In major motivational theory, attributions and ability conceptions have been widely recognized as important constructs that mediate individuals’ motivational patterns in achievement contexts. According to attribution theory, natural ability, effort, and other factors such as luck, mood, and task difficulty are the causes of successful final performance (Graham, 1991). Literature has demonstrated that there were individual differences in attributions and individuals’ attribution patterns would vary according to their ability conceptions (e.g. Hong, Chiu, Dweck, Lin, & Wan, 1999). Because of the unit-sum constraint characteristic, the general approach (e.g., ANOVA) should not be used for the inferential data analysis. The purpose of this study was to use the new approach introduced by Aitchison (1986) to investigate individuals’ attribution patterns and how they would vary with regard to their ability conceptions. The participants in this study were 68 female college students and the data collection procedure was approved by HSIRB. The participants were asked to work on a novel task called lunastix. The goal of lunastix is to lift the baton off ground, catch and release between the control handles in a back and forth manner. The Conceptions of Natural Ability in Object Manipulation Questionnaire (Wang & Biddle, 2001) was administered followed by the 20-minute independent practice session. Then, the percentages of the natural ability, effort, and the mood, luck, and task difficulty were recorded as the composition of the successful final skill performance in Lunastix (Dweck, 2002). Data were analyzed using the compositional data analysis with the SAS software. The additive logistic transformation technique was employed to transform each composition to its logratio composition. The residuals showed that the transformed data met the assumption of the multivariate normal distribution. Overall, participants were more likely to attribute successful final performance to effort (M=64.22, SD=20.71) as compared to natural ability (M=27.89, SD= 20.68), and factors mood, luck, and task difficulty (M=7.89, SD=6.03), which was verified by the ternary diagrams generated from SAS. Participants’ attribution patterns did not significantly vary with regard to either an incremental conception (F (2, 45) =0.33, p>0.72) or the entity conception of ability (F (2, 45) =0.14, p>0.87). The results lead to the conclusion that not only the role of natural ability in the final performance is important, but also the efficacy of effort is also recognized as an important factor. Participants’ attribution patterns in this study did not significantly vary regard to their ability conceptions.Keyword(s): assessment, measurement/evaluation, research