36 ECONOMIC POLICY INSTITUTE | JUNE 17, 2015 PAGE 21
White-Hispanic gaps The observed patterns suggest a significant disadvantage for Hispanic children versus white children across all the cogni- tive variables, with the degree depending on their use of English at home. 37 While both Hispanic subgroups—ELL and English speakers—perform worse than white children, absolute gaps for ELL children are much larger than for English speakers. Indeed, unadjusted ELL Hispanic-white gaps are the largest estimated gaps among any nonwhite-white pair- ing across all cognitive skills and parents’ reported noncognitive skills (the unadjusted differences are between about 0.2 and 0.3 sd for English speakers in reading and math, but near 0.7 sd for ELL). After all controls (SES, family characteristics, etc.) are added, these cognitive gaps relative to white children shrink by about half for English speakers, to 0.08 sd in reading and 0.16 sd in math. And, though they shrink even more for ELL students, because they are so large to begin with, they remain sizeable—0.20 sd in reading and 0.26 sd in math. Both Hispanic subgroups also lag behind their white peers in working memory (their performance is 0.2 sd and 0.4 sd behind their white peers for English speakers and ELL, respectively). Interestingly, the gap between cognitive flexibil- ity of Hispanic ELL children and white children loses its significance after all controls are included (falling from -0.38 sd under the unadjusted model to a statistically insignificant -0.08 sd) but remains significant among English speakers (falling from -0.11 sd to -0.08 sd). Turning to noncognitive skills, after controls are introduced, any differences between English-speaking Hispanics and white children disappear, with three exceptions: compared with white children, they have a relative advantage in cre- ativity and eagerness to learn as perceived by parents and a relative disadvantage in persistence in completing tasks as reported by teachers. 38 In contrast, parents of Hispanic ELL children perceive their children to be substantially disadvantaged compared with whites in social interactions, approaches to learning, and persistence, after all controls are introduced. 39 For this group, too, moderate initial gaps in self-control and creativity as reported by parents vanish after the incorporation of con- trols. 40 Similarly, once controls are introduced to teachers’ assessments, gaps between white students and non-English- speaking Hispanics become insignificant, for all of the considered outcomes. So teacher-parent gaps operate similarly for Hispanic ELL and Asian children, although they apply to different behavioral skills, and somewhat opposite to how they operate for black children, suggesting an interesting avenue for further analysis (see footnote 32).
You've reached the end of your free preview.
Want to read all 93 pages?