When the outcome of a physical test has serious consequences, do participants attribute their success or failure to external factors such as luck and poor testing conditions or internal factors like ability and effort? Roberts and Pascuzzi’s 1979 model for sport attribution categorized attributions along 2 dimensions: locus of control (internal and external) and stability (stable and unstable). The purposes of this study were to determine what elements of the 1979 model were used by subjects following physical fitness testing, to reveal differences that may exist in attributions made following goal achievement versus failure, and to suggest modifications in the model for use in physical testing as necessary. Specifically, 123 participants each completed 14 physical tests assessing health- and skill-related components of physical fitness during a semester-long course. Prior to attempting each test, all participants established a personal goal. After completing the tests, participants reported by open-ended survey the primary reason why they did or did not reach their goals. Analysis of all attributions (N=1393) revealed that the Roberts and Pascuzzi model could not categorize 17% of the responses. Attributions to ability, effort, practice, psychological factors, and task difficulty comprised 69% of all responses, while coaching, unstable ability, teamwork, luck, and officials made up only 5% of responses. For all participants who attained their goals, practice and experience combined for 58% of the responses. When goals were not attained, 76% of attributions were made to ability, task difficulty, and practice factors. Further analysis of course grades showed that good and bad performers created similar attribution profiles. In conclusion, while the Roberts and Pascuzzi model found a dimension for a majority of responses, a significant number of attributions resulting from actual sport experience can be placed into new categories. Genetic ability and experience proved to be useful additions, while categories such as unstable ability and teamwork were hardly applied. Together these findings indicate that real situations can be evaluated according to the elements suggested by Roberts and Pascuzzi, but that even those broad elements failed to encompass the range of attributions created by simple physical tests. The use of an open-ended questionnaire does seem to be appropriate, and the true usefulness of this line of research may be revealed if application can be made for the purpose of finding motivational strategies for athletic performance.