In the United States, low-income 4^th graders score 31 points below their high-income peers on basic literacy tests. This gap decreases just 3 points by 8^th grade, and predicts later differences in educational attainment and employment. Indeed, while nearly 14% of children growing up in the lowest income quartile drop out of high school, just 2% of children born into the highest income quartile do so (NCES, 2011; 2012).

Educational success does not depend on cognitive ability alone. Non-cognitive skills such as perseverance, self-regulation, and self-esteem predict academic achievement and attainment, as well as related outcomes, such as labor-market attachment and wages (Heckman & Rubinstein, 2001; Heckman, Stixrud, & Urzua, 2006). Low-income youth fare worse on these socio-emotional skills than their high-income peers (McLoyd, 1998; Duncan, Yeung, Brooks-Gunn, & Smith, 1998), a difference which further undermines their academic success. Interventions that enhance non-cognitive skills may thus help close the achievement gap by helping low-income youth develop the socio-emotional skills that enable classroom success.

One much-touted approach to closing the achievement gap is the expansion of charter schools (CS). Because CS operate largely outside of government control, advocates argue they respond more flexibly and thus effectively to students’ academic needs. After 15 years of evaluation, research shows small positive effects of CS for academic outcomes, and that these impacts are more pronounced for low-income youth (Betts & Tang, 2011). However, no studies to date have looked at the socio-emotional impacts of CS despite its centrality to academic achievement.

CS may be uniquely positioned to improve students’ socio-emotional skills. School quality in general is associated with students’ socio-emotional outcomes (Bryk & Driscoll, 1988; Silver, Measelle, Armstrong, & Essex, 2005); thus, to the extent that CS may be higher in quality because of flexibility and responsiveness, CS may improve youth socio-emotional functioning. Moreover, previous research documenting the importance of schools for low-income youth suggests that, like academic outcomes, this effect may be more pronounced for low-income students who lack socio-emotional competence (Masten, 2001).

The present study addresses this question using a nationally representative dataset and a matching technique designed to simulate experimental conditions. Data are drawn from the Early Childhood Longitudinal Study- Kindergarten Cohort. The analytic sample is limited to individuals with full data on CS enrollment and outcome variables at 5^th grade, (N=10, 203). Using propensity score matching, a nearest neighbor one-to-one matched sample (N=3,752) is created. Internalizing, externalizing and peer competence outcomes are regressed on a variable indicating length of time spent in a CS, early or concurrent poverty, and interactions between poverty measures and time spent in a CS. All covariates used to generate propensity scores are included in each model. 

Initial results suggest that more time in CS reduces externalizing behavior problems and increases peer competence, but only for children who experience early disadvantage. While effect sizes are modest, about one-tenth of a standard deviation per year in charter, results suggest that CS enhance the socio-emotional functioning of students most at risk for emotional dysfunction.