Searching for a New Standard: The Implications of Using Self-Perceived Social and Emotional Skills as an Indicator of School Performance

Abstract

This study analyzes the relationship between social and emotional (SE) skills and test scores in order to determine whether measures of SE skills should be used as a non-academic indicator of school performance under the Every Student Succeeds Act (ESSA). We examine this relationship using experimental data from Higher Achievement (HA), a demanding, high-standards out-of-school-time (OST) program that provides academic enrichment and high school placement services to underserved youth in the Washington D.C. area. Simple correlations show a positive, direct relationship between self-reported SE scores and test scores suggesting if schools can increase perceived SE skills, student test scores would increase as well. This implies perceived SE skills would make a fitting indicator of school performance. However, there is a set of institutions that believe challenging individuals’ self-perceptions is a critical step in motivating change (e.g. KIPP, National Guard Youth Challenge). This would suggest a temporarily negative relationship between self-reported SE skills and academic performance. Our study will investigate these two hypotheses. The treatment group students who were offered a spot in the HA program experienced a negative shock to perceived SE skills upon entering the program due to the challenging environment. We examine if this exogenous shock to perceived SE skills was causally related to the later increase in test scores for the students. Our data does not provide strong enough evidence to accept either hypothesis. We found no evidence that the negative change in perceived SE skills was causally related to increases in test scores. Instead, we find a numerically small, positive relationship between the SE measures and test scores, but it is not statistically significant. Perhaps there is a positive relationship between SE skills and academic achievement that self-report does not measure, but it is likely to be smaller than the simple correlation shows. Even if there is a positive relationship between actual SE skills and academics, self-reported SE skills might not be a good measure of these true SE skill levels. Thus, we conclude we do not know enough about the relationship and these measures to recommend SE skills as a performance indicator under the ESSA.