Essays on the effectiveness of policies and practices for reducing cognitive gaps between linguistic groups and socioeconomic groups

Academic achievement gaps along racial, socioeconomic, and linguistic lines are perhaps some of the most pressing concerns in the field of education. This dissertation consists of three related papers studying different practices and policies, ranging from before students enter kindergarten though high school, that could conceivably reduce some achievement gaps. In each paper, I explain how the implementation of the practice is expected to differentially affect different children (e.g., children growing up in poverty or not; language-minority or native-language speakers; elementary-school or secondary-school students). This heterogeneity of effects is important because it is through the differential effects that we can close achievement gaps, by capitalizing on instructional practices that work best for certain low-achieving students.;The first paper examines the effects of reading to infants on various measures of language development. Using propensity score matching and a nationally-representative sample of infants, I find significant effects of increased days of parent-child reading on the language development of children not growing up in poverty, but no effects on children growing up in poverty. This finding suggests that increased frequency of current parent-child reading practices for all children would not reduce preschool achievement gaps, and further research into why reading frequency is not as effective for children in poverty is needed.;The second paper tests if the level-appropriate, small-group reading instruction received in ability groups in kindergarten and first-grade is more beneficial to language-minority children than children from homes where English is primarily spoken. The hypothesis is that this type of instruction will act as a compensating factor for limited exposure to English in the home for language-minority children. Results from different specifications of hierarchical linear models all suggest that language-minority Hispanics benefit more from ability grouping than do white children, suggesting ability grouping in the first two years of formal schooling can serve to diminish linguistic achievement gaps.;The third and final paper examines the effects of changing English learners' status to "fluent English proficient"---a process known as "reclassification." Along with the label change, is a discontinuation of English language development (ELD) services. Reclassification is a relatively unstudied topic, although---in some form---it affects the nation's almost 6 million English learners. A common concern with reclassification is the timing of the ELD service removal, such that services are provided as long as children are still benefiting from them, balanced against the concern that ELD services are provided too long and that the instructional time could be better spent on advanced curricula. Results from regression discontinuity analyses show a positive effect of reclassification on achievement in late-elementary and early-middle school, but no effect on high school achievement. Moreover, there is no evidence that reclassification alone has a positive effect on advanced course-taking in high school, contrary to some of the existing literature and concerns of school administrators. Through empirical analyses, this paper illustrates the importance of both (a) removing selection bias from policy decisions related to English learners and (b) creating learning environments that foster a smooth transition from English learner status to fluent English proficient status.