Purpose
Metrics used in admissions, such as GPA and GRE, are insufficient at identifying students that will be most successful. Hence, there is a focus on measuring non-cognitive factors of applicants. One such method, the computer-based assessment for sampling personal characteristics (CASPer), is used by many programs in the admissions process but there is no published evidence of score being predictive of success in PA school.

Methods
This study determined the strength of CASPer in predicting PA program outcomes. For two cycles, applicants were required to take the CASPer; however, the score was not directly used in their evaluation. Pearson correlation and multivariable linear regression were used to determine the association between CASPer score and metrics of program performance.

Results/Outcomes
CASPer had a measurable impact on multiple metrics of program success, independent of measures of cognitive ability. While it was not predictive of scores in traditional didactic courses, it was weakly predictive of performance in our new team-based learning curriculum, a meaningful difference in predictive performance. Additionally, an increase in CASPer score by 10 percentile points was associated with a 2.1-point increase on end of rotation exams during the clinical year with the strongest effect at the start. CASPer, however, was not predictive of success on summative or licensing exams.

Discussion
This study provides the first evidence that it predicts success during PA school. The strongest relationship appears to be on end of rotation exams and coursework using team-based learning. These effects are independent of cognitive metrics of applicant readiness. This suggests that traits assessed by the CASPer, such as motivation, collaboration, and problem solving, are important to performance, especially for independent learning or in the transition to clinical year. While the reported results are strong, they may not be generalizable to all programs.