Association for Public Policy Analysis & Management

Recent research has shown that noncognitive skills (e.g., grit, self-efficacy, growth mindset) are important for shaping young adults' future. This research has contributed to a rapid rise in the profile of noncognitive skills in the field of education, as evidenced by the incorporation of noncognitive skills in school assessments, evaluations, and accountability metrics. The degree to which these skills are important for students’ later success is still an open question, however, largely due to lingering issues regarding how best to measure noncognitive skills. Two broad approaches (student self-report surveys and administrative data proxies, such as grades and attendance) show promise in terms of measurement, but further work is needed given the limitations of each approach. For example, self-report surveys are subject to known biases such as socially desirable responding, and administrative data are subject to problems with accurate recording and grade inflation. This paper compares the predictive power of administrative and survey measures of noncognitive skills for long-term student outcomes, such as high school graduation and college enrollment. In addition, the paper assesses whether determinations about how to best measure noncognitive skills depend upon the outcome that is measured or the grade during which students’ skills are measured.