The purpose of this study was to revise and revalidate the scoring procedure of the DISCOVER Mathematics Assessment to allow evaluators to better measure creative problem solving ability in mathematics, identify gifted students, and evaluate the programs developed for fostering creative problem solving. The data for this study consisted of 233 students selected from five different grade levels. I conducted descriptive statistics and regression analyses to compare the relationships of both the original and revised versions of the scoring system with general creativity. I found that range increased from the original to the revised version of the scoring system for mathematical problem solving performance in semi-open-ended problems, however it decreased for overall performance and performance in open-ended problems. Variance, on the other hand, increased for both overall problem solving performance and performance in semi-open-ended problems, and decreased for performance in open-ended problems from the original to the revised version of the scoring system. Furthermore, in the revised model all three variables of the creative mathematical problem solving performance (overall performance, performance in semi-open-ended problems, and performance in open-ended problems) explained more variance in general creativity than the original version. Statistically, the differences between the original and the revised versions were significant for all three variables, except for creative mathematical problem solving performance in open-ended problems. Across grade levels, I found that for the group Lower Grade Levels (grade levels 1 and 2), the explained variance in general creativity increased from the original to the revised version for both overall performance and performance in semi-open-ended problems. However, it decreased for performance in open-ended problems. On the other hand for the group Higher Grade Levels (grade levels 3, 4, and 5) the explained variance in general creativity increased for all three variables from the original to the revised version. Statistically, the only significant difference between the original and the revised versions was for overall problem solving performance in Higher Grade Levels. I concluded that the revised version of the scoring system was more effective when predicting variance in general creativity for overall problem solving performance, and performance in semi-open-ended problems. Also, it predicted more variance in general creativity for the group Higher Grade Levels than the group Lower Grade Levels. Therefore, I suggested that quality should be considered as well as fluency, flexibility, and originality when scoring assessments for creative problem solving ability in mathematics.