This paper examines the cognitive and non-cognitive skills gaps among children and their evolution between grades Kindergarten to 8^th grade. We analyze educational inequalities depending on the children's race/ethnicity, socio-economic status, immigration background, and language spoken at home. The analysis uses the Early Childhood Longitudinal Study-Kindergarten Class of 1998, which allows measuring how gaps evolve over the school years for Kindergarteners who started school in the Fall of 1998, to Spring of 8^th grade, 2007.

Previous research on race and educational gaps has emphasized the importance of race as a determinant of educational inequalities. The work of Fryer and Levitt (2002, 2005) on the black and White gap in the early years of schooling concluded that Blacks lose substantial ground relative to other races, including Hispanics. Nores (2008) inquired into distributional inequalities in contrast to race showing that SES gaps are larger than race gaps and persist over time (controlling for systematic school differences) and although there is a larger Hispanic-White cognitive gap than a Black-White gap at kindergarten entry, over time this is inverted (as confirmed by Reardon and Galindo, 2009). Yet, the interactions between educational outcomes, national origin, immigrant generational status, and English proficiency, beyond race and economic status, need to be more fully understood. Similarly, a comprehensive assessment of educational gaps that includes cognitive and noncognitive dimensions could contribute to uncover some potential sources of disadvantage among children.

Our paper mostly focuses on Hispanics, for whom both the language and the immigrant backgrounds play important roles, and on the Hispanic-White skills gaps. We further differentiate between native Hispanics (i.e., born in the U.S. to parents also born in the U.S.) from immigrant Hispanics (i.e., those born in the U.S. to at least one immigrant parent or those not born in the U.S.), and English Language Learners Hispanics (who don’t speak English at home, or ELL) from Hispanic English speakers. We compare the outcomes –cognitive and noncognitive- of the different language/immigrant status Hispanic subgroups to White children’s, and assess how they change over the school years.

We find that disparities between ELL non-immigrant Hispanic and White children diminish over time in mathematics, and revert in the non-cognitive dimensions, but ELL non-immigrant Hispanic children continue to underperform White children throughout the years in reading. ELL children with immigrant background evidence the largest disadvantage in skills, and English-speakers with non-immigrant background children evidence a sharp improvement over time. Hispanic ELL non-immigrant and White children diminish over time in mathematics, and revert in the non-cognitive dimensions, but ELL non-immigrant Hispanic children continue to underperform White children throughout the years in reading; Among Hispanics, ELL children with immigrant background evidence the largest disadvantage in skills, and English-speakers with non-immigrant background children evidence a sharp improvement over time. In a context with increasing immigrant youth and Hispanic youth demographics, understanding how Hispanics perform early in school, how gaps evolve over time and how language and immigrant background matters, is increasingly important for education policy and allocation of resources.