution Causality Scale of H. M. Lefcourt and others was studied. While significant differences between the two cross-national samples and within the eight subscales were reported, significant correlations of the rank orderings of the eight subscale means among the American and Polish samples were believed to be evidence for the construct validity of the revised instrument. Additional correlational analyses contrasted these samples with a sample of 87 Anglo and 64 Hispanic American adolescents responding to the original instrument. Since the data support predictions based on the self-worth motive of M. V. Covington (1984), it is argued that construct validation of the revised instrument was achieved. Since construct validity was found in both samples, it is suggested that a successful translation was accomplished. Seven tables and one bar graph present study data. (Author/SLD)